University of Southern California Dissertations and Theses

Advancing energy recovery from food waste using anaerobic biotechnologies: performance and microbial ecology

(USC Thesis Other)

Advancing energy recovery from food waste using anaerobic biotechnologies: performance and microbial ecology

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Yamrot Mulugeta Amha
PI: Dr. Adam L. Smith
Conferring Major/Program: Environmental Engineering
Degree being conferred: Doctor of Philosophy (ENVIRONMENTAL ENGINEERING)
Conferred by: FACULTY OF THE USC GRADUATE SCHOOL
University of Southern California
Degree conferral date: August, 2019
Advancing Energy Recovery
from Food Waste using
Anaerobic Biotechnologies:
Performance and Microbial
Ecology

ii

Acknowledgments
First, I would like to sincerely thank my advisor Dr. Adam L. Smith, for his guidance, sound
evaluation, and strong commitment to my success. Dr. Smith had been an extremely
supportive advisor and he has helped me grow to become a more confident researcher. I
would like to extend my gratitude to my research committee members: Dr. Amy E.
Childress, Dr. Daniel L. McCurry, Dr. Kelly Sanders, and Dr. James Mofett for their constant
input throughout the completion of this project. I would like to thank Dr. Childress for her
mentorship and opening doors for me to advance my career.
My greatest appreciation goes to the undergraduate and graduate students who worked
on this project, Wenjun Zhao, Jia Wang, Pooja Sinha, Juhe Liu, Qin Dong, Rajuan Nelson,
Christina Najm, Alexis Barge, Jewls agman, and Karim Taha. They worked long hours in
the lab and showed great enthusiasm to make sure we succeeded. I want to thank all the
students in Biegler Hall and the Smith Research Group. I would like to especially thank Dr.
Moustapha Harb for his great support when starting my reactors in the lab. Thank you to
my extremely supportive parents, my beloved six siblings, and my partner who foster my
growth in every way. I would like to dedicate this dissertation to my mother, Saba Haddis,
who taught me how to work hard from early age and whose unconditional love is my
inspiration behind everything I do. I would also like to acknowledge the following sources
of financial support for this research: USC Provost Fellowship, Teh Fu Yen Fellowship, Phi
Beta Kappa International Scholarship, and National Science Foundation Grant
Opportunities for Academic Liaison with Industry (GOALI) research fund.
iii

Contents
_____________________________________________________________________

Abstract ...................................................................................................................... 1
1. Introduction ............................................................................................................... 3
1.1 Background ...................................................................................................... 3
1.2 Overview of Dissertation ................................................................................. 7
1.3 Literature Cited .................................................................................................... 7
2. Inhibition of Anaerobic Digestion Processes: Applications of Molecular Tools ........ 9
2.1 Abstract ................................................................................................................ 9
2.2 Introduction .......................................................................................................... 9
2.3 Inhibition of metabolic pathways in AD ............................................................. 13
2.3.1. Hydrolysis and fermentation ...................................................................... 13
2.3.2 Syntrophy and methanogenesis .................................................................. 14
2.4 Targeted nucleic acid biomarkers ...................................................................... 26
2.4.1 Universal target: 16S rRNA gene and metadata analysis ............................ 26
2.4.2 Functional genes .......................................................................................... 32
2.5 Omics studies ..................................................................................................... 33
2.5.1. Metagenomics ............................................................................................ 33
2.5.2. Metatranscriptomics .................................................................................. 35
2.5.3. Metaproteomics ......................................................................................... 37
2.6 Mapping substrate utilization ............................................................................ 38
2.7 Real-time monitoring ......................................................................................... 41
2.7.1. Reporters .................................................................................................... 41
2.7.2. MinION........................................................................................................ 42
iv

2.8 Mitigation strategies .......................................................................................... 44
2.9 Conclusion .......................................................................................................... 47
2.10 Literature Cited ................................................................................................ 49
3. Elucidating Microbial Community Adaptation to Anaerobic Co-Digestion of Fats, Oils,
and Grease and Food Waste ........................................................................................ 58
3.1 Abstract .............................................................................................................. 58
3.2 Introduction ........................................................................................................ 59
3.3 Materials & materials ......................................................................................... 63
3.3.1 Sample collection and chemical assays ....................................................... 63
3.3.2 Bench-scale anaerobic respirometry ........................................................... 64
3.3.3 Nucleic acids extraction and cDNA synthesis .............................................. 66
3.3.4 PCR and sequencing .................................................................................... 66
3.3.5 Reverse transcription-quantitative PCR (RT-qPCR) ..................................... 67
3.3.6 Reagent controls .......................................................................................... 67
3.4 Results and discussion ........................................................................................ 68
3.4.1 Co-digestion of FOG and FW reproducibly increased methane production68
3.4.2 Syntrophic fatty-acid oxidizers were critical to increase methane production
.............................................................................................................................. 72
3.4.3 High FOG addition resulted in significant decline in performance and inhibited
key microbial populations .................................................................................... 81
3.5 Conclusions ......................................................................................................... 90
3.6 Literature Cited .................................................................................................. 92
4. Two-phase improves performance of anaerobic membrane bioreactor treatment of
food waste at high organic loading rates .................................................................... 97
4.1 Abstract .............................................................................................................. 97
4.2 Introduction................................................................................................... 98
4.3 Materials & Methods .................................................................................. 100
4.3.1 Bench-scale AnMBR configuration ............................................................ 100
v

4.3.2. Inoculation and operational parameters ................................................. 101
4.3.3 Chemical assay and sampling .................................................................... 102
4.3.4 Microbial community analysis ................................................................... 103
4.4 Results and Discussion ..................................................................................... 104
4.4.1. TP-AP effectively increased VFAs concentration and enriched a distinct
microbial community from FW ........................................................................... 104
4.4.2. TP AnMBR resulted in improved performance relative to SP AnMBR at high
OLRs and >98% COD removal efficiency was achieved ...................................... 106
4.4.3. TP enriched for syntrophic fatty-acid oxidizing bacteria while stable
methanogen activity suggested functional redundancy .................................... 109
4.4.4. Microbial activity data revealed increased diversity in TP-MP at high OLRs and
distinct community profile ................................................................................. 111
4.4.5. Fouling was severe, but reversible at increased OLRs, with SP and TP-MP
biofilms exhibiting similar community structure and activity profiles ............... 119
4.5 Literature Cited ................................................................................................ 123
5. Co-digestion of FOG improves performance of anaerobic membrane bioreactor
treatment of food waste ............................................................................................ 127
5.1 Abstract ............................................................................................................ 127
5.2 Introduction ...................................................................................................... 128
5.3 Materials and Methods .................................................................................... 130
5.3.1 Bench-scale AnMBR configurations .......................................................... 131
5.3.2 Inoculation and operational parameters .................................................. 132
5.3.3 Chemical assays and sampling ................................................................... 134
5.4 Results and Discussion ..................................................................................... 135
5.5 Literature Cited ................................................................................................ 142
6. Conclusions ............................................................................................................ 144
6.1 Overview........................................................................................................... 144
6.2 Diverse molecular tools are elucidating inhibitor impact on the AD microbiome..
145
vi

6.3 Co-digestion of FW and FOG can increase energy recovery in WWTPs .......... 146
6.4 Two-phase AnMBRs increased energy recovery and microbial community
resiliance at high organic loading rates .................................................................. 147
6.5 Co-digestion of FOG improves performance of anaerobic membrane bioreactor
treatment of food waste ........................................................................................ 150
6.6 Future research ................................................................................................ 151
6.7 Literature Cited ................................................................................................ 154
APPENDIX A ................................................................................................................ 155
A. Supplementary Information ............................................................................... 155
SI 1.0 Chapter 2 ...................................................................................................... 155
SI 1.1 Tables ........................................................................................................ 155
SI 1.2 Figures ....................................................................................................... 157
SI 2.0 Chapter 3 ...................................................................................................... 158
SI 2.1 Materials & Methods ................................................................................ 158
SI 2.2 Figures ....................................................................................................... 164
SI 2.3 Tables ........................................................................................................ 182
SI 3.0 Chapter 4 ...................................................................................................... 184
SI 3.1 Tables ........................................................................................................ 184
SI 3.2 Figures ....................................................................................................... 185

vii

List of Tables
_____________________________________________________________________
Table 1. Studies investigating volatile fatty acid (VFA) inhibition. .............................. 19
Table 2. Studies investigating ammonia inhibition. ..................................................... 20
Table 3. Studies investigating long chain fatty acid (LCFA) inhibition. ........................ 22
Table 4. Miscellaneous studies investigating inhibition. ............................................. 24
Table 5. Substrate volatile solid loading for bench-scale anaerobic respirometry. Each
condition was run in triplicate vessels, containing 10 g L
-1
of TVS. ............................. 65
Table 6. Non-linear regression analysis on all substrate mixtures in Run 1, 2, and 3. 70
Table 7. TVS removal, free ammonia concentration, pH, VFA concentration, sulfate
concentration, and sulfate reduction for end of Run 3. The detection limit for the IC
analyses was 10 mg L
-1
. ................................................................................................ 71
Table 8. TVS removal, free ammonia concentration, pH, VFA concentration, sulfate
concentration, and sulfate reduction for end of Run 5. The detection limit for the IC
analyses was 10 mg L
-1
. The row marked 40% FOG* shows the results of the outlier
sample in the 40% FOG condition................................................................................ 84
Table 9. Spearman rank correlation in Run 5 for methane production and microbial
activity of prominent groups normalized to total 16S rRNA. ...................................... 90
Table 10. Operating conditions for SP and TP-MP..................................................... 134

viii

List of Figures
_____________________________________________________________________
Figure 1. Process flow of full-scale AnMBR plant treating food waste from Ralph’s and
Food4Less....................................................................................................................... 5
Figure 2. Metabolisms and key microbial populations in anaerobic processes. ......... 12
Figure 3. (A) Microbial community structure by class over increasing VFA concentration
(left to right). The y-axis represents relative abundance of OTUS that are 0.1% or greater
in the community. Stacked bars within each class (same color) represent orders. (B)
Microbial community structure by class over increasing LCFA concentration (left to
right). The y-axis represents relative abundance of OTUS that are 0.1% or greater in the
community. Stacked bars within each class (same color) represent orders. .............. 29
Figure 4. (A) Core microbiome by class over increasing LCFA concentration (left to right).
The y-axis represents abundance relative to other populations within the core
microbiome and is therefore independent of populations not represented in the figure
legend. Stacked bars within each class (same color) represent orders (B) Core
microbiome by class over increasing VFA concentration. Table S1 defines the
concentration ranges for LCFAs and VFAs on the x-axis of all graphs. ........................ 30
Figure 5. (A) Cumulative biogas production normalized to initial organic loading in mL/g
TVS for Run 1, Run 2, and Run 3. The error bars indicate the standard deviation every 20
hours for triplicate vessels. (B) Cumulative biogas (solid fill) and methane (pattern fill)
for the substrate mixtures, normalized to initial organic loading in mL/g TVS. Error bars
indicate the combined standard deviation of cumulative biogas/methane production
and initial organic loading TVS measurement for triplicate vessels. Due to experimental
error, one of the triplicate vessels for the following substrate mixtures and runs were
excluded from the mean analysis and standard deviation: PS+TWAS+FOG (in Run 1),
PS+TWAS+FW (in Run 3), and PS+TWAS+FOG+FW (in Run 3). Therefore, each bar
represents the mean of only two vessels for these substrate mixtures and runs. ..... 71
Figure 6. (A) Relative activity based on 16S rRNA sequencing identified at the genus level
where possible for PS+TWAS+FOG+FW for Run 1, Run 2, and Run 3 (B) Relative activity
ix

in at the end of Run 3 for the different substrate mixtures. For Run 1 and Run 3,
triplicate and duplicate samples are shown, respectively, to represent methodological
precision. All data are expressed as a percentage normalized using total 16S rRNA
sequences (Bacteria and Archaea). A y-axis break was used to accentuate differences in
lower activity populations. .......................................................................................... 73
Figure 7. Relative expression of mcrA in all substrate mixtures for Run 1-3. Copies of
mcrA transcripts were normalized to total 16S rRNA copies. Error bars for mcrA
expression represent the standard deviation of the ratio of triplicate RT-qPCR reactions.
Error bars for cumulative methane production (secondary y-axis) represent the standard
deviation for triplicate vessels. .................................................................................... 75
Figure 8 (A) Relative activity of methanogens identified at the genus level where
possible based on 16S rRNA sequencing and (B) relative activity of syntrophic fatty-acid
oxidizers identified at the genus level where possible using 16S rRNA sequencing.
Results are expressed as a percentage normalized using total of 16S rRNA sequences
(Bacteria and Archaea). Truncated y-axes (0 to 1% and 0 to 6% on figure A and B,
respectively) are shown to accentuate differences in abundance. ............................. 77
Figure 9. Cumulative biogas (solid fill) and methane (pattern fill) with increasing FOG
addition normalized to initial TVS loading. One of the triplicates vessels for 40% FOG
showed an outlier in biogas production compared to the other replicates and was
excluded from the mean analysis in this figure. Error bars for cumulative
biogas/methane production indicate the combined standard deviation for gas
production and initial organic loading in g TVS for triplicate vessels for each substrate
mixture. The error bars for relative mcrA gene expression represent the standard
deviation of the ratio of triplicate RT-qPCR reactions. ................................................. 83
Figure 10. Top 30 most active OTUs classified at the genus level for Run 5 (FOG
inhibition study) at the beginning and end of run. Duplicate results are shown for the
end of the run to represent methodological precision. 40% FOG(1)* shows results of a
vessel with low biogas production. All data are expressed as a percentage normalized
using total 16S rRNA sequences (Bacteria and Archaea). ........................................... 86
Figure 11. (A) Relative activity of methanogens identified at the genus level based on
16S rRNA sequencing and (B) relative activity of syntrophic fatty-acid oxidizers identified
at the genus level based on 16S rRNA sequencing. Results are expressed as percentages
normalized to the total 16S rRNA sequences (Bacteria and Archaea). Duplicate results
are shown for the end of the run to represent methodological precision. 40% FOG(1)*
shows results of a vessel with low biogas production. Truncated y-axes (0 to 0.9% and 0
x

to 25% on figure A and B, respectively) are shown to accentuate differences in
abundance. .................................................................................................................. 88
Figure 12. Methane production rate (primary-axis) and methane per OLR (secondary y-
axis). Solid fill shows SP and pattern fill signifies TP-MP. The number after SP or TP-MP
shows the OLRs for each system, for example TP-MP - 2.5 indicates, the mean methane
production for the TP- MP treatment at 2.5 g COD L·d
-1
. SP1 and SP2 are initial runs,
where both AnMBRs were run as SP. The solid error bar indicates 95% confidence
interval of the mean, and the circles indicate standard deviation of mean daily methane
production for each OLRs. The symbol * shows that mean methane production were
significantly different between SP and TP-MP for the specified feeding OLRs, as
indicated by <0.05 p values with the two-tailed t-test. ............................................. 107
Figure 13. (A) Relative activity of syntrophic fatty-acid oxidizers, and (B) relative activity
of methanogens identified at the genus level where possible using 16S rRNA sequencing
for SP and TP-MP, at increasing OLRs. The results shown are average data from three
separate sampling points taken for each OLR. Results are expressed as a percentage
normalized using total of 16S rRNA sequences (Bacteria and Archaea). Truncated y-axes
(0 to 11% and 0 to 50% on figure A and B, respectively) are shown to accentuate
differences in activity. The secondary y-axis and pink diamond shaped data points for
figure A and B signify the ratio of TP-MP to SP for total syntrophs and methanogens,
respectively. ............................................................................................................... 110
Figure 14. Inverse Simpson Index of 16S rRNA gene sequencing results for SP, TP-MP,
and TP-AP at different Organic Loading Rates (OLRs). Inverse Simpson Index measures
richness in community or alpha-diversity. The error bar indicates standard deviation of
triplicate sampling days for each OLR. ....................................................................... 111
Figure 15. (A). Relative activity for genera that showed ≥3% relative activity in at least
one sample in SP, and (B). TP-MP samples. Three distinct groups were formed based on
relative activity change in either 10 or 15 g COD L·day
-1
relative to SP1/SP2, with
increased OLR: (1) decreased by ≥ 50% (red-fill) (2) showed ≤ 50% change (blue-fill),
and (3) increased by ≥ 50% (green-fill). Results are expressed as a percentage
normalized using total of 16S rRNA sequences (Bacteria and Archaea). Truncated y-axes
(0 to 90%) are shown to accentuate differences in activity. ..................................... 114
Figure 16. (A) Taxonomic Cladogram for groups that showed significant differential
relative activity in one of the four categories, low (2.5-3.5 g COD L·day
-1
), medium (5 g
COD L·day
-1
), or high (10 and 15 g COD L·day
-1
) OLRs for SP, and (B) TP-MP samples. All
taxa presented here showed significant differential activity in one of the OLR categories
xi

by resulting in Linear Discriminant Analysis (LDA) score of ≥ 2. The analysis was
conducted with LEfSe tool and relative activity data was used as input for all groups that
showed ≥0.5% relative activity in at least one sample. The highest available
classification level for each OTU is used for labelling. ............................................... 115
Figure 17. (A) Average relative activity of communities that showed significant (p<0.05)
correlation with methane production in SP, and (B). TP-MP, identified at the genus level
where possible based with 16S rRNA sequencing. The relative activity results shown are
average data from three separate sampling points taken for each OLR. Results are
expressed as a percentage normalized using total of 16S rRNA sequences (Bacteria and
Archaea). Truncated y-axes (0 to 50%) are shown to accentuate differences in activity.
The Spearman correlation coefficient values (ρ) are shown as x-axis on the legend. The
secondary y-axis and blue diamond shaped data points for both figure A and B signify
the mean methane production per day for each OLR. For each OLR, samples from three
time points were sequenced, thus, the Spearman rank analysis was conducted using this
triplicate data points for each OLR, i.e. relative activity data for all groups that showed
≥0.5% relative activity in at least one sample with methane production data for that
specific sampling date. ............................................................................................... 118
Figure 18. Schematic of bench-scale AnMBR. ........................................................... 132
Figure 19. Daily methane production (primary y-axis) and biogas composition (secondary
y-axis) for at (A) 0.5 kg m
3
·day
-1
,

(B) 0.75 kg m
3
·day
-1
,

and (C) 1 kg m
3
·day
-1
fats loading
rates. .......................................................................................................................... 137
Figure 20. Methane production rate (primary y-axis) and methane per OLR (secondary y-
axis). Solid fill shows SP and pattern fill signifies TP-MP. The numerical value after SP or
TP-MP shows the fats loading for each system, for example TP-MP - 0.5 indicates, the
mean methane production for the TP-MP treatment at 0.5 kg m
3
·day
-1
. The solid error
bar indicates 95% confidence interval of the mean, and the circles indicate standard
deviation of mean daily methane production for each OLRs. ................................... 138
Figure 21. Methane production rate (primary y-axis) and methane per OLR (secondary y-
axis). Solid dark blue fill shows SP and solid dark red fill signifies TP-MP. The clear fill
show mono-digestion of FW only for SP (blue) and TP-MP (red). The number after SP or
TP-MP shows the fats loading for each system, for example TP-MP - 0.5 indicates, the
mean methane production for the TP-MP treatment at 0.5 kg m
3
·day
-1
fats addition. The
solid error bar indicates 95% confidence interval of the mean, and the circles indicate
standard deviation of mean daily methane production for each OLRs. The symbol *
shows that mean methane production were statistically significantly different between
xii

FW mono-digestion vs. FW and FOG co-digestion for the specified reactor, as indicated
by <0.05 p values with two-tailed t-test. ................................................................... 139
Figure 22. (A) Mass balance analysis for SP based on COD allocation of output relative to
input (%), at different fats addition. (B) Mass balance analysis for TP-MP based on COD
allocation of output relative to input (%), at different fats addition. Complete sulfate
reduction was assumed based on influent sulfate concentration. ........................... 139
Figure 23. Total volatile solids (TVS) concentration (primary y-axis), signified by solid
line, and total solids (TS) concentration (secondary y-axis) signified by broken line for
single-phase (SP) and two-phase methane-phase (TP-MP), at different fats loading. Error
bars for TVS and TS concentrations represent the standard deviation for duplicate
samples. ..................................................................................................................... 140
Figure 24. pH in SP, TP-MP, TP-AP1, and TP-AP2 at different fats loading rates. ..... 140
Figure 25. Chemical oxygen demand (COD) concentration in effluent (primary y-axis) and
COD removal efficiency (secondary y-axis) for different FOG addition for single-phase
(SP) and two-phase methane-phase (TP-MP). ........................................................... 141

xiii

List of SI Tables
_____________________________________________________________________
Table S1. Concentration ranges used to define inhibitory levels of LCFAs and VFAs in
metadata analysis. ..................................................................................................... 155
Table S2. Advantages and limitations of different molecular method. ..................... 156
Table S3. Biomass samples taken for microbial community analyses. The raw samples
were seed, primary sludge (PS), thickened waste activated sludge (TWAS), food waste
(FW) and fats, oils, and grease (FOG). DNA samples are shown in blue and RNA samples
are shown in orange. ................................................................................................. 182
Table S4. Spearman rank correlation analysis for Run 1 – 3, showing significant
correlation (defined as p<0.05) between methane production and microbial community
activity. ....................................................................................................................... 183
Table S5. Experimental conditions for different organic loading rates (OLRs). ........ 184
Table S6. FW characterization. .................................................................................. 184

xiv

List of SI Figures
_____________________________________________________________________
Figure S1. Community structure by class over increasing inhibitory concentration of
volatile fatty acids and long chain fatty acids (left to right). The y-axis represents relative
abundance of OTUS that are 0.1% or greater in the community. Stacked bars within each
class (same color) represent orders. ......................................................................... 157
Figure S2. Relative abundance of genera identified with 16S rRNA gene sequencing for
reagent control samples that were serially diluted using genomic DNA of Thermus
thermophilus. CD0, CD4, CD5, and CD6 indicate undiluted genomic DNA of Thermus
thermophilus, 10
-4
dilution with ultrapure reagent water, 10
-5
dilution, and 10
-6
dilution,
respectively. CLB and CW are control LB and ultrapure reagent water, respectively. CD1,
CD2, and CD3 are not shown due to low sequencing depth. All sequencing results are
shown in percentage normalized to the total 16S rRNA gene sequencing (Bacteria and
Archaea). .................................................................................................................... 164
Figure S3. 16S rRNA gene and 16S rRNA quantification using quantitative PCR (qPCR) and
reverse-transcription RT-qPCR, respectively. The error bar shows the standard deviation
for triplicate runs for each sample. For the highest dilution sample (denoted as 0), 1 mL
of Thermus thermophilus culture in Lysogeny broth (LB) was taken directly, whereas the
subsequent dilution aliquots (denoted dilution-level 1-6) were made by diluting 1:10 the
preceding aliquot with pure LB broth, for each dilution level. In addition, LB and
ultrapure DNase/RNase free water samples were taken for qPCR and RT-qPCR analyses.
.................................................................................................................................... 165
Figure S4. (A) Non-linear regression fitting for PS+TWAS, (B) PS+TWAS+FW, (C)
PS+TWAS+FW, and (D) PS+TWAS+FOG+FW substrate mixtures for Run 2. Vessel 1-3
represent the triplicate substrate mixtures for the run. A is the biogas production
potential (L/kg), λ is the lag phase (d), and the R
2
value shows the non-linear regression
coefficient. ................................................................................................................. 166
xv

Figure S5. Genus-level classification of samples collected from Hyperion Wastewater
Treatment Plant (seed, primary sludge (PS), thickened waste activated sludge (TWAS),
and fats, oils, and grease (FOG)), and from Divert Inc. (food waste (FW)). The DNA-based
and RNA-based sequencing results are shown for all samples, expect for FOG samples,
where only results with RNA-based sequencing are shown. .................................... 167
Figure S6. Relative activity based on 16S rRNA sequencing classified at the phyla level for
Run 1-3 end of run samples with various substrate mixture. All data are expressed as a
percentage normalized using total 16S rRNA sequences (Bacteria and Archaea). A y-axis
break was used to accentuate differences in lower activity populations. ................. 168
Figure S7. Relative abundance based on 16S rRNA gene sequencing classified at the
phyla level for Run 1-3 end of run samples with various substrate mixture. All data are
expressed as a percentage normalized using total 16S rRNA sequences (Bacteria and
Archaea). A y-axis break was used to accentuate differences in lower activity
populations. ............................................................................................................... 169
Figure S8. (A) Relative abundance based on 16S rRNA gene sequencing identified at the
genus level where possible at the end of Run 3. (B) Relative abundance in
PS+TWAS+FOG+FW for Run 1, Run 2, and Run 3. For Run 1 and Run 3, triplicate and
duplicate samples are shown, respectively, to represent methodological precision. All
data are expressed as a percentage normalized using total 16S rRNA gene sequences
(Bacteria and Archaea). A y-axis break was used to accentuate differences in lower
activity populations. ................................................................................................... 170
Figure S9. (A) Relative abundance of methanogens identified at the genus level where
possible based on 16S rRNA gene sequencing and (B) Relative abundance of syntrophic
fatty-acid oxidizers identified at the genus level where possible using 16S rDNA
sequencing. Results are expressed as a percentage normalized using total of 16S rRNA
sequences (Bacteria and Archaea). Truncated y-axes (0 to 0.25% and 0 to 4% on figure A
and B, respectively) are shown to accentuate differences in abundance................. 171
Figure S10. Relative activity of syntrophic acetate oxidizers for Run 1-3, using 16S rRNA
sequencing. Results are expressed as percentages normalized using the total number of
16S rRNA sequencing (including Bacteria and Archaea). A truncated y-axis (0 to 1.6%) is
shown to accentuate differences in abundance. ...................................................... 172
xvi

Figure S11. Methane production for all substrate mixtures in Run 1-3 vs. relative activity
of methanogens and syntrophic fatty-acid oxidizers using 16S rRNA sequencing. The
relative activity data were normalized using total 16S rRNA sequences (including
Archaea and Bacteria). The regression lines are only to aid in visualization of trend and
not to suggest goodness of fit. .................................................................................. 173
Figure S12. Non-metric multidimensional scaling (NMDS) plot for end of Run 1 samples
using (A) 16S rRNA sequencing and (B) 16S rRNA gene sequencing. The error bars for
both plots show the standard deviation for triplicate samples for each substrate
mixture. ...................................................................................................................... 174
Figure S13. Inverse Simpson diversity metric for the total relative activity (Bacteria and
Archaea) based on 16S rRNA sequencing. Error bars are shown only for Run 1 results
and represent the standard deviation of the inverse Simpson metric for triplicate
substrate mixtures (vessels). ..................................................................................... 175
Figure S14. Cumulative biogas production normalized to initial organic loading in
mL/gTVS for Run 5. The error bars indicate the standard deviation every 40 h for
triplicate vessels. One of the triplicate vessels for the 40% FOG+FW condition was
excluded due to an outlier biogas production compared to the remaining replicate
substrate mixtures. .................................................................................................... 176
Figure S15. (A) Relative activity and (B) relative abundance of top 30 most abundant
OTUs identified to the genus level for beginning and end of Run 4. Results are expressed
as percentages normalized using the total number of 16S rRNA sequences and 16S rRNA
gene sequences for A and B, respectively (including Bacteria and Archaea). .......... 177
Figure S16. Relative abundance of top 30 most abundant OTUs identified to the genus-
level for beginning and end of Run 5 using 16S rRNA gene sequencing. Results are
expressed as percentages normalized using the total number of 16S rRNA gene
sequencing (including Bacteria and Archaea). End of run results shown are from
triplicate vessels for each substrate mixture. ............................................................ 178
Figure S17. (A) Relative abundance of methanogens identified at the genus level where
possible based on 16S rRNA gene sequencing. (B) Relative abundance of syntrophic
fatty-acid oxidizers identified at the genus level where possible using 16S rRNA gene
xvii

sequencing. Results are expressed as a percentage normalized using total 16S rRNA
sequences (Bacteria and Archaea). Truncated y-axes (0 to 0.3% and 0 to 9% on A and B,
respectively) are shown to accentuate differences in abundance. ........................... 179
Figure S18. Non-metric multidimensional scaling (NMDS) plot for initial and end of Run 5
samples using 16S rRNA gene sequencing. The error bars show the standard deviation
for triplicate samples for end of run results for each substrate mixture. ................. 180
Figure S19. Comparison of relative activity of methanogens and syntrophic fatty-acid
oxidizers detected in Run 5 with methane production. The total relative activity of
methanogens and syntrophs identified using 16S rRNA sequencing were normalized
using the total number of 16S rRNA sequencing (including Bacteria and Archaea).
Duplicate results for each substrate mixture is shown. The error bar indicates the
standard deviation for methane production where all triplicate vessels for each
substrate mixture were considered (including the outlier biogas production for the 40%
FOG substrate mixture). ............................................................................................ 181
Figure S20. Schematic of TP system consisting of TP-AP and methane-phase (TP-MP). In
single-phase (SP), FW was directly fed (no acid-phase) and it consisted of the
components to the right of the vertical dashed line. ................................................ 185
Figure S21. Volatile fatty acids (VFAs) concentrations in Two-Phase Acid-Phase (TP-AP) at
different Organic Loading Rates (OLRs). The solid line represents VFA concentration
(primary y-axis) and the dashed line represents ratio of VFAs to Chemical Oxygen
Demand (COD) in the FW (secondary y-axis). Error bars for VFA concentrations
represent the standard deviation for triplicate samples. .......................................... 186
Figure S22 (A). pH in Single-Phase (SP), Two-Phase Methane-Phase (TP-MP), and Two-
Phase Acid-Phase (TP-AP) at different Organic Loading Rates (OLRs). (B). Total Ammonia-
Nitrogen concentration in Single-Phase (SP), Two-Phase Methane-Phase (TP-MP), and
Two-Phase Acid-Phase (TP-AP) at different Organic Loading Rates (OLRs). ............. 187
Figure S23. Non-metric multi-dimensional analysis (NMDS) plot showing ordination of
Food Waste (FW) and Two-Phase Acid-Phase (TP-AP) samples (at different Organic
Loading Rates (OLRs)) analyzed using DNA- and RNA-based sequencing. ................ 188
xviii

Figure S24. Relative activity of microbial communities in Food Waste (FW) and Two-
Phase Acid-Phase (TP-AP) based on 16S rRNA sequencing, identified to the genus level
where possible. FW samples were retrieved from the full-scale plant, Divert Inc., weekly
in December, 2017 (denoted as Sample 1-4). TP-AP samples were from the different
Organic Loading Rates (OLRs) at bench-scale and the numbers on the x-axis represent
days after startup of the AnMBR. All data are expressed as a percentage normalized
using total 16S rRNA sequences (Bacteria and Archaea). A y-axis break (at 63%) was
used to accentuate differences in lower activity populations. ................................. 189
Figure S25. (A) Relative activity of methanogens identified at the genus level where
possible based on 16S rRNA sequencing and (B) relative activity of syntrophic fatty-acid
oxidizers identified at the genus level where possible using 16S rRNA sequencing.
Results are expressed as a percentage normalized using total of 16S rRNA sequences
(Bacteria and Archaea). Truncated y-axes (0 to 0.07% and 0 to 0.03% on figure A and B,
respectively) are shown to accentuate differences in abundance. ........................... 190
Figure S26. Mass-balance analysis for SP based on CID allocation of output relative to
input (%), at different OLRs. (B). Mass-balance analysis for TP-MP based on COD
allocation of output relative to input (%), at different Organic Loading Rates (OLRs).
Complete sulfate reduction was assumed based on influent sulfate concentration.191
Figure S27. Total Volatile Solids (TVS) concentration (primary y-axis), signified by solid
line, and Total Solids (TS) concentration (secondary-axis) signified by broken line for
Single-Phase (SP) and Two-Phase Methane-Phase (TP-MP), at different Organic Loading
Rates (OLRs). Error bars for TVS and TS concentrations represent the standard deviation
for duplicate samples. ................................................................................................ 192
Figure S28. (A) Volatile Fatty Acids (VFAs) concentration in Single-Phase (SP) for effluent
(solid-line) and biomass (dashed-line) samples at different Organic Loading Rates (OLRs).
(B). VFAs concentration in Two-Phase Methane-Phase (TP-MP) for effluent (solid-line)
and biomass (dashed-line) samples at different Organic Loading Rates (OLRs). Error bars
for VFAs concentrations represent the standard deviation for triplicate samples. .. 193
Figure S29. Chemical oxygen demand (COD) concentration in effluent (primary axis) and
COD removal efficiency (secondary axis) for different Organic Loading Rates (OLRs) for
Single-Phase (SP) and Two-Phase Methane-Phase (TP-MP). .................................... 194
xix

Figure S30. Relative abundance (primary y-axis) and relative activity (secondary y-axis)
of microbial communities in mixed liquor (ML) samples form full-scale plant (Divert Inc.)
based on 16S rRNA gene and 16S rRNA sequencing, respectively. The microbial
communities were identified to the genus level where possible. ML samples were
retrieved from Divert monthly (June-November, 2017) and weekly on the month of
(denoted as Sample 1-4). Monthly relative abundance data (June - November) are
average values for triplicate samples. All data are expressed as a percentage normalized
using total 16S rRNA gene sequences (Bacteria and Archaea) and 16S rRNA sequences
(Bacteria and Archaea) for relative abundance and relative activity data, respectively.
.................................................................................................................................... 195
Figure S31. Non-metric multi-dimensional analysis (NMDS) plot showing ordination of
microbial community structure (DNA-based) or activity (RNA-based) for SP, TP-MP, and
TP-AP samples. The different fills for SP and TP-MP samples indicate different Organic
Loading Rates (OLRs). The star symbols signify Biofilm samples from SP and TP-MP. The
table shows two-tailed t-test to test the hypothesis that the clustering of ordination in
between different groups is significant, where the p-value<0.05 indicates significant
differential clustering. ................................................................................................ 196
Figure S32. Heat-map showing log Relative Activity (%) of microbial communities in SP
based on 16S rRNA sequencing, identified to the genus level where possible. All data are
expressed as a percentage normalized using total 16S rRNA sequences (Bacteria and
Archaea). .................................................................................................................... 197
Figure S33. Heat-map showing log Relative Activity (%) of microbial communities in TP-
MP based on 16S rRNA sequencing, identified to the genus level where possible. All data
are expressed as a percentage normalized using total 16S rRNA sequences (Bacteria and
Archaea). .................................................................................................................... 198
Figure S34. Relative abundance (for DNA) and relative activity (for RNA) of biofilm
communities in SP and TP-MP at an OLR of 10 g COD L·day
-1
based on 16S rRNA
sequencing, identified to the genus level where possible. All data are expressed as a
percentage normalized using total 16S rRNA sequences (Bacteria and Archaea). ... 199

1

Abstract
Food waste (FW) is one of the largest components of municipal solid waste (MSW), with an
estimated 40% of all food produced in the U.S. being wasted from farm to fork to landfill.
Anaerobic biotechnologies offer an opportunity to recover energy via production of methane-
rich biogas, while also enabling reuse of the embedded nutrients in FW in the form of
biosolids. However, low energy production and vulnerability of these systems to disturbances
are two barriers in their application for FW management at the full scale. In this dissertation,
operating and design strategies were investigated to increase energy recovery and resilience
of anaerobic systems to potentially inhibiting conditions. First, co-digestion of FW and fats,
oils, and grease (FOG), which also originate from food-processing activities, was evaluated as
a strategy to increase energy recovery at wastewater treatment plants. A synergetic increase
in methane production of 26% was observed when FW and FOG were co-digested with
wastewater sludges. RNA-based sequencing indicated that syntrophic fatty-acid oxidizers had
the greatest influence on system performance. Next, the potential for two-phase
(acid/methane) anaerobic membrane bioreactor (AnMBR) treatment of FW was investigated.
We systematically compared single-phase (SP) and two-phase (TP) AnMBR treatment of FW
and characterized the impact of phase separation on microbial community structure and
activity profiles at incrementally increasing organic loading rates (OLRs). The TP system
increased methane production (up to 20.3%) relative to the SP system as OLR increased from
3.5 to 10 g COD L·d-1. At high OLR, activity of syntrophic bacteria in TP was double that of SP.
Our results indicated that AnMBRs in TP mode can effectively treat FW at OLRs up to 10 g
COD∙L day-1 by improving hydrolysis rates, microbial diversity, syntrophic activity, and
enriching more resistant communities to high OLRs relative to AnMBRs in SP mode. Last, FOG
co-digestion was evaluated to improve energy recovery during AnMBR treatment of FW in SP

2

and TP systems, where the upper limits of FOG addition without microbial inhibition and
severe membrane fouling were determined. Anaerobic biotechnologies are expected to play
a major role in FW management in the near future and thus, this body of work provides timely
guidance on operational strategies and design for full-scale anaerobic systems.

3

CHAPTER 1

1. Introduction

1.1 Background
Food waste (FW) is a major component of the organic fraction of municipal solid waste
(MSW), comprising up to 19% of the total solid waste disposed in landfills (Kong et al. 2012).
Recent legislation in California, where an estimated 4 million dry metric tons of food and
processing residuals are landfilled each year (Matteson and Jenkins 2007), dictates that
commercial entities must practice landfill diversion for organic waste. These regulations
impose a minimum threshold of organic waste generation by commercial entities that
decreases over time, resulting in a greater proportion of the commercial sector needing to
comply. These entities will be forced to either compost or anaerobically digest their organic
waste. However, permitting composting facilities within dense cities, such as Los Angeles, is
challenging given the odor and large land requirements. Locating compositing facilities
outside of large cities is also problematic, as transport distances result in environmental
impacts that could potentially outweigh the benefits of landfill diversion altogether. It is
therefore anticipated that anaerobic digestion will play a major role in FW management in
California in the near future. Developing technologies to improve process performance and
enhance biogas production is imperative such that they may be implemented at the full-scale
to maximize the environmental sustainability of landfill diversion. Two potential barriers in
application of anaerobic digestion for FW treatment are low energy recovery and

4

susceptibility to variability of feed substrate. Therefore, in this dissertation we investigated
operational and design strategies that increase performance and improve microbial
community adaptation to operating parameters.
Anaerobic membrane bioreactors (AnMBR) have emerged as a promising technology for FW
landfill diversion relative to composting given their reduced environmental footprint and
ability to recover energy via methane-rich biogas (Becker et al. 2017). Currently, only two full-
scale AnMBRs in the U.S. exist for FW treatment. One of these facilities, operated by Divert,
Inc. (Compton, CA), collects food waste from Ralph’s and Food 4 Less grocery stores, and
produces approximately 120 kWh ton
-1
of food waste processed, offsetting 25-30% of the
energy demands of the distribution center. FW that is unable to be sold or donated is
collected from grocery stores in Los Angeles and transported to the distribution center, where
it is mixed with wastewater from an on-site creamery (Figure 1). The wastewater produced
at the creamery has a high chemical oxygen demand (COD), and if discharged directly to the
sewer system, results in significant quality surcharge fees by the public sewerage district.
Therefore, the advantage of an AnMBR at the facility is twofold, compliance with organic
waste landfill diversion regulations is achieved and fees from creamery wastewater discharge
are minimized.

5

Figure 1. Process flow of full-scale AnMBR plant treating food waste from Ralph’s and
Food4Less.

6

Growing interest in waste-to-energy technologies necessitates fundamental and applied
research on anaerobic treatment of FW. Although co-digestion with sewage sludge at
wastewater treatment plants is an attractive management strategy, decentralized anaerobic
treatment can utilize existing transportation routes in certain applications to reduce overall
environmental impacts. AnMBRs are well-suited for this application given their enhanced
energy recovery, improved effluent quality due to membrane separation, reduced sludge
production, and smaller footprint. Although substantial research has been done to-date on
anaerobic digestion, AnMBRs are still an emerging technology with few full-scale installations
worldwide. In particular, very few studies have evaluated AnMBRs for the sole treatment of
FW. Given the nature of FW (i.e., highly complex organics with temporal variations in strength
and composition), we hypothesized that two-phase (acid/gas) AnMBRs equipped with
ceramic membranes provides performance benefits relative to anaerobic digesters or single-
phase AnMBRs. To our knowledge, this configuration has yet to be evaluated. Additional
fundamental research is also needed to better understand the highly complex microbial
communities driving anaerobic systems. Relatively few studies thus far have evaluated the
microbial community in AnMBRs (Harb et al. 2015, Ma et al. 2013, Smith et al. 2013, 2015, Yu
et al. 2012, Yue et al. 2015) and we are just beginning to apply advanced molecular
microbiological methods such as high-throughput sequencing (e.g., Illumina) to these
systems. A significant opportunity exists to apply these tools and provide a fundamental and
mechanistic understanding linking microbial community structure and function in anaerobic
systems.

7

1.2 Overview of Dissertation
The following chapters investigate important knowledge gaps in anaerobic treatment of FW.
Chapter 2 (Amha et al. 2017) provides a critical review on recent literature studying microbial
inhibition and the use of advanced molecular methods in anaerobic digestion. Chapter 3
(Amha et al. 2017) describes experimental work evaluating FW and fats, oils, and grease (FOG)
co-digestion and is the first study that applied microbial activity analyses to evaluate
community response to co-digestion of these waste streams and during FOG inhibition.
Chapter 4 investigates performance and microbial community structure and activity in single-
phase versus two-phase AnMBRs treating FW. Finally, Chapter 5 evaluates these competing
AnMBR configurations during co-digestion of FW and FOG. Notably, this dissertation research
is in partnership with Divert, Inc., who operate two full-scale AnMBRs in the U.S.
1.3 Literature Cited
Amha, Y.M., Sinha, P., Lagman, J., Gregori, M. and Smith, A.L. (2017) Elucidating microbial
community adaptation to anaerobic co-digestion of fats, oils, and grease and food waste. Water
research.
Becker, A.M., Yu, K., Stadler, L.B. and Smith, A.L. (2017) Co-management of domestic wastewater
and food waste: A life cycle comparison of alternative food waste diversion strategies. Bioresource
Technology 223, 131-140.
Harb, M., Xiong, Y., Guest, J., Amy, G. and Hong, P.-Y. (2015) Differences in microbial communities
and performance between suspended and attached growth anaerobic membrane bioreactors
treating synthetic municipal wastewater. Environmental Science: Water Research & Technology 1(6),
800-813.
Kong, D., Shan, J., Iacoboni, M. and Maguin, S.R. (2012) Evaluating greenhouse gas impacts of
organic waste management options using life cycle assessment. Waste management & research,
0734242X12440479.
Ma, J., Wang, Z., Zou, X., Feng, J. and Wu, Z. (2013) Microbial communities in an anaerobic dynamic
membrane bioreactor (AnDMBR) for municipal wastewater treatment: Comparison of bulk sludge
and cake layer. Process Biochemistry 48(3), 510-516.
Matteson, G.C. and Jenkins, B. (2007) Food and processing residues in California: Resource
assessment and potential for power generation. Bioresource Technology 98(16), 3098-3105.
Smith, A.L., Skerlos, S.J. and Raskin, L. (2013) Psychrophilic anaerobic membrane bioreactor
treatment of domestic wastewater. Water research 47(4), 1655-1665.
Smith, A.L., Skerlos, S.J. and Raskin, L. (2015) Membrane biofilm development improves COD
removal in anaerobic membrane bioreactor wastewater treatment. Microbial biotechnology 8(5),
883-894.

8

Yu, Z., Wen, X., Xu, M. and Huang, X. (2012) Characteristics of extracellular polymeric substances and
bacterial communities in an anaerobic membrane bioreactor coupled with online ultrasound
equipment. Bioresource Technology 117, 333-340.
Yue, X., Koh, Y.K.K. and Ng, H.Y. (2015) Effects of dissolved organic matters (DOMs) on membrane
fouling in anaerobic ceramic membrane bioreactors (AnCMBRs) treating domestic wastewater.
Water research 86, 96-107.

9

CHAPTER 2

2. Inhibition of Anaerobic Digestion Processes: Applications of
Molecular Tools

2.1 Abstract
Inhibition of anaerobic digestion (AD) due to perturbation caused by substrate composition
and/or operating conditions can significantly reduce performance. Such perturbations could
be limited by elucidating microbial community response to inhibitors and devising strategies
to increase community resilience. To this end, advanced molecular methods are increasingly
being applied to study the AD microbiome, a diverse community of microbial populations with
complex interactions. This literature review of AD inhibition studies indicates that inhibitory
concentrations are highly variable, likely stemming from differences in community structure
or activity of inoculum and previous exposure to inhibitors. More recent molecular methods
such as ‘omics’ tools, substrate mapping, and real-time sequencing are helping to unravel the
complexity of AD inhibition by elucidating physiological and ecological significance of key
microbial populations. The AD community must strive towards developing predictive abilities
to avoid system failure (e.g., real-time tracking of an indicator species) to improve resilience
of AD systems.
2.2 Introduction

10

Anaerobic digestion (AD) is a waste management biotechnology that employs a diverse
consortium of microorganisms to convert organics into methane-rich biogas. AD reduces
organic waste landfilling and recovers energy via cogeneration of produced biogas. Given
available organic waste feedstocks worldwide, biogas recovered from AD has the potential to
provide a quarter of the world’s natural gas demand and 6% of primary energy demand (Guo
et al. 2015b). Despite the potential for AD to significantly contribute to our energy portfolio,
biogas remains an underutilized resource with only 47–95 billion kWh of electricity generated
from biogas in 2012, contributing just 0.2-0.4% of global electricity production (De Vrieze and
Verstraete 2016, Enerdata 2015). Improving AD implementation requires that we overcome
existing technological challenges that limit its widespread adoption (e.g., slow startup, low
energy recovery, and inhibition).
AD is an engineered ecosystem where organic waste degradation takes place via a complex
cascade of microbially-driven reactions including hydrolysis, fermentation (i.e., acidogenesis
and acetogenesis), and methanogenesis (Figure 2). Energy recovery thus relies on functional
activity of a wide range of Bacteria and Archaea, making it necessary to configure and operate
AD systems such that conditions are conducive for diverse microbial populations. High
complexity within the AD microbiome makes the process vulnerable to upset due to inhibition
via accumulation of long chain fatty acids (LCFA), volatile fatty acid (VFA), free ammonia, and
other compounds or unfavorable operating conditions, such as temperature and pH (Chen et
al. 2014a). Although AD can recover energy from a wide range of organics (e.g., wastewater
sludges; animal manure; food waste; and fats, oils, and grease (FOG)), feedstock variability
leads to operational uncertainty which can reduce energy recovery or ultimately result in
system failure. Therefore, significant effort has been placed on establishing inhibitory levels

11

of specific compounds in AD, as summarized in Tables 1-4. However, inhibitory levels are
strongly influenced by microbial community structure and activity. For example, functionally
redundant populations within the AD microbiome can prevent upset by limiting accumulation
of inhibitory intermediates. Thus, inoculum selection and temporal adaptation to inhibitors
can prevent process failure (Silva et al. 2014, Silvestre et al. 2011), but we require a better
understanding of the AD microbiome to accurately predict and prevent inhibition.
Recent advances in molecular methods have made it possible to study the structure, function,
and interaction of increasingly complex microbial communities. Nucleic acid–based molecular
methods have revolutionized environmental biotechnology research by making it possible to
study microbial communities without culturing (Kumaraswamy et al. 2014). DNA
fingerprinting methods such as denaturing gradient gel electrophoresis (DGGE) have been
employed widely to study AD systems. However, DNA fingerprinting and other early
molecular methods suffer from limited coverage and depth. This limitation conflicts with the
high diversity and importance of low abundance populations within the AD microbiome,
making it difficult to accurately monitor the activity of rare, but functionally important
microorganisms. Thus, sequencing-based approaches using pyrosequencing, Illumina
sequencing, or other high-throughput platforms are now widely used to study AD systems.
We can also couple high-throughput sequencing with isotope tracing (i.e., DNA- and RNA-
stable isotope probing (SIP)) to link substrate uptake with specific microbial populations
(Werner et al. 2014). Using advanced molecular tools, we now know that syntrophic acetate
oxidation (SAO), a thermodynamically unfavorable reaction (De Vrieze and Verstraete 2016),
is possible by more diverse microbial populations than originally assumed (Lee et al. 2015,
Treu et al. 2016, Werner et al. 2014). However, we continue to struggle with connecting

12

community structure and function in AD systems (Carballa et al. 2015, De Vrieze and
Verstraete 2016) and have not yet unraveled the ‘black-box’ microbial ecology of AD (Nobu
et al. 2015).

Figure 2. Metabolisms and key microbial populations in anaerobic processes.

The AD research community has made considerable progress in understanding inhibitory
impacts and mitigation strategies using advanced molecular methods. However, this work has
not been critically reviewed to summarize recent progress and highlight gaps in our
understanding of inhibition within the AD microbiome. The objective of this manuscript is to

13

review comprehensively how molecular methods have improved our understanding of AD
inhibition and highlight emerging molecular tools that could be used to prevent inhibition.
2.3 Inhibition of metabolic pathways in AD
2.3.1. Hydrolysis and fermentation
Hydrolysis, the first step in AD, is performed by hydrolytic fermentative bacteria that degrade
complex polymers to oligomers and monomers using extracellular enzymes (e.g., cellulases,
proteases, and lipases) (Keating 2015). Complex insoluble polymers in many AD feedstocks
can result in hydrolysis being rate limiting (Pavlostathis and Giraldo-Gomez 1991). Hydrolytic
bacteria in AD are found within five phyla: Firmicutes, Bacteroidetes, Fibrobacter,
Spirochaetes, and Thermotogae (Azman et al. 2015). Firmicutes and Bacteroidetes are
typically the most abundant taxa of hydrolytic bacteria in AD, although relative abundance of
these taxa is often dictated by inoculum and reactor type, as reviewed by Azman et al. (2015).
Following hydrolysis, fermentative bacteria degrade oligomers and monomers into
intermediates such as volatile fatty acids (VFAs) and alcohols (acidogenesis).
Hydrolytic bacteria are inhibited by elevated levels of VFAs, LCFAs, hydrogen partial pressure,
and humic acids (Azman et al. 2017, Azman et al. 2015, Cazier et al. 2015)(Table 1-4).
Inhibition occurs via activity loss, reversible reduction of hydrolases (e.g., when inhibitors bind
to enzyme active sites or substrate-enzyme complexes), or irreversible impacts resulting from
changes in enzyme chemical structure (Azman et al. 2015). The latter is challenging to
mediate, requiring removal of the inhibitor from the system. Siegert and Banks (2005) found
that VFA concentrations of 2 g L
-1
resulted in 75% inhibition in cellulose hydrolysis. Similarly,
another study reported inhibition of cellulose hydrolysis when VFAs exceeded 1.8 g L
-1

(Romsaiyud et al. 2009). High hydrogen partial pressure was also found to be inhibitory to

14

hydrolytic bacteria, reducing degradation of wheat straw with no accumulation of
metabolites from acidogenic bacteria (Cazier et al. 2015). In another study, high humic acid
concentrations resulted in a decrease of hydrolysis by 40% with this decline attributable to a
decrease in relative abundance of hydrolytic/fermentative bacterial populations including
Clostridiales, Bacteroidales, and Anaerolineales (Azman et al. 2017). Inhibition due to LCFA
has also been reported, with inhibition of hydrolytic bacteria occurring at 2.6 - 9.4 kg COD m
-
3
and acidogenic bacteria at a similar concentration range, 2.1-7.9 kg COD m
-3
(Nobu et al.
2015).
2.3.2 Syntrophy and methanogenesis
Syntrophic bacteria in AD convert fatty acids produced by acidogenic bacteria into acetate,
hydrogen, and carbon dioxide. A total of 23 different genera have been identified to date with
the ability to function as syntrophic bacteria, with most syntrophic genera found within
Firmicutes (Schuchmann and Müller 2014). Of these syntrophs, two families,
Syntrophomonadaceae and Syntrophaceae, and 14 species within these families are able to
degrade LCFAs (Baserba et al. 2012, Sousa et al. 2009). Thermodynamically, LCFA
fermentation is both endothermic and nonspontaneous (Chen et al. 2014a). Therefore, LCFA
degradation to acetate and hydrogen is made possible through β-oxidation and syntrophy
with hydrogenotrophic methanogens or acetoclastic methanogens (Chen et al. 2014a, Treu
et al. 2016, Ziels et al. 2017).
Methanogens are commonly categorized as hydrogenotrophic or acetoclastic based on their
electron donor. Many hydrogenotrophic methanogens can metabolize C1 compounds such
as formate and methanol in addition to hydrogen (Demirel and Scherer 2008). Methanosaeta,
obligate acetoclastic methanogens, have been shown to dominate in mesophilic AD (Guo et

15

al. 2015a) and outcompete Methanosarcina at low acetate concentration due to their higher
substrate affinity (Conklin et al. 2006). Methanosarcina, unlike Methanosaeta, are
mixotrophic methanogens that can metabolize acetate, hydrogen, and C1 compounds
(Mladenovska and Ahring 1997). Some studies have suggested that hydrogenotrophic
methanogens are dominant at thermophilic temperatures (Pap et al. 2015) and during
inhibition due to high VFA and ammonia (de Jonge et al. 2017).
Common inhibitors of syntrophs and methanogens are VFAs, ammonia, and LCFAs (Chen et
al. 2008) (Table 1-3). Relative to methanogens, syntrophic bacteria better tolerate high VFA
and ammonia (Li et al. 2016). In fact, high VFA could initially promote the growth of syntrophic
bacteria (Li et al. 2015, 2016). However, inhibition of hydrogenotrophic methanogens, their
syntrophic partners, at elevated VFA concentrations destabilizes syntrophy, eventually
leading to increased hydrogen partial pressure and a decrease in thermodynamic favorability
of the reaction (Li et al. 2015). A study that investigated the effect of VFA accumulation due
to high organic load in a batch reactor treating kitchen waste found that VFAs at 5.8 - 6.9 g L
-
1
were completely inhibitory to methanogens (Xu et al. 2014). In contrast, high VFAs were
shown to be non-inhibitory to methanogens in a full-scale reactor treating cow manure and
food waste, even when propionate concentrations reached 8.7 g L
-1
(Franke-Whittle et al.
2014). Regarding ammonia inhibition, studies have reported that acetoclastic methanogens
are inhibited more severely than hydrogenotrophic methanogens (Hagen et al. 2017, Niu et
al. 2013, Sun et al. 2016, Werner et al. 2014). For example, in one study methanogenesis was
completely inhibited when total ammonia nitrogen (TAN) exceeded 9 g N L
-1
with the critical
threshold for performance decline reported to be 7 g N L
-1
(Sun et al. 2016). Sulfate reducing
bacteria (SRB) can compete with methanogens for substrates while also producing potentially

16

toxic sulfides. Hydrogen sulfide has been shown to diffuse through cell membranes and form
disulfide cross-links between polypeptide chains, thereby denaturing proteins and affecting
cellular function (Chen et al. 2014a, Tursman and Cork 1989).
Syntrophic bacteria have been found to be particularly sensitive to LCFA, with syntrophic
acetogens decreasing in abundance at high LCFA concentrations (Nobu et al. 2015). Ma et al.
(2015) proposed that inhibition resulted from attachment of LCFA on cell surfaces limiting
mass transfer and substrate access. Although fatty acid-based inhibition is not well
understood, researchers have identified several likely mechanisms of inhibition: disruption of
the electron transport chain and oxidative phosphorylation, interference with cellular energy
production, direct lysis of bacterial cells, and decreased cell permeability (Desbois and Smith
2010, Pereira et al. 2005). However, some studies have suggested that LCFA inhibition is
reversible (Kougias et al. 2016, Pereira et al. 2005, Ziels et al. 2016). For example, a study
involving cattle manure digestion with the addition of sodium oleate to simulate high LCFAs
reported that the perturbation observed was reversible (Kougias et al. 2016). Inhibition was
also minimized when inoculating with sludge previously acclimated to high LCFAs (Kougias et
al. 2016), suggesting microbial community adaptation can lead to resilience. Notably,
resilience was found to be an important factor in maintaining a syntrophic population
exposed to disturbances in a full-scale reactor, where syntrophs rebounded following stress
conditions (Werner et al. 2011). Another study also suggested reversibility of LCFA inhibition
due to an increase in relative abundance of syntrophic β-oxidizing bacteria, primarily
Syntrophomonas (Ziels et al. 2016). Further, the same study reported that the abundance and
composition of methanogens was unaffected by addition of 100 - 1570 mg oleic acid g VS
-1
,
where 70% of the methanogens were hydrogenotrophic (Methanomicrobiales) and 30% were

17

acetoclastic (Methanosaeta). Similarly, another study reported that increase of LCFA affected
hydrolytic bacteria more so than methanogens (Nobu et al. 2015). LCFA inhibition is further
complicated by observations of varying inhibition due to type of LCFA, saturated versus
unsaturated LCFAs, and methanogen taxonomy. Sousa et al. (2013) investigated the impact
of high concentrations of oleate (unsaturated LCFA) and palmitate (saturated LCFA) on pure
cultures of acetoclastic methanogens and hydrogenotrophic methanogens, finding that
saturated LCFA had greater inhibitory effects than unsaturated LCFA on methanogens and
that the inhibition mechanism was through damaged membrane integrity. Further,
Methanobacterium formicicum (hydrogenotrophic) was more resilient than Methanospirillum
hungatei (hydrogenotrophic) for both saturated and unsaturated LCFA whereas
Methanosarcina mazei (acetoclastic) and Methanosaeta concilii (acetoclastic) were
completely inhibited by oleate.
Inhibition of acetoclastic methanogens can also drive increases in syntrophic acetate oxidizing
bacteria (SAOB), which oxidize acetate into hydrogen and carbon dioxide, that in turn can
increase hydrogenotrophic methanogenic activity (Carballa et al. 2015, Gao et al. 2015). SAOB
remain poorly studied, with only a few cultured populations to date: Syntrophaceticus spp.,
Thermacetogenium phaeum, Thermotoga lettingae, Tepidanaerobacter acetatoxydans, and
Clostridium ultunense (Müller et al. 2013). Some studies suggest that SAO could be conducted
by more diverse populations than previously thought (Lee et al. 2015, Werner et al. 2014).
For example, a recent study identified cluster II Spirochaetes and members of Clostridia as
potential SAOB using
13
C-labeled acetate (Mosbæk et al. 2016). Ammonia induced
perturbation has also been shown to promote the SAO pathway, with a shift from acetoclastic
to hydrogenotrophic methanogenesis (Werner et al. 2014).

18

The impact of inhibitors may be heavily dependent on microbial community acclimation, with
long-term exposure often leading to stable community function in the presence of inhibitors
(Silva et al. 2014, Silvestre et al. 2011). Therefore, it is unsurprising that a wide range of
inhibitor concentrations have been reported to date (Table 1-4). This variability likely stems
from differences in inoculum (Baserba et al. 2012) or prior exposure to inhibitors. Differences
in microbial community structure make it particularly challenging to compare inhibition in AD.
The majority of studies to date have only applied DNA-based methods to study microbial
community structure during AD inhibition (de Jonge et al. 2017, Li et al. 2015, 2016, Nobu et
al. 2015, Ziels et al. 2016). However, DNA-based methods are relatively insensitive,
particularly in anaerobic communities with low biomass yields. Inhibitory conditions
negatively impact DNA replication rates and DNA may persist in the environment after a cell
ceases activity. De Vrieze et al. (2016) demonstrated that presence does not always correlate
with activity by studying community response to salt perturbation. The authors also used
RNA-based sequencing and found that it was a more sensitive tool than DNA-based
sequencing to quantify microbial community response to high salt concentrations. For
example, Methanosaeta remained the most abundant methanogen based on DNA-
sequencing, but was strongly inhibited by high salt concentration according to RNA-based
sequencing, corroborating performance observations. Similar observations have been
reported during high FOG addition (Amha et al. 2017), providing further evidence supporting
the need for RNA-based methods when evaluating AD inhibition. Recent developments in
molecular methods have resulted in a greater understanding of inhibition in AD, as discussed
in subsequent sections.

19

Table 1. Studies investigating volatile fatty acid (VFA) inhibition.
Reactor Temp. Feed Performance
impact
Microbial
communit
y analysis
tool
Effect on microbial
community
Refe
renc
e
Full-
scale
Mesophilic
and
thermophil
ic
Cow manure
and food
waste
No inhibition
detected even with
high VFA (e.g.,
propionate
concentrations of
8741 mg L
-1
)
ANAEROC
HIP
microarra
y and real-
time qPCR
Methanothermobacter
dominated in
thermophilic reactor
with high VFA
Methanosarcina
increased in
dominance with higher
levels of VFAs
(Fra
nke-
Whit
tle
et al.
2014
)
Batch Thermophil
ic
13
C and
12
C
labelled
acetate
High acetate fed
reactor (8.2 g L
-1
)
showed linear
acetate removal in
the first 120 h (83%
of the amended U-
13
C)

Protein-
SIP and
metageno
me
Methanosarcina,
Methanoculleus, and
five subspecies of
Clostridia (potential
SAOBs) involved in
recovery after
inhibition
Identified Clostridia
groups contained the
fthfs gene
(Mos
bæk
et al.
2016
)
Batch Thermophil
ic
Sodium
acetate
N/A Isotope
labeled
substrate
assays,
protein-
SIP,
metageno
mics
Identified Clostridia,
Methanosarcina, and
Methanoculleus as
being part of the
recovery from high
acetate conditions.
(Mul
at et
al.
2014
)
Batch Mesophilic Cellulose Critical acetate
concentration was
25 mmol L
-1

pH should be
maintained at 7 to
enhance cellulose
hydrolysis rate
N/A N/A (Ro
msai
yud
et al.
2009
)
CSTR Thermophil
ic
Sugar beet
tailings
Lower methane
yield, delayed
methane
production rate,
and propionic acid
accumulation
Pyros
eque
ncing
of
16S
rRNA
gene
Methanogens and
syntrophic bacteria
became less
abundant
Acetanaerobacteri
um and
Ruminococcus
became more
abundant
(Tian
et al.
2015
)
Batch Thermophil
ic
Kitchen
waste
Initial inhibitory
concentration of
acetate was
between 1.5 g L
-1
-
2.5 g L
-1

Methanogenic
activity was
inhibited
PCR-
DGGE
Accumulation of
acetic acid
inhibited
acetoclastic
methanogens
more than
hydrogenotrophic
methanogens
(Xu
et al.
2014
)

20

completely at VFA
concentration of
5.8 g L
-1
- 6.9 g L
-1

Abbreviations: VFA, volatile fatty acids; fthfs, formyltetrahydrofolate synthetase-encoding gene; CSTR, continuous stirred
tank reactor; N/A, not available; PCR-DGGE, polymerase chain reaction - denaturing gradient gel electrophoresis

Table 2. Studies investigating ammonia inhibition.
React
or
Temp. Feed Performance Microbial
community
analysis tool
Effect on
microbial
community
Referenc
e
CSTR Mesoph
ilic
Dewate
red
sludge
VFA reduction
changed from 32% to
21% and biogas
decreased from 12 L
d
-1
to 10 L d
-1
when
ammonia
concentrations
reached 5000-6000
mg N L
-1

Pyrosequencing of
16S rRNA gene and
qPCR targeting
Methanosarcinace
ae
Methanosarcina
dominated
archaeal
population
(resistant to
ammonia stress)
(Dai et
al. 2016)
UASB Mesoph
ilic
Basal
anaerob
ic (BAN)
medium
5 g NH4-N L
-1
led to
25% lower methane
yield
FISH and confocal
laser
Bioaugmentation
of SAO co-culture
was not possible
because
methanogens
used in the co-
culture
(Methanoculleus
spp.) had slow
growth rate
(Fotidis
et al.
2013)
CSTR Thermo
philic
Cattle
manure
Methane production
decreased by 37.7%
after addition of
cattle manure and
accumulation of
ammonia
Ion Torrent PGM
sequencing
targeting 16S rRNA
gene
Inhibition of
methanogens
except
Methanoculleus
(De
Francisci
et al.
2015)
Pilot-
scale
CSTR
Mesoph
ilic
Kitchen
waste
TAN concentration of
4000 mg L
-1
led to
severe inhibition,
acclimated reactor
performed well at
TAN 4293 mg L
-1

MiSeq targeting
16S rRNA gene
Ammonia stress
led to increase in
relative
abundance of
Firmicutes and
hydrogenotrophic
methanogens,
and decrease of
acetolactic
methanogens
(Gao et
al. 2015)
Full-
scale
Thermo
philic
Food
waste
FAN at 367 mg L
-
1
did not inhibit
performance,
but altered
microbial
community
dynamics
Metaproteomics
and metagenomics

Acetoclastic
methanogens
showed low
abundance yet
some metabolic
activity
(Hagen
et al.
2017)

21

Dominate acetate
removal was
through SAOB
Discovered novel
uncultured SAOB
that can also
degrade LCFA
Batch Thermo
philic
Sodium
acetate
Reduced initial
methane production
in the uninhibited
control reactors early
in the experiment,
however, methane
production increased
in the high ammonia
reactors later in the
experiment and
matched uninhibited
control reactors by
day 25.
Isotope labeled
substrate assays
and DNA-SIP with
analysis of 16S
rRNA gene
pyrotags
Methanosarcina
performed
acetoclastic
methanogenesis
at free ammonia
nitrogen
concentrations of
up to 916 mg L
-1

(Hao et
al. 2015)
CSTR Mesoph
ilic
Sodium
propion
ate and
nutrient
medium
TAN at 3.0 g L
-1

inhibited propionate
degradation and
methane recovery
rate dropped from
dropped from
82.91% to 28.09%
FISH Low abundance of
methanogens
under ammonia
stress
(Li et al.
2017b)
CSTR Mesoph
ilic
Thin
stillage
Performance
threshold for
ammonia
concentration was 1
g NH3 L
-1

qPCR targeting
methanogenic
populations, SAOB
and fthfs OTUs, T-
RFLP, and cloning
Methanoculleus
increased and
acetogenic
community
decreased at high
ammonia
(Moeste
dt et al.
2016)
CSTR Thermo
philic
Chicken
manure
Biogas production
almost ceased at high
concentration of TAN
(8000 mg L
-1
) and
VFA (25,000 mg N L
-1
)
16S rRNA gene
cloning and
sequencing
Hydrogenotrophic
methanogens
dominated during
inhibition phase
while acetoclastic
methanogens
were inhibited
(Niu et
al. 2013)
CSTR Mesoph
ilic
Chicken
manure
and
maize
silage
TAN at 7 g N L
-1

critical threshold for
performance
Isotope tracer Methanogens
completely
inhibited at TAN >
9 g N L
-1

(Sun et
al. 2016)
CSTR Mesoph
ilic
Thermal
ly
hydroly
zed
waste
activate
d sludge
Methane production
improved by 54%
when ammonia
concentration
decreased from 630
to 92 mg L
-1

qPCR targeting
methanogens
Six-fold increase
of
Methanosarcinac
eae and doubling
of bacterial
density improved
VFA, protein, and
carbohydrate
removal
(Tao et
al. 2017)

22

Seque
ncing
batch
react
ors
Mesoph
ilic
Swine
waste
TAN at 4.4 g N L
-1
and
FAN at 0.08 g N L
-1

decreased biogas
production rate and
VFA accumulation
occurred
Isotope tracer,
shotgun
sequencing,
cloning-sanger
sequencing, DNA-
SIP, and FISH-
NanoSIMS
Community shift
from acetoclastic
methanogenesis
to SAO
Decrease in
community
evenness
associated with
ammonia-induced
stress
(Werner
et al.
2014)
Inter
mitte
nt
CSTR
Mesoph
ilic
Food
waste
leachat
e
FAN at 700 mg N L
-1

resulted in significant
inhibition
Pyrosequencing of
16S rRNA gene
FAN inhibited
Methanosarcina
and
Methanosaeta
(Yun et
al. 2016)
Abbreviations: UASB, upflow anaerobic sludge blanket; FAN, free ammonia nitrogen; FISH, fluorescent in situ
hybridization; VFA, volatile fatty acids; fthfs, formyltetrahydrofolate synthetase-encoding gene;

Table 3. Studies investigating long chain fatty acid (LCFA) inhibition.
Reactor Temp
.
Feed Performance Microbial
community
analysis tool
Effect on microbial
community
Referenc
e
Batch Ther
moph
ilic
FOG
and
food
waste
>50% FOG volatile
solid loading
addition resulted
in more than 90%
decline in biogas
production
Illumina
MiSeq
targeting
16S rRNA
Increased relative activity
of syntrophs (mostly
Syntrophomonas),
prevented LCFA
accumulation in 30% FOG
(v/v) addition
>50% FOG volatile solid
loading addition was
inhibitory to syntrophs and
methanogens
(Amha et
al. 2017)
CSTR Ther
moph
ilic
Oleate Up to 2 g oleate L
−1

day
−1
did not show
inhibition and
resulted in
methane increase
PCR-DGGE Firmicutes increased
abundance
Decrease in bacterial
diversity
Methanosarcina
and Methanococcus were
the dominant
methanogens
(Baserba
et al.
2012)
CSTR Ther
moph
ilic
Cattle
manure
and
sodium
oleate
In non-acclimated
reactor LCFA
addition of 3 g L
-1

resulted in 95%
reduction in
methane yield
Inhibition was
reversible
Metagenomi
cs
Syntrophomonas and
Methanosarcina increased
in relative abundance
Microbes that responded
positively to LCFA pulse
could encode proteins
related to “chemotaxis”
and “flagellar assembly”
(Kougias
et al.
2016)
Batch Meso
philic
Lipid-
extracte
d algal
biomass
In the high lipid
concentration
digester,
inoculum/substrat
e ratios of <1
resulted in biogas
production decline
Illumina
MiSeq
targeting
16S rRNA
gene
Bacterial community was
affected more than
methanogens at high LCFA
Hydrolytic bacteria and
acetoclastic methanogens
dominated
(Nobu et
al. 2015)

23

Syntrophic acetogens were
sensitive to high LCFA
Various Meso
philic
Skim
milk
oleate
Long term
acclimation to
LCFA was found
necessary to
prevent inhibition
n/a Exposure to LCFA >100
days resulted in acclimated
microbial community and
limited inhibition
Bioaugmentation of
Syntrophomonas zehnderi
and Methanobacterium
formicicum showed no
performance improvement
(Silva et
al. 2014)
Batch Meso
philic
Oleate,
stearate
, and
palmitat
e
Methanogenic
activity decreased
by 50% with 0.3,
0.4 and 1 mM
oleate, stearate,
and palmitate with
M. hungatei
cultures, whereas
50% reduction in
methanogenic
activity was seen
at 1mM oleate and
>4m stearate or
palmitate for M.
formicicum culture
Live/Dead
BacLight
bacterial
viability kit,
cloning and
sequencing,
DGGE
Methanobacterium
formicicum was more
resilient than
Methanospirillum hungatei
for both saturated and
unsaturated LCFA

Methanosarcina mazei and
Methanosaeta concilii
were inhibited in oleate
culture and acetate
accumulated
(Sousa et
al. 2013)
CSTR Ther
moph
ilic
Cattle
manure
and
sodium
oleate
N/A Metatranscr
iptomics
Syntrophomonas
dominated at high LCFA
concentrations
Protective mechanisms
was suggested as
upregulation of genes
involved in peptidoglycan
and lipopolysaccharides
biosynthesis
(Treu et
al. 2016)
CSTR Meso
philic
Manure
and
oleate
Effluent acetate
reached 3000 mg L
-
1
in continuous fed
reactor
Illumina
MiSeq
targeting
16S rRNA
gene
The relative abundance of
Syntrophomonas increased
to ~15%
Methanosaeta and
Methanospirillum were the
dominant methanogens
(Ziels et
al. 2016)
CSTR Meso
philic
Oleic
acid
After acclimation
(on day 204), the
continuously fed
reactor was
inhibited at oleate
concentration of
875 mg L
-1
,
whereas the pulse-
fed reactor was
not inhibited at
1800 mg L
-1
oleate
concentration
Illumina
MiSeq
targeting
16S rRNA
gene and
qPCR
targeting
Syntrophom
onas, total
bacteria,
and total
archaea
LCFA feeding frequency
and OLR impacted
microbial community
composition and
biokinetics
Higher relative and
absolute 16S rRNA gene
concentration of
Syntrophomonas and
Methanosaeta in pulse
feed digester than
continuously fed digester
(Ziels et
al. 2017)
Abbreviations: PCR-DGGE, polymerase chain reaction - denaturing gradient gel electrophoresis;
qPCR, quantitative polymerase chain reaction

24

Table 4. Miscellaneous studies investigating inhibition.
Inhibit
ors
React
or
Temp
.
Feed Performance Microbial
community
analysis tool
Effect on microbial
community
Referen
ce
Humic
acid
Batch Meso
philic
Cellulos
e and
xylan
mixture
Biogas
production
decreased
approx. 33% at
inhibited
conditions
Illumina HiSeq
targeting 16S
rRNA gene
Inhibited the
hydrolysis efficiency
of the digestion by
40%
(Azman
et al.
2017)
H2 Batch Meso
philic
Wheat
straw
Substrate
degradation
decreased
without
accumulation of
metabolites
from acidogenic
bacteria
Illumina MiSeq
targeting 16S
rRNA gene
Hydrolytic activity
was impacted more
than acidogenic
bacteria
(Cazier
et al.
2015)
Humic
acid
Batch Meso
philic
HAL and
FAL
substan
ces
from
maize
and
cow
manure
Tributyrin
hydrolysis was
inhibited by
HAL from 0.5 to
5.0 g L
−1

T-RFLP
fingerprinting
targeting 16S
rRNA gene
Cellulose hydrolysis
was inhibited by 0.5
to 5.0 g L
−1
of HAL
and FAL
(Fernan
des et
al.
2015)
Methyl
Fluorid
e
Batch N/A Sodium
acetate
Reduced biogas
formation
observed with
increasing
concentrations
of CH3F in the
headspace.
Stable isotope
analysis of
biogas
Acetoclastic
methanogenesis
was progressively
inhibited in the
presence of CH3F at
low and middle
concentrations
At the highest
concentration of
CH3F (10% in the
headspace), both
acetoclastic and
hydrogenotrophic
methanogenesis
were inhibited
(Hao et
al.
2011)
Starvat
ion
CSTR Ther
moph
ilic
Cattle
manure
VFAs, H2S, and
ammonia
accumulated
during
starvation
phase
Illumina MiSeq
targeting 16S
rRNA gene
SAOB and
hydrogenotrophic
methanogens
increased after
inhibition phase
(de
Jonge et
al.
2017)
Dissolv
ed
lignin
CSTR Meso
philic
and
ther
moph
ilic
Lignin-
rich
indigen
ous
macrop
hyte
species
Performance
decreased by
approx. 40% in
pretreated
substrate
during
inhibition
Cloning and
sequencing of
16S rRNA gene
Hydrogenotrophic
methanogenic
pathway was a
limiting step for
alkaline treated
reactors, where
increase in partial
(Koyam
a et al.
2017)

25

pressure of H2
caused
accumulation of
VFAs
OLR CSTR Meso
philic
Food
waste
Organic load
increase to 6 g
VS L
-1
d
-1
led to
50% reduction
in methane
production
pyrosequencing
of 16S rRNA
gene
Acid producing
bacteria and
syntrophic fatty acid
oxidizers increased
in relative
abundance at high
OLR
Acetoclastic
methanogens
dominated
(Li et al.
2015)
OLR CSTR Meso
philic
Food
waste
OLR > 6 g VS L
-1
d
-1
resulted in
increase of FAN
(114 mg L
-1
) and
VFA (9443 mg L
-
1
at day 90)
Propionate
increased by
20-fold
pyrosequencing
of 16S rRNA
gene
Acidogenic bacteria
showed functional
redundancy and
adopted to high OLR
Acetogens such as
the genera
Syntrophomonas
and Treponema
increased with high
VFA
Hydrogenotrophic
methanogens were
inhibited
(Li et al.
2016)
Total
solids
CSTR Meso
philic
Sewage
sludge
TS increase
from 10% to
15% resulted in
biogas
production
decrease from
383 mL g VS
-1
to
316 mL g VS
-1

Illumina MiSeq
targeting 16S
rRNA gene
Acetoclastic
methanogens
decreased along
with the increased
TS
Hydrogenotrophic
methanogens
increased
Acidogenic and
acetogenic bacteria
of phylum
Firmicutes
decreased but
phylum
Bacteroidetes
increased in relative
abundance
(Liu et
al.
2016)
OLR Semi-
CSTR
Meso
philic
Spirulin
a
Increasing the
OLR from 0.5 to
1 g Spirulina L
-1

d
-1
led to
inhibition
Metatranscript
omics and
metagenomics
Hydrolysis was
mainly performed
by Bacteroides
while metagenomic
activity was
dominated by
Methanocalculus
(Nolla-
Ardèvol
et al.
2015)
2-BES Meso
cosm
Meso
philic
cow
dung
and
sludge
0.5 mmol L
-1

BES and 10
mmol LBES
-1
led
to 89% and
100% methane
RT-qPCR
targeting mcrA
Methanogenic
activity decreased
with the exposure
to the inhibitors
(Webst
er et al.
2016)

26

from
WWTP
production
reduction
Acetoclastic
methanogens were
more impacted than
hydrogenotrophic
methanogens
Abbreviations: OLR, organic loading rate; T-RFLP, terminal restriction fragment length polymorphism; FAN, free
ammonia nitrogen; 2-BES, 2-bromoethanesulfonate; VFA, volatile fatty acids; TS, total solids; fthfs,
formyltetrahydrofolate synthetase-encoding gene; WWTP, wastewater treatment plant

2.4 Targeted nucleic acid biomarkers
2.4.1 Universal target: 16S rRNA gene and metadata analysis
The 16S rRNA gene is the most widely used biomarker and provides detailed phylogenetic
information about Bacteria and Archaea from mixed microbial communities. Several
limitations of this approach should be noted. Because the number of copies of the 16S rRNA
gene per genome can vary from 1 to 15 and is often not known, differences in relative
abundances can often misrepresent the true abundances of different populations
(Klappenbach et al. 2000). Even within the same organism, the number of copies of rRNA
genes will differ at different stages of development and in the metabolic state at the point of
sampling, i.e. dormant, active or growing (Blazewicz et al. 2013, Sukenik et al. 2012). Relative
abundance information can also be misleading when the overall size of the community is
changing (Props et al. 2016), though quantitative sequencing methods are beginning to be
used (Smets et al. 2016). Despite these limitations, 16S rRNA gene sequencing remains one
of the most convenient and widely used methods to characterize a microbial community.
We aimed to draw on data from existing 16S rRNA gene sequence datasets to investigate how
microbial communities vary across AD over a range of fatty acid concentrations. We collected
16S rRNA gene sequence data and metadata available from four different publicly available
studies that included a total of 99 samples from bench- and full-scale anaerobic digesters
operated at mesophilic and thermophilic temperatures (Amha et al. 2017, Li et al. 2016, Liu

27

et al. 2016, Nobu et al. 2015). These studies were selected based on sequence data and
metadata availability and their comparison of LCFA and VFA inhibition.
Quantitative Insights into Microbial Ecology (QIIME) (Caporaso et al. 2010) was used for the
pre-processing and downstream analyses after the samples were pooled together. Closed-
reference operational taxonomic unit (OTU) calling was performed against Greengenes
(DeSantis et al. 2006) (version 13.8) as a reference database. OTUs were assigned using QIIME
implementation of the UCLUST_ref (Edgar 2010) algorithm for clustering, with a threshold of
97% similarity. Beta-diversity was calculated after adaptive rarefaction (Henschel et al. 2015)
using a Bray Curtis matrix and hierarchical clustering. Samples were pooled based on
categorical levels of LCFA and VFA (Table S1).
In the majority of the compiled samples, Proteobacteria (particularly Gammaproteobacteria
and Deltaproteobacteria), methanogens (Methanobacteria and Methanomicrobia), and
Firmicutes (Clostridia) were present at high relative abundance irrespective of study and LCFA
or VFA concentration (i.e., level of inhibition). The LCFA analysis showed that Synergistia,
Clostridia, and Bacilli were less impacted by high LCFA concentrations relative to other groups
(Figure 3A). In fact, the relative abundance of Bacilli increased with increasing LCFA
concentrations. Thermotogae were highly impacted by high LCFA concentrations and showed
negligible relative abundance in most of the samples with high LCFA concentrations. In
contrast, high VFA concentrations impacted Synergistia, Clostridia, and Bacilli in most of the
samples (Figure 3B). Although there were a few samples that showed high relative abundance
of Clostridia and Bacilli under high VFA concentration, these groups were absent in most
samples at high VFA concentration. Similar to the LCFA analysis, Thermotogae decreased with
increasing VFA concentration, whereas, Methanomicrobia increased in relative abundance

28

with increasing VFA concentration in the majority of the samples. Despite some visual trends
in key populations across LCFA and VFA concentrations, we did not find any statistical
correlation between changes in taxa relative abundance and LCFA or VFA concentration.
To address biases that arise when comparing data from different studies and evaluate the
most shared microbes across the studies a core microbiome approach(Huse et al. 2012) was
utilized. Our analysis was limited by sample size given that sequence data and/or metadata is
not always available in public datasets or made available upon request. It is also important to
note that sequencing technology and the region of the 16S rRNA gene targeted can vary
between studies and introduce biases. The 16S rRNA gene is approximately 1500 bases long
but most next generation sequencing platforms only sequence a small section (e.g., 250 bp
for Illumina MiSeq). Therefore, primers target variable regions of the 16S rRNA gene (V1-V9).
In our analysis, the included sequence datasets targeted different variable regions within V2-
V5 of the 16S rRNA gene. The core microbiome was computed separately for LCFA and VFA
samples using QIIME to identify the OTUs present in at least 60% of the samples.

29

Figure 3. (A) Microbial community structure by class over increasing VFA concentration (left
to right). The y-axis represents relative abundance of OTUS that are 0.1% or greater in the
community. Stacked bars within each class (same color) represent orders. (B) Microbial
community structure by class over increasing LCFA concentration (left to right). The y-axis
represents relative abundance of OTUS that are 0.1% or greater in the community. Stacked
bars within each class (same color) represent orders.
A.
B.

30

Figure 4. (A) Core microbiome by class over increasing LCFA concentration (left to right). The
y-axis represents abundance relative to other populations within the core microbiome and
is therefore independent of populations not represented in the figure legend. Stacked bars
within each class (same color) represent orders (B) Core microbiome by class over increasing
VFA concentration. Table S1 defines the concentration ranges for LCFAs and VFAs on the x-
axis of all graphs.
A.
B.

31

Across the range of LCFA concentrations, four classes were identified in the core microbiome:
Methanobacteria, Clostridia, Bacteroidia, and Synergistia (Figure 4A). The relative abundance
of Bacteroidia and Clostridia increased with increasing LCFA concentration, whereas
Methanobacteria decreased. This is consistent with the expectation that microorganisms with
fatty acid metabolisms would dominate at higher fatty acid concentrations. Synergistia were
the most resistant to increasing LCFA and remained least affected relative to other
populations. The core microbiome for VFA samples included Methanomicrobia,
Methanobacteria, Clostridia, Bacteroidia, and Synergistia. Similar to the LCFA analysis,
Bacteroidia and Clostridia increased in relative abundance at higher VFA concentrations while
methanogen relative abundance (Methanobacteria and Methanomicrobia) decreased.
Bacteroidia were the only population from the core microbiome still present at inhibitory VFA
concentrations (Figure 4B). These results suggest that key phylogenetic groups involved in AD
respond similarly to fatty acid inhibition across different digesters and operational conditions.
Our meta-analysis effort was significantly confounded by two factors: data heterogeneity and
limited data availability. Data heterogeneity largely stems from DNA/RNA extraction protocol,
PCR primer selection, sequencing platform, sequencing depth, etc. Efforts should be made to:
(i) standardize data sharing by depositing raw data in publicly available databases with
supporting information on sample demultiplexing, primers used, etc.; and (ii) measure and
provide metadata describing environmental conditions such as pH, temperature, LCFA, VFA,
ammonia, etc. in standard quantitative units. Better availability of high-quality sequencing
data alongside metadata would help the AD community construct a theoretical framework
describing microbial community dynamics in the presence of inhibitors.

32

2.4.2 Functional genes
Another drawback of the 16S rRNA gene is that it cannot be tied to a specific metabolic
function, and therefore, it can be difficult to infer the contribution of a given population to
changes in the biochemical environment (Blazewicz et al. 2013). However, many functional
genes related to AD have been characterized and can be used to target specific populations.
For hydrogen producing fermentative bacteria, the use of genes encoding the large subunit
of Fe–Fe-hydrogenase (hydA) have been used as biomarkers (Xing et al. 2008, Ziganshin et al.
2016). To study populations that degrade aromatic compounds, which is important when
considering AD of petrochemical industry waste and phenol containing substrates, functional
genes for the benzoylcoenzyme A (benzoyl-CoA) degradation pathway have been designed as
biomarkers (Levén et al. 2012).
Although syntrophic bacteria have diverse metabolisms, the assimilation of CO 2 into biomass
and conservation of energy using the acetyl-CoA pathway has made it possible to use the
formyltetrahydrofolate synthetase-encoding gene (fthfs), a key enzyme for the acetyl-CoA
pathway, also known as the Wood-Ljungdahl pathway as a biomarker (Müller et al. 2013).
SAOB oxidize acetate to carbon dioxide and hydrogen using the reverse Wood-Ljungdahl
pathway, and various studies have used the fthfs gene as a biomarker to identify and quantify
SAOB in AD (Mosbæk et al. 2016, Müller et al. 2013) although it has also been noted that
some non-acetogenic bacteria may also have this gene (Lovell and Leaphart 2005, Mosbæk et
al. 2016).
Methanogens are often targeted in AD as indicators of performance because of their
presence and activity has been shown to correlate with biogas production measurements
(Morris et al. 2014, Webster et al. 2016). The most common functional gene used to study

33

methanogens encodes for the α-subunit of the methyl coenzyme-M reductase (mcrA gene),
and has been suggested as the most important biomarker in AD systems (De Vrieze and
Verstraete 2016). Significant positive correlations between methane production rate and
mcrA gene copy numbers (Morris et al. 2014) and transcripts (Webster et al. 2016) have been
shown. An alternative method to target mcrA without having to extract mRNA, is measuring
the coenzyme F430, a coenzyme of mcrA gene, using liquid chromatography/ mass
spectrometry. This method can detect methanogens in the range of 600 to 10,000 cells
(Kaneko et al. 2014). Although this method has been used to quantify methanogens in
environmental samples (Kaneko et al. 2014, Takano et al. 2013), more studies should apply
the method in AD systems and explore techniques to reduce the detection limit.
2.5 Omics studies
2.5.1. Metagenomics
Metagenomics, “shotgun” sequencing of environmental DNA, is an approach to
characterizing structure and metabolic potential that can provide greater information than
amplicon gene sequencing approaches discussed in the preceding section. New sequencing
platforms have steeply decreased costs, enabling sufficient sequencing depth for
metagenomics on environmental communities (Vanwonterghem et al. 2014). Notably,
metagenomics does not rely on PCR amplification, thus eliminating concerns regarding
amplification efficiency or primer biases. Further, unlike targeting the 16S rRNA gene,
sequence data from metagenomics can be used to infer functional potential of a microbial
community without relying on taxonomy-based physiological characteristics (Shah et al. 2010,
Shakya et al. 2013).

34

Recent studies have applied metagenomics to study the microbial community response to
inhibitors such as ammonia (Gao et al. 2015, Li et al. 2017a, Werner et al. 2014), LCFA (Beale
et al. 2016, Kougias et al. 2016), temperature (Beale et al. 2016, Pap et al. 2015), and VFA
(Mosbæk et al. 2016). A study that used [U-
13
C] labelled acetate and metagenomics to
evaluate VFA inhibition observed that acetate was consumed by Methanosarcina,
Methanoculleus, and five subspecies of Clostridia that contained the ftfhs gene (Mosbæk et
al. 2016). This indicated that the identified species of Clostridia were potential SAOB, as the
ftfhs gene is a key enzyme for reductive acetogenesis. Metagenomics has also been used to
characterize temperature-based competition between acetoclastic and hydrogenotrophic
methanogens (Pap et al. 2015). At mesophilic temperature, acetoclastic methanogens
dominated the archaeal community. However, a gradual increase to thermophilic
temperature enriched hydrogenotrophic methanogens alongside an increased abundance of
hydrogen producing Fe-hydrogenases associated with syntrophic and fermentative bacteria.
Metagenomics has similarly been used to evaluate methanogenic pathways in response to
elevated ammonia, with a shift from acetoclastic to hydrogenotrophic methanogenesis and
SAO (Gao et al. 2015, Li et al. 2017a, Werner et al. 2014). Beyond revealing the shifts in
microbial community composition and metabolic potential, metagenomics has also been
effective at elucidating communities that would have remained uncharacterized without
using non-targeted approach (Guermazi et al. 2008, Kougias et al. 2017).
More recently, Pacific Biosciences (PacBio) has developed the single-molecule real time
(SMRT) sequencing platform that is capable of long read lengths compared to other second
generation sequencing platforms. According to a recent review, SMRT sequencing is capable
of producing sequence reads averaging 10 kb, and as high as 60 kb (Rhoads and Au 2015).

35

Combining long read SMRT sequence data with high-throughput and high-accuracy
sequences from second generation platforms such as HiSeq2000 can facilitate the
construction of longer and more accurate metagenome assemblies (Frank et al. 2016). These
hybrid assemblies have also been used to study low-abundance and difficult to sequence
phylotypes in AD systems (Hagen et al. 2017), and this technique could allow researchers to
study the role that these phylotypes play in AD systems during inhibition.
A major limitation of metagenomics is that it characterizes communities based on phylogeny
and functional potential, not function according to gene expression or translated proteins. It
is important to note that AD systems rely on low abundance/rare populations (e.g., syntrophic
bacteria in AD), which can be challenging to accurately quantify using molecular approaches
(Shah et al. 2010). A study comparing metagenomics and 16S rRNA gene sequencing found
higher sensitivity and resolution using metagenomics, while 16S rRNA gene sequencing only
captured broad shifts in microbial diversity over time (Poretsky et al. 2014). Another study
used synthetic communities to compare metagenomics and 16S rRNA gene sequencing and
reported that both Illumina and 454 metagenomics outperformed 16S rRNA gene sequencing
(Shakya et al. 2013). Despite the drastic increase in nucleotide database coverage, the
bottleneck for metagenomics remains genome assembly and gene prediction from assembled
reads, particularly because environmental annotations are lacking (Cabezas et al. 2015). It is
likely that future advancements in sequencing technology and algorithms for analysis will aid
genome assembly from complex communities (Vanwonterghem et al. 2014).
2.5.2. Metatranscriptomics
Metatranscriptomics is an RNA-based molecular method that uses “shotgun” sequencing of
reverse transcribed environmental RNA to characterize functional activity of a microbial

36

community (Cabezas et al. 2015). The first study that used metatranscriptomics in AD
compared 16S rRNA sequences from the metatranscriptome, 16S rRNA gene sequences from
the metagenome, and 16S rRNA gene sequences generated via amplicon sequencing in a full-
scale AD (Zakrzewski et al. 2012). In general, the most abundant microbes retrieved using
metagenomics and amplicon sequencing also contributed the majority of 16S rRNA
sequences. However, Archaea represented 2.4% and 12.9% of 16S rRNA gene relative
abundance using metagenomics and amplicon sequencing, respectively, but contributed 24%
of the transcriptional activity. This indicates that archaeal populations may have high
transcriptional activity, even when constituting a relatively small fraction of the community
based on gene abundance. More effective mRNA enrichment methods are needed to better
analyze functional activity, as only 2.6% of the metatranscriptome reads were mRNA, with
more than 90% of the reads rRNA. For this reason, pre-treatment to remove rRNA prior to
sequencing has been investigated as a more targeted approach to understand community
function (He et al. 2010). Further, RNA extraction efficiencies have been found to have biases
in quantitative analysis of metatranscriptomics data (Stark et al. 2014).
Few studies to date have used metatranscriptomics to evaluate AD during inhibition (Nolla-
Ardèvol et al. 2015, Treu et al. 2016). Treu et al. (2016) investigated inhibition by simulating
high LCFA concentrations with oleic acid. Similar to other studies using 16S rRNA gene
sequencing (Ziels et al. 2017, Ziels et al. 2016), Syntrophomonas increased activity at high
LCFA concentrations. However, metatranscriptomics enabled elucidation of two potential
mechanisms of adaptation to high LCFA concentration: (1) upregulation of genes involved in
peptidoglycan and lipopolysaccharides biosynthesis that possibly result in membrane
modification and (2) transcriptional activation of the chemotaxis genes that enable responses

37

to fatty acids gradient (Treu et al. 2016). Further application of metatranscriptomics is likely
to expand our understanding of the mechanisms of inhibition and potential options to avoid
digester failure.
2.5.3. Metaproteomics
Post-translational regulation of proteins prevents the accurate prediction of all activities
based on gene expression measured by metatranscriptomics, a drawback that is avoided by
metaproteomics (Cabezas et al. 2015). In metaproteomics, expressed proteins are
characterized using three main steps: protein extraction, followed by
separation/fractionation, and subsequent detection with mass spectrometry
(Vanwonterghem et al. 2014). Metaproteomics is particularly useful in identifying novel
functional systems and obtaining direct functional insights (Lü et al. 2014, Siggins et al. 2012).
Further, application of metaproteomics with metabolomics, which measures intermediate
cellular products, can provide new information related to changes in microbial activity,
system functioning, and mechanisms of adaptation to inhibitory conditions (Siggins et al.
2012).
Some methodological challenges to widespread application of metaproteomics remain,
including the extraction of high quality protein at sufficient amounts, interference of co-
extracted compounds, and the need for metagenomics data (unless de novo peptide
sequencing is conducted) (Siggins et al. 2012). One metaproteomic study on AD of cellulose
identified more than 500 non-redundant protein functions (Lü et al. 2014). The only study to
our knowledge that used metaproteomics under potentially inhibitory conditions (free
ammonia of 367 mg NH 3-N L
-1
), found that acetoclastic methanogens were present at low
abundance and enzymes associated with Methanosaeta thermophile were detected (Hagen

38

et al. 2017). However, the dominant mechanism of acetate removal was found to be through
SAO by groups closely related to Thermacetogenium phaeum. In fact, this study discovered
two novel uncultured bacteria with necessary genes for both SAO and β-oxidation of LCFA.
Although a number of recent studies have applied metaproteomics to study the functional
activity of AD with various substrates (Jing et al. 2017, Kohrs et al. 2014) more studies are
needed to elucidate changes in protein expression during inhibition.
2.6 Mapping substrate utilization
Anaerobic metabolisms form a complex web of substrate utilization mediated by a broad
consortium of microorganisms. Inhibition of a single organism or group of organisms can
prevent an intermediate substrate from being formed, which in turn may hinder subsequent
metabolisms and ultimately biogas production. It is therefore useful to develop tools that
allow researchers to pair substrate utilization with specific organisms to better understand
the chain of substrate utilization in AD. Studying substrate utilization patterns and community
dynamics under stressed and inhibited conditions provides information that can be used to
develop more robust and reliable AD systems.
Stable isotope and radio isotope labeled compounds are used to link specific substrates with
degradation products. Lettinga et al. (1999) fed batch anaerobic reactors with
14
C labeled
acetate,
14
C labeled bicarbonate, and unlabeled propionate and analyzed the radio isotope
composition of the produced methane to determine the relative contributions of each carbon
source to methane production. The authors used this method to study the effect of low-
temperature inhibition on the propionate degradation pathway. Single-carbon labeled
acetate has been used extensively to study the relative activity of acetoclastic and SAO
metabolisms in AD. Single-carbon labeled acetate produces unique isotope signatures in

39

produced biogas depending on the relative activity of each pathway. This technique has been
used to study methanogenic mechanisms under different temperatures (Karakashev et al.
2006, Nozhevnikova et al. 2007) and during ammonia-induced inhibition (Hao et al. 2015,
Werner et al. 2014). Labeled substrate experiments have also been used in conjunction with
other molecular methods to correlate different methanogenic pathways with specific
organisms and communities (Hao et al. 2015, Ito et al. 2011, Mosbæk et al. 2016, Werner et
al. 2014). Mulat et al. (2014) used membrane inlet quadrupole mass spectrometry (MIMS)
and isotope labeled acetate to monitor the relative activity of acetoclastic and SAO pathways
in near-real-time in a bench-scale reactor.
Naturally occurring isotope fractionation in biogas has also been used to estimate the relative
importance of acetoclastic methanogenic and SAO pathways, however, it has typically been
used in natural environments where the addition of large quantities of isotope labeled
substrates is not practical. This approach is difficult to use because it requires knowledge of
the isotope fractionation of naturally occurring acetate and system specific isotope
fractionation factors (Conrad, 2005). In spite of these limitations, this method has been used
in AD by comparing the relative fractionation of
13
C in biogas produced in uninhibited and
inhibited bioreactors and inferring the naturally occurring isotope fractionation (Hao et al.
2011, Hao et al. 2017).
Stable isotopes can be combined with molecular techniques through DNA-, RNA-, and protein
stable isotope probing (SIP) to reveal information about the phylogeny and activity of specific
organisms responsible for the transformation of a particular substrate. DNA-SIP requires
cellular growth for labelled elements to be incorporated into the DNA and subsequently
detected (Lueders et al. 2016). Due to variations in density based on G-C content, labelled

40

samples must be compared to unlabeled controls (Youngblut and Buckley 2014). In AD, DNA-
SIP has been used to identify the acetoclastic methanogens that dominate under high
ammonia conditions (Hao et al. 2015) and identify cellulose degraders (Li et al. 2009, Limam
et al. 2014). RNA-SIP, in contrast to DNA-SIP, can be used to track labels in both rRNA and
mRNA and does not require cellular replication or growth (Lueders et al. 2016). Applied to
AD, RNA-SIP has been used to trace labelled glucose through glucose-, propionate-, and
acetate-degrading bacteria and acetoclastic methanogens (Ito et al. 2011, 2012) and reveal
the diversity of fatty acid degrading bacteria (Hatamoto et al. 2007). Protein-SIP tracks label
incorporation into proteins, providing information about cell activity as well as phylogeny
(Jehmlich et al. 2010). To maximize the information obtained, it is best to combine protein-
SIP with metagenomics (Jehmlich et al. 2010). This was done in AD to evaluate short-term
changes resulting from high and low acetate concentrations (Mosbæk et al. 2016). These
results revealed the importance of the SAO pathway under high acetate conditions (Mosbæk
et al. 2016).
The fate of labeled substrates can be visualized using techniques such as
microautoradiography (MAR) (Talbot et al. 2008) and nano-scale secondary ion mass
spectrometry (NanoSIMS) (Musat et al. 2016). MAR involves the use of radioisotope labeled
substrates and allows for visualization of actively metabolizing cells. Radioactive decay of the
labeled substrate can be observed by the microbes that have taken up the substrate.
Combined with fluorescence in situ hybridization (FISH), the actively metabolizing microbes
can be identified. MAR-FISH has been used in anaerobic systems to identify novel acetate-
utilizing bacteria (Ito et al. 2011), characterize the propionate oxidizing community (Ariesyady
et al. 2007b), and identify low-abundance, highly active bacterial and archaeal populations

41

(Ariesyady et al. 2007a, Ito et al. 2012). MAR-FISH is limited by radioisotope labeled substrates
with suitable half-lives. While both organic substrates and carbon dioxide can be radiolabeled
to target both heterotrophs and autotrophs, radioactive N (
13
N) cannot be used because it
has a very short half-life.
NanoSIMS overcomes the issues of radioisotope labeled substrates by using stable isotopes
for visualization. NanoSIMS can also be combined with FISH (Chapleur et al. 2013) and has
been used to visualize the spatial arrangement and isotopic enrichment of specific
microorganisms from AD (Li et al. 2008, Limam et al. 2014). NanoSIMS has also been
combined with phylogenetic micro-arrays to measure isotopic enrichment of rRNA at much
lower enrichment levels than traditionally used for RNA-SIP (Mayali et al. 2012). Although not
yet used to study AD inhibition, these powerful new molecular tools could help identify
metabolic pathways most sensitive to specific inhibitors.
2.7 Real-time monitoring
2.7.1. Reporters
Fluorescent protein reporters are commonly used to identify the activity of specific proteins
in vivo. A fluorescent protein is encoded onto a vector along with a protein of interest, and
this vector is inserted into the microbes of interest. When the protein is expressed, the
attached fluorescent marker is activated. Green fluorescent protein (GFP) is a widely used
reporter protein, however GFP requires molecular oxygen in order to fluoresce, and is
therefore not suitable for monitoring in anaerobic environments (Reid and Flynn 1997).
Anaerobic GFP (AnGFP) was developed to overcome this challenge and is capable of
producing fluorescence in both aerobic and anaerobic environments (Drepper et al. 2007).
AnGFP has been used to study microbes in the human gut (Landete et al. 2014) and lactic acid

42

producing bacteria (Landete et al. 2015). Another reporter system used in anaerobic systems
is the proprietary SNAP-tag
TM
system, which uses a modified protein derived from human
DNA to tag a protein. The SNAP-tag
TM
protein then bonds with a fluorescent probe (Regoes
and Hehl 2005). SNAP-tag is suitable for anaerobic environments and has been used to label
the nuclei of Giardia organisms (Regoes and Hehl 2005) and to monitor the activity of specific
pathogens in samples of dental plaque (Nicolle et al. 2010). These anaerobic-capable
fluorescent probes allow in vivo studies of specific microorganisms in anaerobic cultures and
could be used to study inhibition.
In addition to fluorescent reporters, other reporters can be used to track activity of specific
organisms in anaerobic systems. Cheng et al. (2016) demonstrated a gas reporter system that
uses the methyl halide transferase gene to produce a halogenated gas (e.g., CH 3F) when a
specific gene is expressed. This gas is measured and correlated to the expression of the gene
in question. The use of this system works in both aerobic and anaerobic environments, and
thus is a promising reporter for use in AD systems. One potential drawback of using reporters
is that they rely on engineered organisms that may be difficult to propagate in complex
microbial communities such as those found in AD. Further, the reporters may constitute a
metabolic burden for the host microorganism, and thus be lost or shed if present on a plasmid
over generations of growth. Despite these limitations, reporters represent a powerful tool for
providing information in near real-time about the activity of specific groups of
microorganisms and their function in response to inhibitors.
2.7.2. MinION
The Oxford Nanopore MinION (MinION) sequencer is a small, portable, low-cost single-
molecule sequencing device. The MinION platform is cost effective and capable of rapidly

43

sequencing DNA and RNA. The rapid sequencing capability of the MinION may be used in real-
time to increase understanding of reactor dynamics at greater temporal resolution. In the
future, it may be possible to use this platform to perform rapid sequencing of microbial
communities in bench-, pilot-, and full-scale AD systems, providing a near-real-time profile of
the microbial community and rapid identification of potential upset conditions in the reactor
(e.g., by tracking abundance/activity of an indicator species).
Early reviews of the MinION indicated that it is capable of long read lengths but was prone to
error rates as high as 38% (Laver et al. 2015, Mikheyev and Tin 2014). According to a more
recent review of the technology, improvements have resulted in a sequencing error rate of
8% (Ip et al. 2015). This error rate is still high compared with other next generation sequencing
platforms. However, the Minion is capable of read lengths of 60 to 300 kbp compared with
roughly 250 bp reads from other next generation sequencers. Researchers are hopeful
improvements in chemistry and bioinformatics associated with the device will improve
sequencing accuracy (Jain et al. 2016).
The platform has been used to characterize mixed microbial communities based on 16S rRNA
gene sequencing in a mock microbial community with low diversity (Benitez-Paez et al. 2016),
a mixed community from produced hydraulic fracturing wastewater (Nobu et al. 2015), and
the microbial community in a mouse gut (Shin et al. 2016). Karst et al. (2016) used MinION
sequencing in conjunction with Illumina sequencing and a molecular tagging method to
reduce error associated with the MinION sequencer when sequencing full-length 16S rRNA
genes and SSU rRNA fragments collected from AD and several other complex environments.
The method used by Karst et al. is primer independent and therefore reduces primer bias.
The authors report a significant increase in species diversity in anaerobic communities when

44

compared with primer-dependent methods. With further development, MinION sequencing
is likely to be an important molecular tool to characterize AD inhibition in near real-time.
2.8 Mitigation strategies
Inhibition can reduce energy recovery or necessitate reseeding under extreme instances,
negatively impacting the economic favorability of AD systems. New molecular methods have
elucidated many inhibition mechanisms in AD and evaluated how common inhibitors impact
microbial community structure and activity. Various mitigation strategies have been studied
to reduce inhibitory effects on the AD microbiome. For example, bioaugmentation, the
addition of key enriched cultures, has been applied as a strategy to increase performance of
AD systems and decrease sensitivity to inhibitors. However, there are mixed reports on the
effectiveness of bioaugmentation, as reviewed by De Vrieze and Verstraete (2016). A study
that bioaugmented via addition of a methanogenic propionate degrading community (0.3 g
dry cell weight L
-1
d
-1
) at high ammonia stress conditions (3.0 g N L
-1
), reported that methane
recovery rate increased by 21% and propionic acid degradation increased by 51% after 45
days, compared to a non-bioaugmented reactor (Li et al. 2017b). The increased performance
was partly attributed to enrichment of Methanosaetaceae, the most abundant methanogenic
population in the bioaugmentation culture (> 90% relative abundance). Further, recovery of
the non-bioaugmented reactor after near failure (almost no methane production after 75
days) was demonstrated by routinely adding a double dosage (0.6 g dry cell weight L
-1
d
-1
) of
the bioaugmentation culture. Similarly, another study reported that bioaugmentation of
Methanoculleus bourgensis MS2 in a CSTR with elevated ammonia concentration of 5 g NH 3
L
-1
led to a 31.3% increase in methane production compared to a non-bioaugmented reactor
(Fotidis et al. 2014). Relative abundance of Methanoculleus increased by 5-fold after 39 days

45

after bioaugmentation, suggesting that the bioaugmented culture was functionally active in
the CSTR. An earlier study by the same authors, however, reported that bioaugmentation of
an ammonia tolerant SAOB co-culture, Clostridium ultunense spp. nov., and Methanoculleus
spp. strain MAB1 in a UASB reactor subjected to ammonia stress, did not prevent system
failure (Fotidis et al. 2013). The authors hypothesized that slow growth of methanogens in
the co-culture limited success of these experiments. Another study that tested
bioaugmentation to mitigate ammonia inhibition was similarly unsuccessful (Westerholm et
al. 2012).
An alternative strategy to prevent inhibition in AD is temporal acclimation to inhibitors which
can result in microbial community adaptation (Dai et al. 2016, Gao et al. 2015, Silva et al.
2014, Silvestre et al. 2011). A study that used this approach for digestion of a protein-rich
substrate in a CSTR, reported that in situ acclimation led to tolerance to ammonia
concentrations of up to 4.2 g L
-1
, where relative abundance of Firmicutes and
hydrogenotrophic methanogens increased in response to elevated ammonia (Gao et al.
2015). Another study that compared bioaugmentation and acclimation as a strategy to
decrease inhibition of LCFA showed that long-term acclimation (>100 days) by applying an
increasing load of oleate resulted in reduced lag phase in biogas production, whereas
bioaugmentation of a co-culture of Syntrophomonas zehnderi and Methanobacterium
formicicum did not have any significant impact (Silva et al. 2014). Co-digestion of different
substrates has also been suggested by some studies as a relatively simple method to dilute
substrate feed that might contain inhibitors (Astals et al. 2014, Pagés-Díaz et al. 2014). For
example, a study that used slaughterhouse waste feed reported that dilution of inhibitory
compounds with co-digestion led to improved methane yield compared to mono-digestion

46

reactors (Astals et al. 2014). Other methods for inhibition mitigation include thermal pre-
treatment (Ennouri et al. 2016), alkaline pre-treatment (Koyama et al. 2017), and enzyme
addition (Meng et al. 2017). A study that evaluated the addition of three lipases to hydrolyze
food waste rich in crude lipid revealed that using two of the lipases increased methane
production by 81-158% in animal fat, 27-54% in vegetable oil, and 37-41% in floatable grease
waste digestions (Meng et al. 2017).
Electrically conductive support media could also be an approach to enhance performance in
AD systems via DIET, an electron exchange mechanism that does not require diffusive
molecules (H 2 and formate) for electron transfer (Summers et al. 2010). Researchers have
thus considered its application as a strategy to increase performance (Kato et al. 2012, Lin et
al. 2017) and potentially mitigate inhibition. DIET can occur through biosynthesized
nanowires and pili or through addition of a semi-conductive compound such as activated
carbon (Liu et al. 2012), biochar (Chen et al. 2014c), nano-magnetite (Jing et al. 2017), or
graphene (Lin et al. 2017). DIET has the potential to enhance performance because traditional
electron transfer between syntrophic bacteria and methanogens can be rate limiting in AD
(Stams 1994). It has been demonstrated that both Methanosarcina (Rotaru et al. 2014a) and
Methanosaeta (Rotaru et al. 2014b) can receive electrons via DIET. Various semi-conductive
minerals have been shown to facilitate electron transfer in mineral-based DIET. For example,
supplying haematite or magnetite resulted in an increased abundance of Geobacter spp., a
common exoelectrogen, in a study that used rice paddy field soil to enrich methanogens (Kato
et al. 2012). The study noted that when methanogenic inhibitors were added, the growth of
Geobacter spp. also declined, suggesting that Geobacter only grew in syntrophy with
methanogens (Kato et al. 2012). Further, the study reported that the supplementation of the

47

iron oxides resulted in faster methane production rate and reduced lag phase in
methanogenesis (Kato et al. 2012). Similarly, the use of graphene nanomaterials resulted in a
25% increase in methane yield (Lin et al. 2017). Theoretical calculations indicated that DIET
facilitated higher electron transfer flux than was achievable by electron transfer via diffusive
molecules. Microbial community analysis revealed Geobacter and Pseudomonas as electron
donors and Methanobacterium and Methanospirillum as possible electron receivers. Another
study demonstrated mineral based DIET using carbon cloth in co-cultures of Geobacter
metallireducens and Methanosarcina barkeri (Chen et al. 2014b). The use of mutant
Geobacter metallireducens strains in the co-culture, lacking electrically conductive pili or pili
associated cytochromes, facilitated the distinction from pili-based DIET. Supplying AD with
semi-conductive minerals could be an effective strategy to increase the efficiency of
methanogenesis, but further studies need to be conducted to confirm the prospects of using
DIET for AD inhibition prevention.
2.9 Conclusion
Researchers are beginning to use a diverse set of molecular tools to elucidate microbial
community interactions in AD systems to better understand and devise strategies to prevent
inhibition. The AD community must continue to develop and employ advanced molecular
tools to gain a mechanistic understanding of how inhibitors influence stability and standardize
sequencing methods along with better metadata reporting to facilitate cross-study analysis.
Techniques that map substrate consumption and offer real-time feedback could provide
breakthroughs in inhibition prevention. It is important that we couple these advanced tools
with hypothesis-driven research to improve resiliency and broaden implementation of AD
systems.

48

Acknowledgements
YMA was supported by a Provost Fellowship from the University of Southern California and
the National Science Foundation under Grant No. CBET-1605715. MZA was supported by the
European Union’s Horizon 2020 research and innovation programme under the Marie
Skłodowska-Curie project MicroArctic under grant agreement No 675546. AMB was
supported by a fellowship from the Rice University Civil and Environmental Engineering
Department.
Appendix A. S1 Supplementary data

49

2.10 Literature Cited
Amha, Y.M., Sinha, P., Lagman, J., Gregori, M. and Smith, A.L. (2017) Elucidating microbial
community adaptation to anaerobic co-digestion of fats, oils, and grease and food waste. Water
research.
Ariesyady, H.D., Ito, T. and Okabe, S. (2007a) Functional bacterial and archaeal community structures
of major trophic groups in a full-scale anaerobic sludge digester. Water research 41(7), 1554-1568.
Ariesyady, H.D., Ito, T., Yoshiguchi, K. and Okabe, S. (2007b) Phylogenetic and functional diversity of
propionate-oxidizing bacteria in an anaerobic digester sludge. Applied Microbiology and
Biotechnology 75(3), 673-683.
Astals, S., Batstone, D.J., Mata-Alvarez, J. and Jensen, P.D. (2014) Identification of synergistic impacts
during anaerobic co-digestion of organic wastes. Bioresource Technology 169, 421-427.
Azman, S., Khadem, A.F., Plugge, C.M., Stams, A.J., Bec, S. and Zeeman, G. (2017) Effect of humic acid
on anaerobic digestion of cellulose and xylan in completely stirred tank reactors: inhibitory effect,
mitigation of the inhibition and the dynamics of the microbial communities. Applied Microbiology
and Biotechnology 101(2), 889-901.
Azman, S., Khadem, A.F., Van Lier, J.B., Zeeman, G. and Plugge, C.M. (2015) Presence and role of
anaerobic hydrolytic microbes in conversion of lignocellulosic biomass for biogas production. Critical
Reviews in Environmental Science and Technology 45(23), 2523-2564.
Baserba, M.G., Angelidaki, I. and Karakashev, D. (2012) Effect of continuous oleate addition on
microbial communities involved in anaerobic digestion process. Bioresource Technology 106, 74-81.
Beale, D.J., Karpe, A.V., McLeod, J.D., Gondalia, S.V., Muster, T.H., Othman, M.Z., Palombo, E.A. and
Joshi, D. (2016) An ‘omics’ approach towards the characterisation of laboratory scale anaerobic
digesters treating municipal sewage sludge. Water research 88, 346-357.
Benitez-Paez, A., Portune, K.J. and Sanz, Y. (2016) Species-level resolution of 16S rRNA gene
amplicons sequenced through the MinION™ portable nanopore sequencer. GigaScience 5(1), 4.
Blazewicz, S.J., Barnard, R.L., Daly, R.A. and Firestone, M.K. (2013) Evaluating rRNA as an indicator of
microbial activity in environmental communities: limitations and uses. The ISME journal 7(11), 2061-
2068.
Cabezas, A., de Araujo, J.C., Callejas, C., Galès, A., Hamelin, J., Marone, A., Sousa, D.Z., Trably, E. and
Etchebehere, C. (2015) How to use molecular biology tools for the study of the anaerobic digestion
process? Reviews in Environmental Science and Bio/Technology 14(4), 555-593.
Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., Fierer, N.,
Peña, A.G., Goodrich, J.K. and Gordon, J.I. (2010) QIIME allows analysis of high-throughput
community sequencing data. Nature methods 7(5), 335-336.
Carballa, M., Regueiro, L. and Lema, J.M. (2015) Microbial management of anaerobic digestion:
exploiting the microbiome-functionality nexus. Current opinion in biotechnology 33, 103-111.
Cazier, E.A., Trably, E., Steyer, J.-P. and Escudié, R. (2015) Biomass hydrolysis inhibition at high
hydrogen partial pressure in solid-state anaerobic digestion. Bioresource Technology 190, 106-113.
Chapleur, O., Wu, T.-D., Guerquin-Kern, J.-L., Mazéas, L. and Bouchez, T. (2013) SIMSISH Technique
Does Not Alter the Apparent Isotopic Composition of Bacterial Cells. PloS one 8(10), e77522.
Chen, J.L., Ortiz, R., Steele, T.W.J. and Stuckey, D.C. (2014a) Toxicants inhibiting anaerobic digestion:
a review. Biotechnology advances 32(8), 1523-1534.
Chen, S., Rotaru, A.-E., Liu, F., Philips, J., Woodard, T.L., Nevin, K.P. and Lovley, D.R. (2014b) Carbon
cloth stimulates direct interspecies electron transfer in syntrophic co-cultures. Bioresource
Technology 173, 82-86.
Chen, S., Rotaru, A.-E., Shrestha, P.M., Malvankar, N.S., Liu, F., Fan, W., Nevin, K.P. and Lovley, D.R.
(2014c) Promoting interspecies electron transfer with biochar. Scientific Reports 4, 5019.
Chen, Y., Cheng, J.J. and Creamer, K.S. (2008) Inhibition of anaerobic digestion process: a review.
Bioresource Technology 99(10), 4044-4064.

50

Cheng, H.-Y., Masiello, C.A., Bennett, G.N. and Silberg, J.J. (2016) Volatile Gas Production by Methyl
Halide Transferase: An In Situ Reporter of Microbial Gene Expression In Soil. Environmental science
& technology 50(16), 8750-8759.
Conklin, A., Stensel, H.D. and Ferguson, J. (2006) Growth kinetics and competition between
Methanosarcina and Methanosaeta in mesophilic anaerobic digestion. Water Environment Research
78(5), 486-496.
Dai, X., Yan, H., Li, N., He, J., Ding, Y., Dai, L. and Dong, B. (2016) Metabolic adaptation of microbial
communities to ammonium stress in a high solid anaerobic digester with dewatered sludge.
Scientific Reports 6.
De Francisci, D., Kougias, P.G., Treu, L., Campanaro, S. and Angelidaki, I. (2015) Microbial diversity
and dynamicity of biogas reactors due to radical changes of feedstock composition. Bioresource
Technology 176, 56-64.
de Jonge, N., Moset, V., Møller, H.B. and Nielsen, J.L. (2017) Microbial population dynamics in
continuous anaerobic digester systems during start up, stable conditions and recovery after
starvation. Bioresource Technology 232, 313-320.
De Vrieze, J. and Verstraete, W. (2016) Perspectives for microbial community composition in
anaerobic digestion: from abundance and activity to connectivity. Environmental Microbiology 18(9),
2797-2809.
Demirel, B. and Scherer, P. (2008) The roles of acetotrophic and hydrogenotrophic methanogens
during anaerobic conversion of biomass to methane: a review. Reviews in Environmental Science
and Biotechnology 7(2), 173-190.
DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K., Huber, T., Dalevi, D., Hu,
P. and Andersen, G.L. (2006) Greengenes, a chimera-checked 16S rRNA gene database and
workbench compatible with ARB. Applied and environmental microbiology 72(7), 5069-5072.
Desbois, A.P. and Smith, V.J. (2010) Antibacterial free fatty acids: activities, mechanisms of action
and biotechnological potential. Applied Microbiology and Biotechnology 85(6), 1629-1642.
Drepper, T., Eggert, T., Circolone, F., Heck, A., Krauß, U., Guterl, J.-K., Wendorff, M., Losi, A., Gärtner,
W. and Jaeger, K.-E. (2007) Reporter proteins for in vivo fluorescence without oxygen. Nature
biotechnology 25(4), 443-445.
Edgar, R.C. (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics
26(19), 2460-2461.
Enerdata (2015) Global Energy Statistical Yearbook 2015.
Ennouri, H., Miladi, B., Diaz, S.Z., Güelfo, L.A.F., Solera, R., Hamdi, M. and Bouallagui, H. (2016) Effect
of thermal pretreatment on the biogas production and microbial communities balance during
anaerobic digestion of urban and industrial waste activated sludge. Bioresource Technology 214,
184-191.
Fernandes, T.V., van Lier, J.B. and Zeeman, G. (2015) Humic acid-like and fulvic acid-like inhibition on
the hydrolysis of cellulose and tributyrin. BioEnergy Research 8(2), 821-831.
Fotidis, I.A., Karakashev, D. and Angelidaki, I. (2013) Bioaugmentation with an acetate-oxidising
consortium as a tool to tackle ammonia inhibition of anaerobic digestion. Bioresource Technology
146, 57-62.
Fotidis, I.A., Wang, H., Fiedel, N.R., Luo, G., Karakashev, D.B. and Angelidaki, I. (2014)
Bioaugmentation as a solution to increase methane production from an ammonia-rich substrate.
Environmental science & technology 48(13), 7669-7676.
Frank, J.A., Pan, Y., Tooming-Klunderud, A., Eijsink, V.G., McHardy, A.C., Nederbragt, A.J. and Pope,
P.B. (2016) Improved metagenome assemblies and taxonomic binning using long-read circular
consensus sequence data. Scientific Reports 6.
Franke-Whittle, I.H., Walter, A., Ebner, C. and Insam, H. (2014) Investigation into the effect of high
concentrations of volatile fatty acids in anaerobic digestion on methanogenic communities. Waste
Manag 34, 2080-2089.

51

Gao, S., Zhao, M., Chen, Y., Yu, M. and Ruan, W. (2015) Tolerance response to in situ ammonia stress
in a pilot-scale anaerobic digestion reactor for alleviating ammonia inhibition. Bioresource
Technology 198, 372-379.
Guermazi, S., Daegelen, P., Dauga, C., Rivière, D., Bouchez, T., Godon, J.J., Gyapay, G., Sghir, A.,
Pelletier, E. and Weissenbach, J. (2008) Discovery and characterization of a new bacterial candidate
division by an anaerobic sludge digester metagenomic approach. Environmental Microbiology 10(8),
2111-2123.
Guo, J., Peng, Y., Ni, B.-J., Han, X., Fan, L. and Yuan, Z. (2015a) Dissecting microbial community
structure and methane-producing pathways of a full-scale anaerobic reactor digesting activated
sludge from wastewater treatment by metagenomic sequencing. Microbial cell factories 14(1), 33.
Guo, M., Song, W. and Buhain, J. (2015b) Bioenergy and biofuels: History, status, and perspective.
Renewable and Sustainable Energy Reviews 42, 712-725.
Hagen, L.H., Frank, J.A., Zamanzadeh, M., Eijsink, V.G.H., Pope, P.B., Horn, S.J. and Arntzen, M.Ø.
(2017) Quantitative metaproteomics highlight the metabolic contributions of uncultured phylotypes
in a thermophilic anaerobic digester. Applied and environmental microbiology 83(2), e01955-01916.
Hao, L.-P., Lü, F., He, P.-J., Li, L. and Shao, L.-M. (2011) Quantification of the inhibitory effect of
methyl fluoride on methanogenesis in mesophilic anaerobic granular systems. Chemosphere 84(9),
1194-1199.
Hao, L., Lü, F., Mazéas, L., Desmond-Le Quéméner, E., Madigou, C., Guenne, A., Shao, L., Bouchez, T.
and He, P. (2015) Stable isotope probing of acetate fed anaerobic batch incubations shows a partial
resistance of acetoclastic methanogenesis catalyzed by Methanosarcina to sudden increase of
ammonia level. Water research 69, 90-99.
Hao, L., Mazéas, L., Lü, F., Grossin-Debattista, J., He, P. and Bouchez, T. (2017) Effect of ammonia on
methane production pathways and reaction rates in acetate-fed biogas processes. Water Science
and Technology 75(8), 1839-1848.
Hatamoto, M., Imachi, H., Yashiro, Y., Ohashi, A. and Harada, H. (2007) Diversity of Anaerobic
Microorganisms Involved in Long-Chain Fatty Acid Degradation in Methanogenic Sludges as Revealed
by RNA-Based Stable Isotope Probing. Applied and environmental microbiology 73(13), 4119-4127.
He, S., Wurtzel, O., Singh, K., Froula, J.L., Yilmaz, S., Tringe, S.G., Wang, Z., Chen, F., Lindquist, E.A.
and Sorek, R. (2010) Validation of two ribosomal RNA removal methods for microbial
metatranscriptomics. Nature methods 7(10), 807-812.
Henschel, A., Anwar, M.Z. and Manohar, V. (2015) Comprehensive meta-analysis of ontology
annotated 16S rRNA profiles identifies beta diversity clusters of environmental bacterial
communities. PLoS computational biology 11(10), e1004468.
Huse, S.M., Ye, Y., Zhou, Y. and Fodor, A.A. (2012) A core human microbiome as viewed through 16S
rRNA sequence clusters. PloS one 7(6), e34242.
Ip, C.L., Loose, M., Tyson, J.R., de Cesare, M., Brown, B.L., Jain, M., Leggett, R.M., Eccles, D.A.,
Zalunin, V. and Urban, J.M. (2015) MinION Analysis and Reference Consortium: Phase 1 data release
and analysis. F1000Research 4.
Ito, T., Yoshiguchi, K., Ariesyady, H.D. and Okabe, S. (2011) Identification of a novel acetate-utilizing
bacterium belonging to Synergistes group 4 in anaerobic digester sludge. ISME J 5(12), 1844-1856.
Ito, T., Yoshiguchi, K., Ariesyady, H.D. and Okabe, S. (2012) Identification and quantification of key
microbial trophic groups of methanogenic glucose degradation in an anaerobic digester sludge.
Bioresource Technology 123, 599-607.
Jain, M., Olsen, H.E., Paten, B. and Akeson, M. (2016) The Oxford Nanopore MinION: delivery of
nanopore sequencing to the genomics community. Genome Biology 17(1), 239.
Jehmlich, N., Schmidt, F., Taubert, M., Seifert, J., Bastida, F., von Bergen, M., Richnow, H.-H. and
Vogt, C. (2010) Protein-based stable isotope probing. Nat. Protocols 5(12), 1957-1966.
Jing, Y., Wan, J., Angelidaki, I., Zhang, S. and Luo, G. (2017) iTRAQ quantitative proteomic analysis
reveals the pathways for methanation of propionate facilitated by magnetite. Water research 108,
212-221.

52

Kaneko, M., Takano, Y., Chikaraishi, Y., Ogawa, N.O., Asakawa, S., Watanabe, T., Shima, S., Krüger,
M., Matsushita, M. and Kimura, H. (2014) Quantitative analysis of coenzyme F430 in environmental
samples: a new diagnostic tool for methanogenesis and anaerobic methane oxidation. Analytical
chemistry 86(7), 3633-3638.
Karakashev, D., Batstone, D.J., Trably, E. and Angelidaki, I. (2006) Acetate oxidation is the dominant
methanogenic pathway from acetate in the absence of Methanosaetaceae. Applied and
environmental microbiology 72(7), 5138-5141.
Karst, S.M., Dueholm, M.S., McIlroy, S.J., Kirkegaard, R.H., Nielsen, P.H. and Albertsen, M. (2016)
Thousands of primer-free, high-quality, full-length SSU rRNA sequences from all domains of life.
bioRxiv, 070771.
Kato, S., Hashimoto, K. and Watanabe, K. (2012) Methanogenesis facilitated by electric syntrophy via
(semi) conductive iron-oxide minerals. Environmental Microbiology 14(7), 1646-1654.
Keating, C. (2015) Hydrolysis, methanogenesis and bioprocess performance during low-temperature
anaerobic digestion of dilute wastewater.
Klappenbach, J.A., Dunbar, J.M. and Schmidt, T.M. (2000) rRNA operon copy number reflects
ecological strategies of bacteria. Applied and environmental microbiology 66(4), 1328-1333.
Kohrs, F., Heyer, R., Magnussen, A., Benndorf, D., Muth, T., Behne, A., Rapp, E., Kausmann, R.,
Heiermann, M. and Klocke, M. (2014) Sample prefractionation with liquid isoelectric focusing
enables in depth microbial metaproteome analysis of mesophilic and thermophilic biogas plants.
Anaerobe 29, 59-67.
Kougias, P.G., Campanaro, S., Treu, L., Zhu, X. and Angelidaki, I. (2017) A novel archaeal species
belonging to Methanoculleus genus identified via de-novo assembly and metagenomic binning
process in biogas reactors. Anaerobe.
Kougias, P.G., Treu, L., Campanaro, S., Zhu, X. and Angelidaki, I. (2016) Dynamic functional
characterization and phylogenetic changes due to Long Chain Fatty Acids pulses in biogas reactors.
Scientific Reports 6.
Koyama, M., Watanabe, K., Kurosawa, N., Ishikawa, K., Ban, S. and Toda, T. (2017) Effect of alkaline
pretreatment on mesophilic and thermophilic anaerobic digestion of a submerged macrophyte:
Inhibition and recovery against dissolved lignin during semi-continuous operation. Bioresource
Technology 238, 666.
Kumaraswamy, R., Amha, Y.M., Anwar, M.Z., Henschel, A., Rodríguez, J. and Ahmad, F. (2014)
Molecular analysis for screening human bacterial pathogens in municipal wastewater treatment and
reuse. Environmental science & technology 48(19), 11610-11619.
Landete, J.M., Langa, S., Revilla, C., Margolles, A., Medina, M. and Arqués, J.L. (2015) Use of
anaerobic green fluorescent protein versus green fluorescent protein as reporter in lactic acid
bacteria. Applied Microbiology and Biotechnology 99(16), 6865-6877.
Landete, J.M., Peirotén, Á., Rodríguez, E., Margolles, A., Medina, M. and Arqués, J.L. (2014)
Anaerobic green fluorescent protein as a marker of Bifidobacterium strains. International journal of
food microbiology 175, 6-13.
Laver, T., Harrison, J., O’neill, P.A., Moore, K., Farbos, A., Paszkiewicz, K. and Studholme, D.J. (2015)
Assessing the performance of the oxford nanopore technologies minion. Biomolecular detection and
quantification 3, 1-8.
Lee, S.-H., Park, J.-H., Kim, S.-H., Yu, B.J., Yoon, J.-J. and Park, H.-D. (2015) Evidence of syntrophic
acetate oxidation by Spirochaetes during anaerobic methane production. Bioresource Technology
190, 543-549.
Lettinga, G., Rebac, S., Parshina, S., Nozhevnikova, A., Van Lier, J.B. and Stams, A.J. (1999) High-rate
anaerobic treatment of wastewater at low temperatures. Applied and environmental microbiology
65(4), 1696-1702.
Levén, L., Nyberg, K. and Schnürer, A. (2012) Conversion of phenols during anaerobic digestion of
organic solid waste–a review of important microorganisms and impact of temperature. Journal of
environmental management 95, S99-S103.

53

Li, L., He, Q., Ma, Y., Wang, X. and Peng, X. (2015) Dynamics of microbial community in a mesophilic
anaerobic digester treating food waste: relationship between community structure and process
stability. Bioresource Technology 189, 113-120.
Li, L., He, Q., Ma, Y., Wang, X. and Peng, X. (2016) A mesophilic anaerobic digester for treating food
waste: process stability and microbial community analysis using pyrosequencing. Microbial cell
factories 15(1), 65.
Li, N., He, J., Yan, H., Chen, S. and Dai, X. (2017a) Pathways in bacterial and archaeal communities
dictated by ammonium stress in a high solid anaerobic digester with dewatered sludge. Bioresource
Technology.
Li, T., Mazéas, L., Sghir, A., Leblon, G. and Bouchez, T. (2009) Insights into networks of functional
microbes catalysing methanization of cellulose under mesophilic conditions. Environmental
Microbiology 11(4), 889-904.
Li, T., Wu, T.-D., Mazéas, L., Toffin, L., Guerquin-Kern, J.-L., Leblon, G. and Bouchez, T. (2008)
Simultaneous analysis of microbial identity and function using NanoSIMS. Environmental
Microbiology 10(3), 580-588.
Li, Y., Zhang, Y., Sun, Y., Wu, S., Kong, X., Yuan, Z. and Dong, R. (2017b) The performance efficiency of
bioaugmentation to prevent anaerobic digestion failure from ammonia and propionate inhibition.
Bioresource Technology 231, 94-100.
Limam, R.D., Chouari, R., Mazéas, L., Wu, T.-D., Li, T., Grossin-Debattista, J., Guerquin-Kern, J.-L.,
Saidi, M., Landoulsi, A., Sghir, A. and Bouchez, T. (2014) Members of the uncultured bacterial
candidate division WWE1 are implicated in anaerobic digestion of cellulose. Microbiologyopen 3(2),
157-167.
Lin, R., Cheng, J., Zhang, J., Zhou, J., Cen, K. and Murphy, J.D. (2017) Boosting biomethane yield and
production rate with graphene: the potential of direct interspecies electron transfer in anaerobic
digestion. Bioresource Technology.
Liu, C., Li, H., Zhang, Y., Si, D. and Chen, Q. (2016) Evolution of microbial community along with
increasing solid concentration during high-solids anaerobic digestion of sewage sludge. Bioresource
Technology 216, 87-94.
Liu, F., Rotaru, A.-E., Shrestha, P.M., Malvankar, N.S., Nevin, K.P. and Lovley, D.R. (2012) Promoting
direct interspecies electron transfer with activated carbon. Energy & Environmental Science 5(10),
8982-8989.
Lovell, C.R. and Leaphart, A.B. (2005) Community-Level Analysis: Key Genes of CO 2-Reductive
Acetogenesis. Methods in enzymology 397, 454-469.
Lü, F., Bize, A., Guillot, A., Monnet, V., Madigou, C., Chapleur, O., Mazéas, L., He, P. and Bouchez, T.
(2014) Metaproteomics of cellulose methanisation under thermophilic conditions reveals a
surprisingly high proteolytic activity. The ISME journal 8(1), 88-102.
Lueders, T., Dumont, M.G., Bradford, L. and Manefield, M. (2016) RNA-stable isotope probing: from
carbon flow within key microbiota to targeted transcriptomes. Current opinion in biotechnology 41,
83-89.
Mayali, X., Weber, P.K., Brodie, E.L., Mabery, S., Hoeprich, P.D. and Pett-Ridge, J. (2012) High-
throughput isotopic analysis of RNA microarrays to quantify microbial resource use. ISME J 6(6),
1210-1221.
Meng, Y., Luan, F., Yuan, H., Chen, X. and Li, X. (2017) Enhancing anaerobic digestion performance of
crude lipid in food waste by enzymatic pretreatment. Bioresource Technology 224, 48-55.
Mikheyev, A.S. and Tin, M.M.Y. (2014) A first look at the Oxford Nanopore MinION sequencer.
Molecular ecology resources 14(6), 1097-1102.
Mladenovska, Z. and Ahring, B.K. (1997) Mixotrophic growth of two thermophilic Methanosarcina
strains, Methanosarcina thermophila TM-1 and Methanosarcina sp. SO-2P, on methanol and
hydrogen/carbon dioxide. Applied Microbiology and Biotechnology 48(3), 385-388.

54

Moestedt, J., Müller, B., Westerholm, M. and Schnürer, A. (2016) Ammonia threshold for inhibition
of anaerobic digestion of thin stillage and the importance of organic loading rate. Microbial
biotechnology.
Morris, R., Schauer-Gimenez, A., Bhattad, U., Kearney, C., Struble, C.A., Zitomer, D. and Maki, J.S.
(2014) Methyl coenzyme M reductase (mcrA) gene abundance correlates with activity
measurements of methanogenic H2/CO2-enriched anaerobic biomass. Microbial biotechnology 7(1),
77-84.
Mosbæk, F., Kjeldal, H., Mulat, D.G., Albertsen, M., Ward, A.J., Feilberg, A. and Nielsen, J.L. (2016)
Identification of syntrophic acetate-oxidizing bacteria in anaerobic digesters by combined protein-
based stable isotope probing and metagenomics. The ISME journal.
Mulat, D.G., Ward, A.J., Adamsen, A.P.S., Voigt, N.V., Nielsen, J.L. and Feilberg, A. (2014) Quantifying
contribution of synthrophic acetate oxidation to methane production in thermophilic anaerobic
reactors by membrane inlet mass spectrometry. Environmental science & technology 48(4), 2505-
2511.
Müller, B., Sun, L. and Schnürer, A. (2013) First insights into the syntrophic acetate-oxidizing
bacteria–a genetic study. Microbiologyopen 2(1), 35-53.
Musat, N., Musat, F., Weber, P.K. and Pett-Ridge, J. (2016) Tracking microbial interactions with
NanoSIMS. Current opinion in biotechnology 41, 114-121.
Nicolle, O., Rouillon, A., Guyodo, H., Tamanai-Shacoori, Z., Chandad, F., Meuric, V. and Bonnaure-
Mallet, M. (2010) Development of SNAP-tag-mediated live cell labeling as an alternative to GFP in
Porphyromonas gingivalis. FEMS Immunology & Medical Microbiology 59(3), 357-363.
Niu, Q., Qiao, W., Qiang, H. and Li, Y.-Y. (2013) Microbial community shifts and biogas conversion
computation during steady, inhibited and recovered stages of thermophilic methane fermentation
on chicken manure with a wide variation of ammonia. Bioresource Technology 146, 223-233.
Nobu, M.K., Narihiro, T., Rinke, C., Kamagata, Y., Tringe, S.G., Woyke, T. and Liu, W.-T. (2015)
Microbial dark matter ecogenomics reveals complex synergistic networks in a methanogenic
bioreactor. The ISME journal 9(8), 1710-1722.
Nolla-Ardèvol, V., Strous, M. and Tegetmeyer, H.E. (2015) Anaerobic digestion of the microalga
Spirulina at extreme alkaline conditions: biogas production, metagenome, and metatranscriptome.
Frontiers in microbiology 6, 597.
Nozhevnikova, A.N., Nekrasova, V., Ammann, A., Zehnder, A.J., Wehrli, B. and Holliger, C. (2007)
Influence of temperature and high acetate concentrations on methanogenensis in lake sediment
slurries. FEMS microbiology ecology 62(3), 336-344.
Pagés-Díaz, J., Pereda-Reyes, I., Taherzadeh, M.J., Sárvári-Horváth, I. and Lundin, M. (2014)
Anaerobic co-digestion of solid slaughterhouse wastes with agro-residues: synergistic and
antagonistic interactions determined in batch digestion assays. Chemical Engineering Journal 245,
89-98.
Pap, B., Györkei, Á., Boboescu, I.Z., Nagy, I.K., Bíró, T., Kondorosi, É. and Maróti, G. (2015)
Temperature-dependent transformation of biogas-producing microbial communities points to the
increased importance of hydrogenotrophic methanogenesis under thermophilic operation.
Bioresource Technology 177, 375-380.
Pavlostathis, S. and Giraldo-Gomez, E. (1991) Kinetics of anaerobic treatment: a critical review.
Critical Reviews in Environmental Science and Technology 21(5-6), 411-490.
Pereira, M.A., Pires, O.C., Mota, M. and Alves, M.M. (2005) Anaerobic biodegradation of oleic and
palmitic acids: evidence of mass transfer limitations caused by long chain fatty acid accumulation
onto the anaerobic sludge. Biotechnology and bioengineering 92(1), 15-23.
Poretsky, R., Rodriguez-R, L.M., Luo, C., Tsementzi, D. and Konstantinidis, K.T. (2014) Strengths and
limitations of 16S rRNA gene amplicon sequencing in revealing temporal microbial community
dynamics. PloS one 9(4), e93827.

55

Props, R., Kerckhof, F.-M., Rubbens, P., De Vrieze, J., Sanabria, E.H., Waegeman, W., Monsieurs, P.,
Hammes, F. and Boon, N. (2016) Absolute quantification of microbial taxon abundances. The ISME
journal.
Regoes, A. and Hehl, A.B. (2005) SNAP-tag™ mediated live cell labeling as an alternative to GFP in
anaerobic organisms. Biotechniques 39(6), 809.
Reid, B.G. and Flynn, G.C. (1997) Chromophore formation in green fluorescent protein. Biochemistry
36(22), 6786-6791.
Rhoads, A. and Au, K.F. (2015) PacBio sequencing and its applications. Genomics, proteomics &
bioinformatics 13(5), 278-289.
Romsaiyud, A., Songkasiri, W., Nopharatana, A. and Chaiprasert, P. (2009) Combination effect of pH
and acetate on enzymatic cellulose hydrolysis. Journal of Environmental Sciences 21(7), 965-970.
Rotaru, A.-E., Shrestha, P.M., Liu, F., Markovaite, B., Chen, S., Nevin, K.P. and Lovley, D.R. (2014a)
Direct interspecies electron transfer between Geobacter metallireducens and Methanosarcina
barkeri. Applied and environmental microbiology 80(15), 4599-4605.
Rotaru, A.-E., Shrestha, P.M., Liu, F., Shrestha, M., Shrestha, D., Embree, M., Zengler, K., Wardman,
C., Nevin, K.P. and Lovley, D.R. (2014b) A new model for electron flow during anaerobic digestion:
direct interspecies electron transfer to Methanosaeta for the reduction of carbon dioxide to
methane. Energy & Environmental Science 7(1), 408-415.
Schuchmann, K. and Müller, V. (2014) Autotrophy at the thermodynamic limit of life: a model for
energy conservation in acetogenic bacteria. Nature Reviews Microbiology 12(12), 809-821.
Shah, N., Tang, H., Doak, T.G. and Ye, Y. (2010) Comparing bacterial communities inferred from 16S
rRNA gene sequencing and shotgun metagenomics, pp. 165-176.
Shakya, M., Quince, C., Campbell, J.H., Yang, Z.K., Schadt, C.W. and Podar, M. (2013) Comparative
metagenomic and rRNA microbial diversity characterization using archaeal and bacterial synthetic
communities. Environmental Microbiology 15(6), 1882-1899.
Shin, J., Lee, S., Go, M.-J., Lee, S.Y., Kim, S.C., Lee, C.-H. and Cho, B.-K. (2016) Analysis of the mouse
gut microbiome using full-length 16S rRNA amplicon sequencing. Scientific Reports 6.
Siegert, I. and Banks, C. (2005) The effect of volatile fatty acid additions on the anaerobic digestion
of cellulose and glucose in batch reactors. Process Biochemistry 40(11), 3412-3418.
Siggins, A., Gunnigle, E. and Abram, F. (2012) Exploring mixed microbial community functioning:
recent advances in metaproteomics. FEMS microbiology ecology 80(2), 265-280.
Silva, S.A., Cavaleiro, A.J., Pereira, M.A., Stams, A.J.M., Alves, M.M. and Sousa, D.Z. (2014) Long-term
acclimation of anaerobic sludges for high-rate methanogenesis from LCFA. biomass and bioenergy
67, 297-303.
Silvestre, G., Rodríguez-Abalde, A., Fernández, B., Flotats, X. and Bonmatí, A. (2011) Biomass
adaptation over anaerobic co-digestion of sewage sludge and trapped grease waste. Bioresource
Technology 102(13), 6830-6836.
Smets, W., Leff, J.W., Bradford, M.A., McCulley, R.L., Lebeer, S. and Fierer, N. (2016) A method for
simultaneous measurement of soil bacterial abundances and community composition via 16S rRNA
gene sequencing. Soil Biology and Biochemistry 96, 145-151.
Sousa, D.Z., Salvador, A.F., Ramos, J., Guedes, A.P., Barbosa, S., Stams, A.J.M., Alves, M.M. and
Pereira, M.A. (2013) Activity and viability of methanogens in anaerobic digestion of unsaturated and
saturated long-chain fatty acids. Applied and environmental microbiology 79(14), 4239-4245.
Sousa, D.Z., Smidt, H., Alves, M.M. and Stams, A.J.M. (2009) Ecophysiology of syntrophic
communities that degrade saturated and unsaturated long-chain fatty acids. FEMS microbiology
ecology 68(3), 257-272.
Stams, A.J. (1994) Metabolic interactions between anaerobic bacteria in methanogenic
environments. Antonie van Leeuwenhoek 66(1), 271-294.
Stark, L., Giersch, T. and Wünschiers, R. (2014) Efficiency of RNA extraction from selected bacteria in
the context of biogas production and metatranscriptomics. Anaerobe 29, 85-90.

56

Sukenik, A., Kaplan-Levy, R.N., Welch, J.M. and Post, A.F. (2012) Massive multiplication of genome
and ribosomes in dormant cells (akinetes) of Aphanizomenon ovalisporum (Cyanobacteria). The
ISME journal 6(3), 670-679.
Summers, Z.M., Fogarty, H.E., Leang, C., Franks, A.E., Malvankar, N.S. and Lovley, D.R. (2010) Direct
exchange of electrons within aggregates of an evolved syntrophic coculture of anaerobic bacteria.
Science 330(6009), 1413-1415.
Sun, C., Cao, W., Banks, C.J., Heaven, S. and Liu, R. (2016) Biogas production from undiluted chicken
manure and maize silage: A study of ammonia inhibition in high solids anaerobic digestion.
Bioresource Technology 218, 1215-1223.
Takano, Y., Kaneko, M., Kahnt, J., Imachi, H., Shima, S. and Ohkouchi, N. (2013) Detection of
coenzyme F430 in deep sea sediments: A key molecule for biological methanogenesis. Organic
geochemistry 58, 137-140.
Talbot, G., Topp, E., Palin, M. and Massé, D. (2008) Evaluation of molecular methods used for
establishing the interactions and functions of microorganisms in anaerobic bioreactors. Water
research 42(3), 513-537.
Tao, B., Donnelly, J., Oliveira, I., Anthony, R., Wilson, V. and Esteves, S.R. (2017) Enhancement of
microbial density and methane production in advanced anaerobic digestion of secondary sewage
sludge by continuous removal of ammonia. Bioresource Technology 232, 380-388.
Tian, Z., Zhang, Y., Li, Y., Chi, Y. and Yang, M. (2015) Rapid establishment of thermophilic anaerobic
microbial community during the one-step startup of thermophilic anaerobic digestion from a
mesophilic digester. Water research 69, 9-19.
Treu, L., Campanaro, S., Kougias, P.G., Zhu, X. and Angelidaki, I. (2016) Untangling the Effect of Fatty
Acid Addition at Species Level Revealed Different Transcriptional Responses of the Biogas Microbial
Community Members. Environmental science & technology 50(11), 6079-6090.
Tursman, J. and Cork, D. (1989) Influence of sulfate and sulfate-reducing bacteria on anaerobic
digestion technology. Advances in Biotechnological processes 12, 273-285.
Vanwonterghem, I., Jensen, P.D., Ho, D.P., Batstone, D.J. and Tyson, G.W. (2014) Linking microbial
community structure, interactions and function in anaerobic digesters using new molecular
techniques. Current opinion in biotechnology 27, 55-64.
Webster, T.M., Smith, A.L., Reddy, R.R., Pinto, A.J., Hayes, K.F. and Raskin, L. (2016) Anaerobic
microbial community response to methanogenic inhibitors 2-bromoethanesulfonate and propynoic
acid. Microbiologyopen 5(4), 537-550.
Werner, J.J., Garcia, M.L., Perkins, S.D., Yarasheski, K.E., Smith, S.R., Muegge, B.D., Stadermann, F.J.,
DeRito, C.M., Floss, C. and Madsen, E.L. (2014) Microbial community dynamics and stability during
an ammonia-induced shift to syntrophic acetate oxidation. Applied and environmental microbiology
80(11), 3375-3383.
Werner, J.J., Knights, D., Garcia, M.L., Scalfone, N.B., Smith, S., Yarasheski, K., Cummings, T.A., Beers,
A.R., Knight, R. and Angenent, L.T. (2011) Bacterial community structures are unique and resilient in
full-scale bioenergy systems. Proceedings of the National Academy of Sciences 108(10), 4158-4163.
Westerholm, M., Levén, L. and Schnürer, A. (2012) Bioaugmentation of syntrophic acetate-oxidizing
culture in biogas reactors exposed to increasing levels of ammonia. Applied and environmental
microbiology 78(21), 7619-7625.
Xing, D., Ren, N. and Rittmann, B.E. (2008) Genetic diversity of hydrogen-producing bacteria in an
acidophilic ethanol-H2-coproducing system, analyzed using the [Fe]-hydrogenase gene. Applied and
environmental microbiology 74(4), 1232-1239.
Xu, Z., Zhao, M., Miao, H., Huang, Z., Gao, S. and Ruan, W. (2014) In situ volatile fatty acids influence
biogas generation from kitchen wastes by anaerobic digestion. Bioresource Technology 163, 186-
192.
Youngblut, N.D. and Buckley, D.H. (2014) Intra-genomic variation in G+ C content and its implications
for DNA stable isotope probing. Environmental microbiology reports 6(6), 767-775.

57

Yun, Y.M., Kim, D.H., Cho, S.K., Shin, H.S., Jung, K.W. and Kim, H.W. (2016) Mitigation of ammonia
inhibition by internal dilution in high-rate anaerobic digestion of food waste leachate and evidences
of microbial community response. Biotechnology and bioengineering.
Zakrzewski, M., Goesmann, A., Jaenicke, S., Jünemann, S., Eikmeyer, F., Szczepanowski, R., Al-Soud,
W.A., Sørensen, S., Pühler, A. and Schlüter, A. (2012) Profiling of the metabolically active community
from a production-scale biogas plant by means of high-throughput metatranscriptome sequencing.
Journal of biotechnology 158(4), 248-258.
Ziels, R.M., Beck, D.A.C. and Stensel, H.D. (2017) Long-chain fatty acid feeding frequency in
anaerobic codigestion impacts syntrophic community structure and biokinetics. Water research 117,
218-229.
Ziels, R.M., Karlsson, A., Beck, D.A.C., Ejlertsson, J., Yekta, S.S., Bjorn, A., Stensel, H.D. and Svensson,
B.H. (2016) Microbial community adaptation influences long-chain fatty acid conversion during
anaerobic codigestion of fats, oils, and grease with municipal sludge. Water research 103, 372-382.
Ziganshin, A.M., Schmidt, T., Lv, Z., Liebetrau, J., Richnow, H.H., Kleinsteuber, S. and Nikolausz, M.
(2016) Reduction of the hydraulic retention time at constant high organic loading rate to reach the
microbial limits of anaerobic digestion in various reactor systems. Bioresource Technology 217, 62-
71.

58

CHAPTER 3

3. Elucidating Microbial Community Adaptation to Anaerobic Co-
Digestion of Fats, Oils, and Grease and Food Waste

3.1 Abstract
Despite growing interest in co-digestion and demonstrated process improvements (e.g.,
enhanced stability and biogas production), few studies have evaluated how co-digestion impacts
the anaerobic digestion (AD) microbiome. Three sequential bench-scale respirometry
experiments were conducted at thermophilic temperature (50°C) with various combinations of
primary sludge (PS); thickened waste activated sludge (TWAS); fats, oils, and grease (FOG); and
food waste (FW). Two additional runs were then performed to evaluate microbial inhibition at
higher organic fractions of FOG (30-60% volatile solids loading (VSL; v/v)). Co-digestion of PS,
TWAS, FOG, and FW resulted in a 26% increase in methane production relative to digestion of PS
and TWAS. A substantial lag time was observed in biogas production for vessels with FOG addition
that decreased by more than half in later runs, likely due to adaptation of the microbial
community. 30% FOG with 10% FW showed the highest increase in methane production,
increasing 53% compared to digestion of PS and TWAS. FOG addition above 50% VSL was found

59

to be inhibitory with and without FW addition and resulted in volatile fatty acid (VFA)
accumulation. Methane production was linked with high relative activity and abundance of
syntrophic fatty-acid oxidizers alongside hydrogenotrophic methanogens, signaling the
importance of interspecies interactions in AD. Specifically, relative activity of Syntrophomonas
was significantly correlated with methane production. Further, methane production increased
over subsequent runs along with methyl coenzyme M reductase (mcrA) gene expression, a
functional gene in methanogens, suggesting temporal adaptation of the microbial community to
co-digestion substrate mixtures. The study demonstrated the benefits of co-digestion in terms of
performance enhancement and enrichment of key active microbial populations.
Highlights
 Co-digestion of FW and FOG synergistically increased performance
 FOG addition led to high lag phase that decreased with microbial adaptation
 The activity of syntrophic fatty-acid oxidizers showed the highest influence on
performance
 30% FOG + 10% FW volatile solid loading increased methane production by 53%
 >50% FOG volatile solid loading addition was inhibitory to syntrophs & methanogens
3.2 Introduction
Anaerobic digestion (AD) of sludge generated during wastewater treatment recovers energy in
the form of methane-rich biogas and reduces sludge volume, lessening further treatment and
disposal. However, AD requires long retention times and can be sensitive to feed and operational
conditions, particularly high free ammonia concentration, relative to aerobic processes (Duan et
al. 2012, Li et al. 2011). One option to enhance the performance of AD is to improve biogas

60

production via co-digestion of multiple substrates (e.g., food waste (FW) or fats, oils, and grease
(FOG) along with primary and waste activated sludges). Co-digestion can increase the organic
loading rate (OLR) but also helps dilute concentrations of potentially toxic compounds,
synergistically improve microbial activity, and enhance the nutrient balance (Dai et al. 2013).
FW is a major component of the organic fraction of municipal solid waste (OFMSW), comprising
up to 19% of landfill solid waste disposal (Kong et al. 2012). An estimated 4 million dry metric
tons per year of landfilled food and processing residuals could potentially generate 248 MWe in
an energy recovery process such as AD (Matteson and Jenkins 2007). In response, California has
introduced new legislation, AB1826, that requires the diversion of organic waste from landfills
for commercial FW generators thus making AD and composting the leading treatment options
(Platt et al. 2014). Under the California law, the minimum threshold of organic waste generation
by businesses will decrease over time, resulting in a greater proportion of the commercial sector
needing to comply. Other states (e.g., Massachusetts, New York, Connecticut, and Vermont) have
instated similar FW diversion regulations (Leibrock 2017). Permitting composting facilities within
dense cities, such as Los Angeles, is challenging given potential odor concerns and large land
requirements. Locating compositing facilities outside of large cities is also problematic, as
transport distances result in environmental impacts that could potentially outweigh the benefits
of landfill diversion altogether. We have documented via life cycle assessment that food waste
composting has significant environmental impacts, particularly with regards to eutrophication
and respiratory impacts, relative to food waste co-digestion (Becker et al. 2017). Therefore, it is
anticipated that AD will play a major role in FW management in the near future.

61

FOG is a lipid rich waste material that originates from cooking and food processing related
industries, such as slaughterhouses and dairy farms (Alqaralleh et al. 2016). Due to the negative
impacts of FOG in sewer systems (e.g., formation of hardened deposits that reduce conveyance
capacity), grease traps and grease interceptors are placed in collection systems to remove FOG
from waste streams (Long et al. 2012). Conventionally, the collected FOG is landfilled, however,
newer disposal methods such as AD are becoming more attractive to recover energy and reduce
environmental impacts related to disposal (Alqaralleh et al. 2016). In addition, co-digestion of
FOG with domestic wastewater sludges has shown a significant increase in biogas production
(Long et al. 2012). One concern is that key microbial populations such as acetogens and
methanogens could be inhibited by high levels of FOG (Long et al. 2012). Accumulation of long
chain fatty acids (LCFA) are thought to damage cell membranes, reduce nutrient transport, and
decrease cell permeability affecting the cell’s ability to regulate pH (Long et al. 2012, Palatsi et al.
2010, Sousa et al. 2013). Another concern in co-digestion of high strength organics alongside
domestic wastewater sludges is the fluctuating characteristics of these substrates and the
potential impacts on populations susceptible to toxins, such as methanogens (Karakashev et al.
2005). Therefore, it is crucial to understand how co-digestion affects the AD microbiome to
prevent full-scale performance upsets.
The present study was motivated by an initial full-scale test by LA Sanitation’s Hyperion
Wastewater Treatment Plant (WWTP) in which an increase in biogas production was observed
when FW and vegetable cooking oil (VCO) were co-digested with primary sludge (PS) and
thickened waste activated sludge (TWAS) relative to digesters receiving only PS and TWAS (Amha
et al. 2015a). Co-digestion of VCO at 2% volatile solids loading (VSL) and food waste at 5% VSL

62

led to a 10-15% increase in biogas production, while co-digestion of either substrate alone at the
same VSL produced a nominal 0% increase in biogas production (Amha et al. 2015a). In addition,
a separate experiment showed that FOG addition at 15% VSL led to a 30% increase in biogas
production compared to a control digester only fed with PS and TWAS (Amha et al. 2015a). In the
present study, bench-scale respirometry was used as a more controlled approach to confirm full-
scale observations by evaluating a range of substrate mixtures and loading rates. Other bench-
scale studies have previously demonstrated performance-based benefits of FW co-digestion
(Koch et al. 2015, Koch et al. 2016, Li et al. 2016), FOG co-digestion (Davidsson et al. 2008, Girault
et al. 2012, Kabouris et al. 2009, Noutsopoulos et al. 2013, Wang et al. 2013, Ziels et al. 2016),
and simultaneous co-digestion of FW and FOG (Li et al. 2011, Wu et al. 2016, Xu et al. 2015) . Few
studies have evaluated the impact of co-digestion of FOG on the AD microbiome (Girault et al.
2012, Martín-González et al. 2011, Xu et al. 2015, Yang et al. 2016, Ziels et al. 2015, Ziels et al.
2016). The majority of these studies were done at mesophilic temperatures (Davidsson et al.
2008, Koch et al. 2015, Li et al. 2011, Li et al. 2016, Wang et al. 2013, Xu et al. 2015, Yang et al.
2016, Ziels et al. 2016), with only a few studies done at thermophilic temperatures (Alqaralleh et
al. 2016, Kabouris et al. 2009, Martín-González et al. 2011, Xu et al. 2015, Zhu et al. 2015).
Recent advances in molecular biological tools have led to an influx in research evaluating
microbial community structure and activity on AD performance. AD is a relatively complex chain
of microbial processes including hydrolysis, fermentation, acetogenesis, and methanogenesis,
where methanogenic archaea are divided into acetoclastic and hydrogenotrophic depending on
electron donor. A few studies have evaluated AD community dynamics during co-digestion of
FOG and/or FW using traditional molecular methods (Hatamoto et al. 2007, Martín-González et

63

al. 2011, Palatsi et al. 2010) and high-throughput DNA-based sequencing (Li et al. 2016, Sundberg
et al. 2013, Yang et al. 2016, Ziels et al. 2016). However, DNA-based sequencing approaches
detect inactive cells (Amha et al. 2015b) and can be inaccurate during perturbation in AD (De
Vrieze et al. 2016), making it difficult to link function and microbial community structure.
Therefore, the current study aims to understand how co-digestion of FW and FOG influences
microbial community structure and activity using both DNA- and RNA-based Illumina sequencing
of the 16S rRNA gene and 16S rRNA, respectively, and reverse transcription quantitative PCR (RT-
qPCR) of the methyl coenzyme M reductase (mcrA) gene in methanogens. To our knowledge, this
is the first study to use both DNA and RNA-based Illumina sequencing to study co-digestion of
FW and FOG. These molecular tools were employed across three sequential bench-scale
respirometery runs conducted under varying substrate mixtures at thermophilic conditions and
during two additional runs evaluating FOG inhibition at high VSLs. Outcomes of the research are
expected to inform decisions at the full-scale regarding substrate ratios and maximum advisable
OLRs.
3.3 Materials & materials
3.3.1 Sample collection and chemical assays
PS, TWAS, FOG, and seed (biomass from a full-scale thermophilic AD) samples were collected
from Hyperion WWTP (Los Angeles, CA). FW samples were obtained from Divert, Inc. (Compton,
CA). Biomass samples for DNA and RNA analyses were processed onsite during sample collection
by centrifuging and decanting the supernatant to pelletize biomass in 2 mL centrifuge tubes. For
RNA preservation, biomass was stabilized using DNA/RNA shield (Zymo Research, Irvine, CA).

64

Preserved biomass samples were transported on ice to the University of Southern California
(USC) after which they were stored at -80
o
C until further processing.
3.3.2 Bench-scale anaerobic respirometry
Bench-scale respirometry experiments were conducted using a MPA-200 Methane Potential
Analyzer (Challenge Technology, Springdale, AR) containing 15 replicate 500 mL continuously-
stirred glass reactor vessels. Three sequential batch runs (Runs 1-3) were conducted containing
PS, TWAS, FOG, and FW according to Table 5 (% v/v) at a TVS concentration of approximately 10
g/L. The effective liquid volume was maintained at 400 mL by adding PBS buffer (137 mM NaCl,
2.7 mM KCl, 10 mM Na 2HPO 4, and 2 mM KH 2PO 4) as necessary. For Run 1, all reactor vessels were
inoculated with 200 mL of seed from a full-scale thermophilic AD at Hyperion WWTP. However,
for the subsequent two runs (Run 2 and 3), biomass from the preceding run fed with all substrates
(PS, TWAS, FOG, and FW) was used as seed to evaluate if microbial community adaptation was
beneficial to performance in other substrate mixtures considered. Although not reseeding in this
way may have led to more dramatic changes in microbial community structure and relative
activity, this approach enabled evaluation of each run independent of differences in inoculum.
The control vessels were only inoculated with 200 mL of seed and PBS buffer to attain the
effective liquid volume. Therefore, these vessels did not contain other substrates (Table 5). Prior
to startup, the vessels were purged with nitrogen gas for 10 min to create anaerobic conditions.
Subsequently, the vessels were partially submerged in a recirculating water bath to maintain a
thermophilic temperature of 50
o
C to replicate conditions of the full-scale AD at Hyperion WWTP.
Vessels were continuously stirred at 250 rpm using magnetic stir bars. Biogas composition,

65

chemical assay, and lag-phase (using non-linear regression) were analyzed as described in
Appendix A S2.1.
Table 5. Substrate volatile solid loading for bench-scale anaerobic respirometry. Each condition
was run in triplicate vessels, containing 10 g L
-1
of TVS.
% Volatile solid loading
Run 1-3 PS WAS FW FOG
Control (1-3) 0 0 0 0
PS+TWAS (4-6) 50 50 0 0
PS+TWAS+FOG (7-9) 45 45 0 10
PS+TWAS+FW (10-12) 45 45 10 0
PS+TWAS+FOG+FW (13-15) 40 40 10 10
Run 4 (FOG Inhibition)
Control (1-3) 0 0 0 0
30%FOG+FW (4-6) 30 30 10 30
30%FOG (7-9) 35 35 0 30
60%FOG+FW (10-12) 15 15 10 60
60%FOG (13-15) 20 20 0 60
Run 5 (FOG Inhibition)
Control (1-3) 0 0 0 0
30%FOG+FW (4-6) 30 30 10 30
40%FOG+FW (7-9) 35 35 0 30
50%FOG+FW (10-12) 15 15 10 60
60%FOG+FW (13-15) 20 20 0 60

Two additional runs (Run 4 and 5) were conducted to assess potential inhibition during high FOG
addition. For Run 4, 30% and 60% FOG TVS addition was evaluated with and without 10% FW
TVS, whereas in Run 5, 30%, 40%, 50%, and 60% FOG TVS addition were evaluated with 10% FW
TVS (Table 5) and the balance from PS and TWAS (i.e., identical TVS addition to all vessels). For
Run 4, sludge from a full-scale thermophilic AD at Hyperion WWTP was used as seed, whereas
for Run 5, biomass from Run 4 30% FOG+FW condition was used as seed for all vessels. In both
runs, 200 mL of seed was used to inoculate the vessels and all conditions were run in triplicates.

66

As in the first phase of the study, the vessels were kept at 50
o
C and continuously stirred at 250
rpm.
3.3.3 Nucleic acids extraction and cDNA synthesis
Biomass samples were taken at the beginning and end of each run, centrifuged at 5,000 x g for 5
min at 4
o
C, decanted, and preserved at -80
o
C until further processing. DNA and RNA extraction
from preserved biomass was performed using Maxwell extraction kits (Promega, Madison, WI),
as further described in Appendix S1. For the RNA extracts, an additional DNase treatment was
conducted using DNA-free™ DNA Removal Kit (Invitrogen, Carlsbad, CA) to remove DNA
contamination. Reverse transcription to generate single-stranded complementary DNA (cDNA)
from RNA extracts was performed using GoScript™ Reverse Transcription System according to
manufacturer’s instructions (Promega, Madison, WI). 100 ng of RNA was taken from each sample
for cDNA synthesis.
3.3.4 PCR and sequencing
Due to the potential for differences in microbial community structure and activity within replicate
vessels under the same substrate mixtures, triplicate DNA and RNA samples for Run 1, and
triplicate DNA and duplicate RNA samples for Run 3 were taken for analysis (Table S3). PCR for
16S rRNA gene sequencing and 16 rRNA (cDNA transcript) sequencing was conducted using a
universal 16S rRNA gene primer set targeting the V4 region (Appendix S1). The sequencing results
were analyzed using mothur (Schloss et al. 2009) according to the Schloss MiSeq SOP. Resulting
sequences from the first phase of the study (Run 1-3) and the FOG inhibition study (Run 4-5) were
analyzed separately, as described in detail in Appendix S1. We define “relative abundance” as the
percentage of 16S rRNA gene sequences for a given population out of the total 16S rRNA gene

67

sequences and define “relative activity” as the percentage of 16S rRNA sequences for a given
population out of the total 16S rRNA sequences. Spearman’s rank was used for non-parametric
analysis to correlate the microbial community profile with methane production using MaxStat
3.6 (Germany). Significant correlation was defined as p<0.05. All raw sequences form this study
are available in NCBI’s Sequence Read Archive (SRA) database (Leinonen et al. 2010) (SRA Study
SRP102955).
3.3.5 Reverse transcription-quantitative PCR (RT-qPCR)
The 16S rRNA and mcrA transcripts were quantified using RT-qPCR. Positive controls for 16S rRNA
and mcrA were isolated from a pool of 10 cDNA samples used in the study and pooled by equal
mass. PCR to prepare the RT-qPCR standards was performed as previously described (Smith et al.
2015) (Appendix S1). Genomic DNA of Thermus thermophilus isolated from pure culture
(American Type Culture Collection (ATCC) 27634) was used as a negative control and ultra-pure
DNase/RNase free water was used as a no-template control. Melt curve analysis was conducted
to check for specificity of amplifications. The R
2
value for the 16S rRNA standard curve was 1.00
with an average efficiency of 78.0%. The R
2
value for the mcrA standard curve was also 1.00 and
the average efficiency was 84.0%.
3.3.6 Reagent controls
To evaluate potential contamination from extraction kits and reagents, sequencing was
performed on serial dilutions of a pure culture using a similar approach to Salter et al. (Salter et
al. 2014). Thermus thermophilus (ATCC 27634) was used as an internal standard to make six, ten-
fold serial dilutions, which were taken for DNA extraction and subsequent sequencing targeting

68

the 16S rRNA gene. In addition, qPCR was conducted on DNA and reverse transcribed cDNA from
RNA extracts of the different dilutions of Thermus thermophilus to quantify the observed
contamination. Further details on methodology and results from these observations can be found
in SI Section II.
3.4 Results and discussion
3.4.1 Co-digestion of FOG and FW reproducibly increased methane production
Bench-scale respirometery of the substrate mixtures indicated reproducible improvement in
methane production, with Run 3 PS+TWAS+FOG+FW and PS+TWAS+FOG exhibiting 26.0 ± 8.1%
and 21.1 ± 6.6% increase in methane production relative to PS+TWAS, respectively. While
PS+TWAS+FOG and PS+TWAS+FOG+FW also showed an increase in methane production in the
preceding runs (Figure 5), PS+TWAS+FW only showed an increase in methane production relative
to PS+TWAS by Run 3 (18.4 ± 3.2%). In Run 3, mean methane production for PS+TWAS+FW,
PS+TWAS+FOG, and PS+TWAS+FOG+FW was significantly greater than PS+TWAS (unpaired one-
tail t-test, p<0.05). The differences in mean methane production were 19.1 mL CH 4 g TVS
-1

(p=0.0063) for PS+TWAS+FW, 22.1 mL CH 4 g TVS
-1
(p=0.0145) for PS+TWAS+FOG, and 27.0 mL
CH 4 g TVS
-1
(p=0.0206) for PS+TWAS+FOG+FW. Therefore, FOG addition was associated with
greater improvement in methane production than FW addition, confirming full-scale
observations at Hyperion WWTP (Amha et al. 2015a). Full-scale experiments observed a 30%
increase in digester gas production during 15% VSL FOG co-digestion compared to a control
digester fed with only PS and TWAS. Conversely, co-digestion of FW at 10% VSL only increased
digester gas production by 10% (Amha et al. 2015a). Bench-scale PS+TWAS+FW showed a higher
improvement in performance in Run 3 than was obtained at the full scale (Amha et al. 2015a),

69

but overall PS+TWAS+FW performance was inconsistent with insignificant performance
improvements in Run 1 and 2. The inconsistency observed may have resulted from fluctuating
FW characteristics, which is a potential risk in full-scale facilities, particularly when FW comprises
a large fraction of the OLR. Methane production increased in subsequent runs across all substrate
mixtures, with the highest production observed in Run 3 consistently (Figure 5). For example,
mean methane production for PS+TWAS+FOG+FW increased by 38.8 mL CH 4 g TVS
-1
(p=0.0125)
in Run 3 relative to Run 1. It is important to note that all vessels in Run 2 and 3 were seeded with
the same inoculum, biomass from the preceding run from the PS+TWAS+FOG+FW substrate
mixture. Therefore, microbial community adaptation to the co-digestion substrate mixture
improved performance in Run 2 and 3 across all substrate mixtures (vessels) suggesting that
temporary co-digestion can be beneficial in improving process performance.
Non-linear regression analysis of biogas production indicated a decrease in lag phase (λ) from
Run 1 to 3 for substrate mixtures containing FOG (Table 6, Figure S4). The lowest lag phase was
observed for PS+TWAS, ranging from 40.6 to 62.3 h, of all substrate mixtures, followed by
PS+TWAS+FW ranging from 53.0-88.6 h. The largest decrease in lag phase was observed for
PS+TWAS+FOG, which decreased from 118 ± 2 h in Run 1 to 46.3 ± 12.5 h in Run 3. A more
moderate decrease in lag phase was observed for PS+TWAS+FOG+FW, which decreased from 114
± 10 h in Run 1 to 85.0 ± 23.1 h in Run 3. Conversely, PS+TWAS+FW showed an increase in lag
phase, comparing Run 3 and Run 1. These results suggest that addition of FOG increased lag
phase initially but was significantly diminished in subsequent runs due to adaptation of the
microbial community. Acclimation of the microbial community to FOG has been shown to shorten
lag phase and increase methane activity (Silva et al. 2014, Silvestre et al. 2011, Ziels et al. 2016),

70

corroborating our observed significant lag phase reduction with the reseeding of acclimated
biomass from previous runs.
Run 3 PS+TWAS+FOG+FW showed the highest TVS reduction (41.3 ± 7.1%) and lowest free
ammonia concentration (24.9 ± 1.2 mg NH 3 L
-1
), correlating well with the high methane
production observed (Table 7). TVS reduction ranged from 35.7-41.3% in Run 3 for the various
substrate mixtures. Free ammonia concentrations ranged from 24.9 - 38.9 mg NH 3 L
-1
, well under
the inhibitory concentration of 100-150 mg NH 3 L
-1
for AD communities (Hansen et al. 1998).
Sulfate removal ranged from 64-67% under all substrate mixtures and was thus not influenced
by co-digestion. Low concentrations of VFAs were observed at the end of each run for most
vessels. However, PS+TWAS+FW vessels contained elevated propionate concentrations at the
end of Run 3.
Table 6. Non-linear regression analysis on all substrate mixtures in Run 1, 2, and 3.
Run 1 Estimated A
Biogas (mL g
TVS
-1
)
λ (h) Rm (mL g TVS
-
1
.h)
R
2

PS+TWAS 114 ± 2 55.4 ± 3.7 1.09 ± 0.12 0.980 ± 0.001
PS+TWAS+FOG 123 ± 3 118 ± 2 1.32 ± 0.001 0.943 ± 0.004
PS+TWAS+FW 115 ± 7 53.0 ± 3.2 1.10 ± 0.07 0.943 ± 0.012
PS+TWAS+FOG+FW 116 ± 12 114 ± 10 1.30± 0.03 0.909 ± 0.045
Run 2

PS+TWAS 127.0 ± 3.885 62.3 ± 16.5 1.15 ± 0.16 0.989 ± 0.007
PS+TWAS+FOG 145.9 ± 16.22 115 ± 10 1.41 ± 0.32 0.971 ± 0.023
PS+TWAS+FW 120.6 ± 1.672 74.9 ± 16.6 1.13 ± 0.11 0.976 ± 0.012
PS+TWAS+FOG+FW 147.8 ± 32.35 103 ± 9 1.66 ± 0.05 0.980 ± 0.016
Run 3

PS+TWAS 141.9 ± 3.862 40.6 ± 4.3 1.27 ± 0.01 0.993 ± 0.004
PS+TWAS+FOG 156.0 ± 10.55 46.3 ± 12.5 1.64 ± 0.04 0.967 ± 0.031
PS+TWAS+FW 139.2 ± 22.16 88.6 ± 13.0 1.45 ± 0.25 0.952 ± 0.014
PS+TWAS+FOG+FW 142.1 ± 38.28 85.0 ± 23.1 1.53 ± 0.25 0.965 ± 0.030

71

Table 7. TVS removal, free ammonia concentration, pH, VFA concentration, sulfate
concentration, and sulfate reduction for end of Run 3. The detection limit for the IC analyses
was 10 mg L
-1
.

Figure 5. (A) Cumulative biogas production normalized to initial organic loading in mL/g TVS for
Run 1, Run 2, and Run 3. The error bars indicate the standard deviation every 20 hours for
triplicate vessels. (B) Cumulative biogas (solid fill) and methane (pattern fill) for the substrate
mixtures, normalized to initial organic loading in mL/g TVS. Error bars indicate the combined
standard deviation of cumulative biogas/methane production and initial organic loading TVS
measurement for triplicate vessels. Due to experimental error, one of the triplicate vessels for
the following substrate mixtures and runs were excluded from the mean analysis and standard
deviation: PS+TWAS+FOG (in Run 1), PS+TWAS+FW (in Run 3), and PS+TWAS+FOG+FW (in Run
3). Therefore, each bar represents the mean of only two vessels for these substrate mixtures
and runs.
Table 3. TVS removal, free ammonia concentration, pH, VFA concentration, sulfate concentration, and
sulfate reduction for end of Run 3. The detection limit for the IC analyses was 10 mg/L.

TVS
removal
(%)
Free
ammonia
(mg/L)
pH Acetate
(mg/L)
Propionate
(mg/L)
Butyrate
(mg/L)
Formate
(mg/L)
Valerate
(mg/L)
Sulfate
(mg/L)
Sulfate
reduction
(%)
Control 7.80 ±8.1 228 ±14 8.26 ±0.01 84.7 ±26.5 14.9 ±0.3 < 10 < 10 < 10 21.0 ±0.4 81.7 ±2.5
PS+TWAS 35.7 ±2.3 38.9 ±2.4 7.12 ±0.03 16.1 ±0.4 < 10 < 10 < 10 < 10 29.5 ±0.3 64.3 ±1.2
PS+TWAS+FOG 38.0 ±0.6 35.7 ±4.4 7.09 ±0.05 36.7 ±0.3 47.7 ±0.2 < 10 < 10 < 10 57.5 ±1.5 65.9 ±0.1
PS+TWAS+FW 34.1 ±0.9 33.6 ±0.2 7.05 ±0.01 29.7 ±0.1 179 ±0.8 17.5 ±0.0 < 10 73.8±0.7 29.0 ±0.2 67.4 ±1.2
PS+TWAS+FOG+FW 41.3 ±7.1 24.9 ±1.2 6.94 ±0.02 25.6 ±0.3 < 10 94.3 ±0.6 < 10 < 10 33.6 ±0.1 64.1 ±0.9
B.

A.

72

3.4.2 Syntrophic fatty-acid oxidizers were critical to increase methane production
RNA-based sequencing was more sensitive than DNA-based sequencing in evaluating microbial
population dynamics across various substrate mixtures. For example, 16S rRNA sequence data
indicated that Coprothermobacter and Thermotogales were relatively inactive at 50% FOG and
60% FOG during the FOG inhibition study (Run 5). However, 16S rRNA gene sequence data
showed only a relatively minor reduction in their relative abundance. Further, Mycobacterium
had relatively high abundance in all substrate mixtures based on 16S rRNA gene sequence data
(1.61-3.99% relative abundance at end of Run 3; Figure S7), but they were not detected in 16S
rRNA sequencing, suggesting they were inactive (Figure 6). Mycobacterium are known to have
DNA that persists in the environment (Kumaraswamy et al. 2014). Therefore, their detection with
DNA sequencing could be due to inactive cells, highlighting the main shortcoming of DNA-based
analyses. Similar observations have been reported in a previous study under inhibitory
conditions (De Vrieze et al. 2016). Based on these concerns, more emphasis was given to the
RNA-based sequencing in the following discussion on the microbial community. DNA-based
sequencing results can be found in Appendix S1 (Figure S7-Figure S9 and Figure S15-Figure S17).

73

Figure 6. (A) Relative activity based on 16S rRNA sequencing identified at the genus level where
possible for PS+TWAS+FOG+FW for Run 1, Run 2, and Run 3 (B) Relative activity in at the end
of Run 3 for the different substrate mixtures. For Run 1 and Run 3, triplicate and duplicate
samples are shown, respectively, to represent methodological precision. All data are expressed
as a percentage normalized using total 16S rRNA sequences (Bacteria and Archaea). A y-axis
break was used to accentuate differences in lower activity populations.

The Spearman rank correlation analysis showed significant (p<0.05) but weak correlation
(R=0.475, p=0.0189) with total methanogenic relative activity (Table S4), which was obtained by
summing the relative activity of methanogens identified to the genus-level for each substrate
mixture. The relative activity of methanogens increased from Run 1 to Run 3 in all substrate
mixtures except for PS+TWAS (Figure 8A). The increase was highest in PS+TWAS+FOG+FW, where
the relative activity increased from 0.410% at the end of Run 1 to 0.873% at the end of Run 3.
PS+TWAS+FOG+FW also showed the highest methane production increase with subsequent runs

74

(Figure 5). PS+TWAS exhibited a higher relative activity of methanogens in Run 1 and 2 than any
other substrate mixture, but this likely resulted from an increase in hydrolytic bacteria and
fermenter relative activity during co-digestion in all other substrate mixtures rather than an
actual decrease in methanogenic relative activity. The biogas production data supports this
notion. Methanogens also exhibited high relative activity in the control vessels, likely stemming
from relative activity decreases of other populations. According to Spearman rank, the mcrA/16S
rRNA ratio was not significantly correlated with methane production (p=0.6743). Similarly, a
previous study reported that mcrA expression did not correlate with methane production (Morris
et al. 2014). Nonetheless, considering the four different substrate mixtures, the ratio of mcrA
transcripts to 16S rRNA copies increased substantially with increasing run (Figure 7) and with
increasing methane production except for PS+TWAS, where Run 2 ratios exceeded Run 3. In the
control where methane production was lowest in all runs, mcrA gene expression was high, likely
due to decreased relative activity of other populations such as hydrolytic bacteria and fermenters
that were not receiving substrates. Methanogenic relative activity was thus likely artificially
inflated due to endogenous decay. These observations highlight the need for absolute
abundance/activity data to more accurately monitor engineered systems, as discussed in a recent
study that concluded that increases in relative abundance do not necessarily reflect increases in
absolute abundance/outgrowth of taxa (Props et al. 2016). It is also important to note some of
the shortcomings of using 16S rRNA genes and 16S rRNA to infer microbial community structure
and relative activity, respectively. Analysis of 16S rRNA is not linked with a specific cellular
function and does not always relate with activity (Blazewicz et al. 2013). In addition, the
relationship between 16S rRNA copies and growth rate can differ between populations

75

(Blazewicz et al. 2013). Further, using 16S rRNA genes to infer community structure is
complicated by 16S rRNA operon numbers varying from 1-15 per genome (Klappenbach et al.
2000), resulting in potential under- or over-representation of certain taxa.

Figure 7. Relative expression of mcrA in all substrate mixtures for Run 1-3. Copies of mcrA
transcripts were normalized to total 16S rRNA copies. Error bars for mcrA expression represent
the standard deviation of the ratio of triplicate RT-qPCR reactions. Error bars for cumulative
methane production (secondary y-axis) represent the standard deviation for triplicate vessels.

The relative activity of two methanogenic genera, Methanoculleus (R=0.7488, p<0.001) and
Methanosarcina (R=0.521, p=0.009), showed significant positive correlation with methane
production (Table S4). Methanosarcina spp. showed the highest relative activity (Figure 8A) in
most samples. Methanosarcina are mixotrophic methanogens that are able to metabolize
hydrogen, acetate, and C1 compounds (Mladenovska and Ahring 1997, Welander and Metcalf

76

2005), which can be considered an advantage over other methanogens by providing metabolic
flexibility. In addition, obligatory hydrogenotrophic methanogens such as Methanobacterium,
Methanobrevibacter, Methanosphaera, and Methanothermobacter were detected.
Methanosaeta spp., obligate acetoclastic methanogens, were only detected in the control in Run
1 (0.013%) and PS+TWAS+FOG+FW in Run 3 (0.013%). Methanosaeta are known to have higher
substrate affinity for acetate and lower growth rate than Methanosarcina, which means they
typically outcompete Methanosarcina at low acetate concentrations (Conklin et al. 2006). The
extremely low presence of Methanosaeta spp. suggests that acetate oxidation was performed
exclusively by Methanosarcina spp. or via other metabolic processes (e.g., syntrophic acetate
oxidation or sulfate reduction). Notably, acetate did not accumulate in any of the substrate
mixtures (Table 7). Further, sulfate concentrations were relatively low in all of the substrate
mixtures. For example, the influent sulfate concentration for the substrate mixtures averaged
only 87.0 ± 5.2 mg L
-1
for Run 3. Thus, there was limited enrichment of sulfate reducing bacteria,
accounting for less than 0.1% of the relative activity across all substrate mixtures at the end of
Run 3.

77

Figure 8 (A) Relative activity of methanogens identified at the genus level where possible based
on 16S rRNA sequencing and (B) relative activity of syntrophic fatty-acid oxidizers identified at
the genus level where possible using 16S rRNA sequencing. Results are expressed as a
percentage normalized using total of 16S rRNA sequences (Bacteria and Archaea). Truncated y-
axes (0 to 1% and 0 to 6% on figure A and B, respectively) are shown to accentuate differences
in abundance.

The absence of Methanosaeta spp. and low presence of sulfate reducing bacteria could indicate
that syntrophic acetate oxidizers such as Thermacetogenium, Syntrophaceticus, and
Tepidanaerobacter (Westerholm et al. 2011), played a significant role in acetate oxidation. For
example, Tepidanaerobacter showed relative activity of 1.05% and 1.20% in PS+TWAS+FOG and
PS+TWAS+FW, respectively (Figure 8B). Tepidanaerobacter (R=0.419, p=0.042) and
B.

A.

78

Syntrophaceticus (R=0.424, p=0.039) relative activity was also significantly correlated with
methane production. In addition, there was an enrichment of all three syntrophic acetate
oxidizers over sequential runs (Figure S10). Unclassified sequences belonging to the family
Thermoanaerobacteraceae showed a relative activity of 0.026-0.212% in the various substrate
mixtures. Thermoanaerobacteraceae includes known syntrophic acetate oxidizers, such as
Thermacetogenium and Tepidanaerobacter (Mosbæk et al. 2016), but it is difficult to conclude
the role of these groups at this phylogenetic resolution. Syntrophic acetate oxidizers have a
thermodynamic advantage at higher temperatures (Hao et al. 2010) and the thermophilic
temperature in this study could have aided enrichment of these groups over time. It should be
noted that there are few known syntrophic acetate oxidizers (i.e., Syntrophaceticus spp.,
Thermacetogenium phaeum, Thermotoga lettingae, Tepidanaerobacter acetatoxydans, and
Clostridium ultunense (Westerholm et al. 2011)). Recent studies have demonstrated that cluster
II Spirochaetes could perform syntrophic acetate oxidation with hydrogenotrophic methanogens
(Lee et al. 2015), but OTUs classifying as Spirochaetes comprised less than 0.1% relative activity
for most samples here. We did detect high relative activity of OTUs belonging to Clostridium,
however, no OTUs classified within Clostridium XII in Run 1-3, which is the cluster that includes
the known syntrophic acetate oxidizer Clostridium ultunense (Schnürer et al. 1996). Syntrophic
acetate oxidation is likely performed by yet to be described microbial populations rather than
known cultures of syntrophic acetate oxidizers (Werner et al. 2014). Therefore, these
undescribed populations likely play a significant role in acetate consumption here and in other
AD systems. Further research is needed using DNA/RNA-stable isotope probing with
13
C -labeled
acetate to identify and better understand these key populations.

79

The high relative activity of hydrogenotrophic methanogens in all substrate mixtures and runs
indicated the importance of syntrophy in improving biogas production. Oxidation of reduced
compounds, such as fatty acids, by secondary fermenting bacteria is thermodynamically
unfavorable when not coupled with a hydrogen-consuming syntrophic partner to keep hydrogen
partial pressure low (Hattori 2008). Our sequencing results reinforce the importance of this
interspecies interaction. Plotting methane production vs. total relative activity of syntrophic
fatty-acid oxidizers and methanogens for all substrate mixtures (Figure S11) gave a visual
representation of the increased methane production with increasing relative activity of these
groups. Further, syntrophic fatty-acid oxidizers were enriched over time in PS+TWAS+FOG+FW,
although relative activity of specific syntrophs shifted in sequential runs. For example,
Pelotomaculum accounted for the highest relative activity in Run 1 PS+TWAS+FOG+FW but this
population was completely replaced with Syntrophothermus and Syntrophomonas by Run 3.
Syntrophomonas are syntrophic fatty acid β-oxidizing bacteria that are able to oxidize LCFA, and
have been shown to be highly active during FOG co-digestion (Ziels et al. 2016). Pelotomaculum
remained dominant in PS+TWAS across all runs. In PS+TWAS+FOG, these populations resurfaced
in Run 3 after disappearing in Run 2. Tepidanaerobacter significantly increased in relative activity
in all substrate mixtures other than the control from Run 1 to Run 3 (Figure 8B). Overall, observed
increase in relative activity of hydrogenotrophic methanogens and syntrophic fatty-acid oxidizers
with increasing biogas production highlight the significance of syntrophy in co-digestion. Similar
observations have been reported by other researchers (Ziels et al. 2015, Ziels et al. 2016). For
example, a study on the impact of high FOG addition in a mesophilic, semi-continuous anaerobic
digester found an increase of hydrogenotrophic methanogen Methanospirillum from a relative

80

abundance of 1.3% to 34% over 138 days along with an increase in relative abundance of
Syntrophomonas (Ziels et al. 2016).
The observed increase in the sum of the relative activity of Syntrophomonas and
Syntrophothermus in PS+TWAS+FOG+FW relative to PS+TWAS+FOG and PS+TWAS+FW suggests
that these populations of syntrophic fatty-acid oxidizers (Narihiro et al. 2015, Sekiguchi et al.
2000, Sousa et al. 2009) may be key to the synergistic increase in biogas production observed at
the bench and full scale. Relative activity data for Run 3 showed that Syntrophomonas and
Syntrophothermus were not significant in PS+TWAS, each comprising less than 1.00% of the
relative activity. PS+TWAS+FW showed substantially higher relative activity of these key
populations, accounting for 2.53% and 1.44% of the relative activity, respectively. The
enrichment of these populations in PS+TWAS+FW by Run 3 may be the reason for the enhanced
performance observed compared to PS+TWAS. In addition, other populations with significant
positive correlation with methane production (Table S4) also showed high relative activity in
PS+TWAS+FW, such as Clostridium sensu stricto (R=0.571, p=0.004), Clostridium XI (R=0.655,
p=0.0005), and Tepidanaerobacter (R=0.419, p=0.042), with relative activity of 5.50%, 5.65%, and
1.20%, respectively (Figure 6). In PS+TWAS+FOG+FW, Syntrophomonas and Syntrophothermus
were also relatively more active compared to PS+TWAS, comprising 2.51% and 2.25% of the
relative activity, respectively. Interestingly, these populations had low relative activity in
PS+TWAS+FOG in Run 3. Notably, there was a positive correlation between methane production
and total relative activity of syntrophs (R=0.475, p=0.0192) and relative activity of
Syntrophomonas (R=0.406, p=0.0491) (Table S4). A correlation between performance
enhancement and increase in relative abundance of Syntrophomonas was also reported

81

previously (Ziels et al. 2016). In addition, a more recent study reported that the kinetics of LCFA
degradation rate correlated to the relative abundance of Syntrophomonas (Ziels et al. 2017).
Diversity analysis using the inverse Simpson metric revealed that the highest diversity was in
PS+TWAS+FOG+FW and that diversity increased with sequential runs (Figure S13). The inverse
Simpson metric for PS+TWAS+FOG+FW increased by 38.7% from Run 1 to Run 3 while
PS+TWAS+FOG and PS+TWAS+FW showed a more modest temporal increase in diversity. The
increase in diversity was largely driven by upper-level fermenters, which remain somewhat
poorly described in the AD microbiome and deserve further study.
3.4.3 High FOG addition resulted in significant decline in performance and inhibited key
microbial populations
Run 4 demonstrated that 60% FOG addition was severely inhibitory and FW addition at high FOG
only moderately improved performance. A steep decline in methane production occurred at 60%
FOG addition (Figure 9). FW addition increased methane production in the 30% FOG mixtures by
31.0 ± 19.9%, relative to FOG addition without FW. Relative to Run 1 PS+TWAS (Figure 5),
methane production increased by 47.5 ± 9.8% and 12.6 ± 16.0% for 30% FOG+FW and 30% FOG,
respectively. The mean methane production compared to Run 1 PS+TWAS increased by 42.1 mL
CH 4 g TVS
-1
(p=0.0034) for the 30% FOG+FW condition, and 11.2 mL CH 4 g TVS
-1
(p=0.24) for the
30% FOG condition. The increase in mean methane production for the 30% FOG condition
compared to Run 1 PS+TWAS was insignificant (p >0.05) due to high variability within the
triplicate vessels for the substrate mixture.

82

For Run 5, 10% FW addition was applied uniformly to 30%, 40%, 50%, and 60% FOG (Table 5),
since Run 4 suggested that co-digestion of FW at high FOG addition increased performance. All
conditions were seeded with biomass from Run 4 30% FOG+FW. Relative to Run 1 PS+TWAS,
methane production was highest at 30% FOG with a 53.3 ± 18.2% increase. Methane production
also increased at 40% FOG but by only 27.4 ± 19.6% relative to Run 1 PS+TWAS. At higher FOG
loading, methane production decreased relative to Run 1 PS+TWAS by 90.3 ± 1.6 % and 94.4 ±
2.3 % at 50% and 60% FOG, respectively. The increase in mean methane production for 30% FOG
and 40% FOG in Run 5 compared to Run 1 PS+TWAS was 47.2 (p=0.02) and 24.2 (p=0.11) mL CH 4
g TVS
-1
, respectively. Conversely, the 50% FOG and 60% FOG showed a decrease of 79.9
(p<0.0001) and 83.6 (p<0.0001) mL CH 4 g TVS
-1
. One of the triplicate vessels for 40% FOG showed
an outlier in methane production compared to the other replicates suggesting less stability in
performance at this FOG loading. Excluding this outlier vessel, 40% FOG showed similar biogas
production to 30% FOG, although at a lower methane content. Both scenarios showed high
variability within triplicate vessels and long lag phase (Figure S14). Due to the long lag phase, Run
5 was conducted for a total of 800 h.

83

Figure 9. Cumulative biogas (solid fill) and methane (pattern fill) with increasing FOG addition
normalized to initial TVS loading. One of the triplicates vessels for 40% FOG showed an outlier
in biogas production compared to the other replicates and was excluded from the mean analysis
in this figure. Error bars for cumulative biogas/methane production indicate the combined
standard deviation for gas production and initial organic loading in g TVS for triplicate vessels
for each substrate mixture. The error bars for relative mcrA gene expression represent the
standard deviation of the ratio of triplicate RT-qPCR reactions.

High concentrations of VFAs at 50% and 60% FOG suggested inhibition of methanogens and
syntrophic fatty-acid oxidizers. Acetate concentrations at 50% and 60% FOG were 353 ± 166 and
711 ± 227 mg L
-1
, respectively (Table 8). Propionate concentrations were 650 ± 13 mg L
-1
and 600
± 60 mg L
-1
for 50% and 60% FOG, respectively. Despite high VFAs, pH remained above 6.9 due
to adequate carbonate buffering. Direct inhibition by VFAs was unlikely as other studies observed

84

inhibition at significantly higher concentrations (Franke-Whittle et al. 2014). Therefore, VFA
degrading populations were likely inhibited by LCFA or other components of FOG.
Table 8. TVS removal, free ammonia concentration, pH, VFA concentration, sulfate
concentration, and sulfate reduction for end of Run 5. The detection limit for the IC analyses
was 10 mg L
-1
. The row marked 40% FOG* shows the results of the outlier sample in the 40%
FOG condition.

Using analysis of molecular variance (AMOVA), a significant variation (p=0.027) in microbial
community profiles was observed between substrate mixtures with high and low methane
production. A non-metric multidimensional scaling (NMDS) plot of microbial community
structure for 50% and 60% FOG clustered together (Figure S18). Given that all substrate mixtures
for Run 5 were seeded from Run 4 30% FOG+FW, similar microbial activities were observed at
the beginning of Run 5 (Figure 10). End of run samples from duplicate vessels submitted for RNA-
based sequencing in Run 5 showed reproducibility of relative activity profiles, except at 40% FOG
which also had inconsistent biogas production. Genus-level classification showed that
unclassified Clostridiales, unclassified Thermotogales, Anaerobaculum, and Syntrophomonas
were the most active OTUs at the lowest FOG loading. Coprothermobacter was also highly active
at 30% and 40% FOG with high methane production, but relatively inactive at 50% and 60% FOG.
Coprothermobacter, proteolytic anaerobic fermenters that have an optimum growth rate at 55
o
C
(Etchebehere et al. 1998), were positively correlated (R=0.890, p=0.003) with methane

TVS
reduction
(%)
free
ammonia
(mg/L)
pH Acetate
(mg/L)
Propionate
(mg/L)
Butyrate
(mg/L)
Formate
(mg/L)
Valerate
(mg/L)
Sulfate
(mg/L)
Sulfate
reduction (%)
Control 13.4 ± 4.4 93.2 ± 24.9 7.70 ± 0.1 < 10 160 ± 59 < 10 < 10 < 10 28.8 ± 0.6 87.5 ± 0.279
30%FOG 51.8 ± 10.9 134 ± 50 7.70 ± 0.1 < 10 < 10 < 10 < 10 < 10 40.0 ± 18.4 76.5 ± 9.91
40%FOG 53.2 ± 10.8 132 ± 2 7.50 ± 0.3 < 10 < 10 < 10 < 10 < 10 75.7 ± 34.3 49.3 ± 27.2
50% FOG 57.7 ± 4.3 33.5 ± 7.0 7.10 ± 0.1 353 ± 166 650 ± 13 < 10 121 ± 10.8 92.0 ± 4.10 14.5 ± 0.3 88.7 ± 0.2
60% FOG 60.8 ± 0.3 22.2 ± 12.3 6.90 ± 0.2 717 ± 227 600 ± 60 < 10 139 ± 4.58 103 ± 1.60 14.3 ± 1.8 88.4 ± 1.7
40%FOG* 54.7 41.1 ± 6.3 7.13 325 ± 10.8 660 ± 2 < 10 60.0 ± 65.7 80.1 ± 0.2 15.0 ± 0.03 90.0 ± 0.02

85

production (Table 9). Thermotogales followed a similar trend with high relative activity at 30%
and 40% FOG and no activity at 50% and 60% FOG.
Several populations significantly increased in relative activity at high FOG OLR. For example,
Lactobacillus and Tepidimicrobium had high relative activity at 50% and 60% FOG. Further, these
populations were highly active at 40% FOG when biogas production was low (outlier in data set)
but relatively inactive when biogas production was high (Figure 9 and Figure 10). Spearman rank
analysis showed a negative correlation between methane production and relative activity of
Lactobacillus (R=-0.762, p=0.028). Lactobacillus have been shown to dominate in AD with FW
(Shin et al. 2010), and also comprised 77.9% of the relative abundance and 86.0% of the relative
activity in the raw FW samples sequenced in our study (Figure S5). Tepidimicrobium spp., protein
and amino acid degraders (Tang et al. 2011), may have been active at high FOG OLR because
they had less competition from other populations inhibited by high LCFA. A previous study found
a positive correlation between acetate and propionate concentrations and the presence of
Tepidimicrobium (Regueiro et al. 2015), corroborating the observations here. LCFA accumulation
likely occurred at 50% and 60% FOG because syntrophic-fatty acid oxidizers did not sufficiently
increase activity, thereby resulting in inhibition of methanogens, similar to what was observed in
a previous study (Ziels et al. 2016). The mechanism of LCFA toxicity on syntrophic-fatty acid
oxidizers (e.g., biochemical inhibition or mass transfer limitations leading to decrease of cell
permeability (Nobu et al. 2015)), requires further investigation.

86

Figure 10. Top 30 most active OTUs classified at the genus level for Run 5 (FOG inhibition study)
at the beginning and end of run. Duplicate results are shown for the end of the run to represent
methodological precision. 40% FOG(1)* shows results of a vessel with low biogas production.
All data are expressed as a percentage normalized using total 16S rRNA sequences (Bacteria
and Archaea).

A significant positive correlation between the sum of the relative activity of syntrophic fatty-acid
oxidizers and methane production was observed (R=0.909 p=0.0017;Table 9). Syntrophic fatty-
acid oxidizers showed a remarkably high increase in relative activity at both 30% and 40% FOG
(Figure S19). These populations increased from 0.629% in the beginning of Run 5 to 19.5% and
21.3% in the two duplicate samples for 30% FOG at the end of Run 5 (Figure 11B).
Syntrophomonas and Syntrophothermus comprised the majority of this syntrophic relative
activity. Syntrophomonas likely prevented the accumulation of LCFA at high FOG OLR. It has been

87

previously proposed that prediction of FOG-loading capacity of a digester is possible by
monitoring abundance of Syntrophomonas (Ziels et al. 2016) and that higher LCFA degradation
rates were achieved with increased abundance of Syntrophomonas (Ziels et al. 2017).
Syntrophomonas and Syntrophothermus showed significant positive correlation with methane
production, both with R values of 0.792 (Table 9). Syntrophic fatty-acid oxidizers increased from
0.467% at the beginning of Run 5 in biogas producing 40% FOG to 16.9% at the end of Run 5. The
40% FOG with low biogas production only showed an increase in relative activity of these
populations to 5.83% at the end of Run 5, suggesting their importance. In sharp contrast, the
relative activity of syntrophic fatty-acid oxidizers at 50% FOG was only 3.65% and 2.25% and even
lower at 60% FOG, 1.3% and 1.47%. Further, neither Syntrophomonas nor Syntrophothermus
were active at high FOG. The majority of syntrophic relative activity at 50% and 60% FOG was
Tepidanaerobacter, a syntrophic acetate oxidizer. The significant increase in relative activity of
key syntrophs at 30% and 40% FOG suggests that high FOG loading could be an important
bioaugmentation strategy to boost these populations in AD. A study that used a CSTR to study
the impact of FOG co-digestion reported up to 52% FOG (w/w) addition was not inhibitory
throughout 198 days of operation (Ziels et al. 2016). Although Run 5 in our study was conducted
for a shorter period (33 days), we reseeded biomass from the preceding run that had already
acclimated to the high FOG addition. Therefore, reseeding likely compensated for the long lag
phase needed at high concentrations of FOG. Notably, our study run time was similar to other
batch reactors that investigated the impact of high FOG addition (Li et al. 2011, Martínez et al.
2011, Xu et al. 2015). Further, the use of RNA-based methods also added sensitivity relative to
other studies that used DNA-based method (Ziels et al. 2016) to evaluate changes in the microbial

88

community. Nevertheless, longer retention times may have eventually led to recovery in 50%
FOG and 60% FOG vessels.

Figure 11. (A) Relative activity of methanogens identified at the genus level based on 16S rRNA
sequencing and (B) relative activity of syntrophic fatty-acid oxidizers identified at the genus
level based on 16S rRNA sequencing. Results are expressed as percentages normalized to the
total 16S rRNA sequences (Bacteria and Archaea). Duplicate results are shown for the end of
the run to represent methodological precision. 40% FOG(1)* shows results of a vessel with low
biogas production. Truncated y-axes (0 to 0.9% and 0 to 25% on figure A and B, respectively)
are shown to accentuate differences in abundance.

Relative activity of methanogens decreased significantly at high FOG OLR (Figure 11A). However,
Methanoculleus increased in relative activity suggesting more resilience or adaptation potential
to high FOG. The only methanogen that showed significant positive correlation with methane
production in the FOG inhibition study was Methanosarcina (R=0.909, p=0.0017) (Table 9). RT-
B.

A.

89

qPCR targeting methanogen activity via mcrA gene expression indicated a steep decline in mcrA
gene expression normalized to 16S rRNA at high FOG OLR, with total inhibition at 50% FOG (Figure
9). A previous study demonstrated that high LCFA can impact the membrane integrity of
methanogens (Sousa et al. 2013), and this could have been the mechanism of inhibition in our
study. RT-qPCR showed similar trends to relative activity inferred from 16S rRNA sequencing and
thus confirmed that 16S rRNA sequencing is a useful tool in profiling active methanogens.
The majority of previous studies evaluating FOG co-digestion have been at mesophilic
temperatures (Long et al. 2012, Silvestre et al. 2011, Ziels et al. 2016), although thermophilic
temperature has been reported to increase LCFA degradation, increase reaction rates, and
promote faster liquid-solid separation (Yenigün and Demirel 2013). The benefits of operating at
thermophilic temperature compared to the increased energy demand for heating should be
taken into consideration in full-scale applications (Long et al. 2012). The OLR of FOG addition
without incurring inhibition has been a subject of various studies, with differing reports
(Alqaralleh et al. 2016, Cirne et al. 2007, Rasit et al. 2015). For example, a study that conducted
thermophilic and two stage hyper-thermophilic reactors with co-digestion of TWAS and FOG
reported inhibition at 80% FOG addition (TVS) (Alqaralleh et al. 2016). Some studies have
suggested the addition of lipase-enzyme (Donoso-Bravo and Fdz-Polanco 2013) and two-stage
reactors (Xu et al. 2015) to mitigate inhibition. Mitigation strategies should be further evaluated
by including microbial community analyses to characterize the impact on active microbial
populations, particularly on syntrophic bacteria that were highly impacted by high FOG addition
in this study.

90

Table 9. Spearman rank correlation in Run 5 for methane production and microbial activity of
prominent groups normalized to total 16S rRNA.
Group R p Group R p
Planctomycetaceae 0.943 0.0004 Sporosarcina -0.962 0.0001
Methanosarcina 0.909 0.002 Tissierella -0.866 0.005
Coprothermobacter 0.890 0.003 Clostridium -0.847 0.008
Syntrophomonadaceae 0.875 0.004 Porphyromonadaceae -0.837 0.010
Thermogymnomonas 0.833 0.010 Clostridium XI -0.812 0.014
Syntrophomonas 0.792 0.019 Turicibacter -0.800 0.017
Syntrophothermus 0.792 0.019 Ruminococcus -0.769 0.026
Clostridiales 0.788 0.020 Lactobacillus -0.762 0.028
Petrimonas 0.752 0.032 Tepidiphilus -0.762 0.028
Syntrophaceticus 0.749 0.032

unclassified Bacteria 0.738 0.037

Total Syntrophs 0.909 0.002

Total Methanogens 0.738 0.037

3.5 Conclusions
The impact of co-digestion of FW and FOG on performance and microbial community dynamics
during bench-scale respirometry was evaluated using high-throughput DNA and RNA-based
sequencing and RT-qPCR. The following conclusions were made based on observations during
the study:
 Co-digestion of FOG and FW synergistically increased performance by enhancing TVS
removal and methane production.
 FOG addition increased lag-phase significantly, which was reduced in later runs due to
microbial community adaptation.
 Syntrophic fatty-acid oxidizers, such as Syntrophomonas, were enriched during co-
digestion and were highly linked with increases in methane production, particularly at
high FOG addition.

91

 30% VSL FOG addition was optimal, where methane production was >50% that of
PS+TWAS likely due to a 34 fold increase in relative activity of syntrophic fatty-acid
oxidizers. However, FOG addition resulted in unstable performance and extreme
inhibition at 50-60% VSL.
 Overall, AD performance in this study was better linked with syntrophic fatty-acid
oxidizers than methanogens, suggesting the importance of evaluating these populations
in full-scale reactors that perform co-digestion with FOG.

Acknowledgements
The authors wish to thank Judy Opp for help with Illumina sequencing. We would like to thank
LA Sanitation and Divert, Inc. for their participation in the study. YMA was supported by a
Provost Fellowship from the University of Southern California.

92

3.6 Literature Cited
Alqaralleh, R.M., Kennedy, K., Delatolla, R. and Sartaj, M. (2016) Thermophilic and hyper-thermophilic
co-digestion of waste activated sludge and fat, oil and grease: Evaluating and modeling methane
production. Journal of environmental management 183, 551-561.
Amha, Y., Gregori, M., Johnson, B., Palacios, R.N. and Smith, A.L. (2015a) Bench-and Full-Scale Anaerobic
Co-Digestion of Fats Oil and Grease, Food Waste, and Vegetable Cooking Oil for Enhanced Biogas
Production. Proceedings of the Water Environment Federation 2015(15), 5304-5311.
Amha, Y.M., Kumaraswamy, R. and Ahmad, F. (2015b) A probabilistic QMRA of Salmonella in direct
agricultural reuse of treated municipal wastewater. Water Science and Technology 71(8), 1203-1211.
Becker, A.M., Yu, K., Stadler, L.B. and Smith, A.L. (2017) Co-management of domestic wastewater and
food waste: A life cycle comparison of alternative food waste diversion strategies. Bioresource
Technology 223, 131-140.
Blazewicz, S.J., Barnard, R.L., Daly, R.A. and Firestone, M.K. (2013) Evaluating rRNA as an indicator of
microbial activity in environmental communities: limitations and uses. The ISME journal 7(11), 2061-
2068.
Cirne, D., Paloumet, X., Björnsson, L., Alves, M. and Mattiasson, B. (2007) Anaerobic digestion of lipid-
rich waste—effects of lipid concentration. Renewable Energy 32(6), 965-975.
Conklin, A., Stensel, H.D. and Ferguson, J. (2006) Growth kinetics and competition between
Methanosarcina and Methanosaeta in mesophilic anaerobic digestion. Water Environment Research
78(5), 486-496.
Dai, X., Duan, N., Dong, B. and Dai, L. (2013) High-solids anaerobic co-digestion of sewage sludge and
food waste in comparison with mono digestions: Stability and performance. Waste Management 33(2),
308-316.
Davidsson, Å., Lövstedt, C., la Cour Jansen, J., Gruvberger, C. and Aspegren, H. (2008) Co-digestion of
grease trap sludge and sewage sludge. Waste Management 28(6), 986-992.
De Vrieze, J., Regueiro, L., Props, R., Vilchez-Vargas, R., Jáuregui, R., Pieper, D.H., Lema, J.M. and
Carballa, M. (2016) Presence does not imply activity: DNA and RNA patterns differ in response to salt
perturbation in anaerobic digestion. Biotechnology for biofuels 9(1), 244.
Donoso-Bravo, A. and Fdz-Polanco, M. (2013) Anaerobic co-digestion of sewage sludge and grease trap:
Assessment of enzyme addition. Process Biochemistry 48(5), 936-940.
Duan, N., Dong, B., Wu, B. and Dai, X. (2012) High-solid anaerobic digestion of sewage sludge under
mesophilic conditions: feasibility study. Bioresource Technology 104, 150-156.
Etchebehere, C., Pavan, M., Zorzopulos, J., Soubes, M. and Muxi, L. (1998) Coprothermobacter platensis
sp. nov., a new anaerobic proteolytic thermophilic bacterium isolated from an anaerobic mesophilic
sludge. International Journal of Systematic and Evolutionary Microbiology 48(4), 1297-1304.
Franke-Whittle, I.H., Walter, A., Ebner, C. and Insam, H. (2014) Investigation into the effect of high
concentrations of volatile fatty acids in anaerobic digestion on methanogenic communities. Waste
Manag 34, 2080-2089.
Girault, R., Bridoux, G., Nauleau, F., Poullain, C., Buffet, J., Peu, P., Sadowski, A.G. and Béline, F. (2012)
Anaerobic co-digestion of waste activated sludge and greasy sludge from flotation process: Batch versus
CSTR experiments to investigate optimal design. Bioresource Technology 105, 1-8.
Hansen, K.H., Angelidaki, I. and Ahring, B.K. (1998) Anaerobic digestion of swine manure: inhibition by
ammonia. Water research 32(1), 5-12.
Hao, L.-P., Lü, F., He, P. -J., Li, L. and Shao, L.-M. (2010) Predominant contribution of syntrophic acetate
oxidation to thermophilic methane formation at high acetate concentrations. Environmental science &
technology 45(2), 508-513.

93

Hatamoto, M., Imachi, H., Yashiro, Y., Ohashi, A. and Harada, H. (2007) Diversity of Anaerobic
Microorganisms Involved in Long-Chain Fatty Acid Degradation in Methanogenic Sludges as Revealed by
RNA-Based Stable Isotope Probing. Applied and environmental microbiology 73(13), 4119-4127.
Hattori, S. (2008) Syntrophic acetate-oxidizing microbes in methanogenic environments. Microbes and
Environments 23(2), 118-127.
Kabouris, J.C., Tezel, U., Pavlostathis, S.G., Engelmann, M., Dulaney, J.A., Todd, A.C. and Gillette, R.A.
(2009) Mesophilic and thermophilic anaerobic digestion of municipal sludge and fat, oil, and grease.
Water Environment Research 81(5), 476-485.
Karakashev, D., Batstone, D.J. and Angelidaki, I. (2005) Influence of environmental conditions on
methanogenic compositions in anaerobic biogas reactors. Applied and environmental microbiology
71(1), 331-338.
Klappenbach, J.A., Dunbar, J.M. and Schmidt, T.M. (2000) rRNA operon copy number reflects ecological
strategies of bacteria. Applied and environmental microbiology 66(4), 1328-1333.
Koch, K., Helmreich, B. and Drewes, J.E. (2015) Co-digestion of food waste in municipal wastewater
treatment plants: Effect of different mixtures on methane yield and hydrolysis rate constant. Applied
Energy 137, 250-255.
Koch, K., Plabst, M., Schmidt, A., Helmreich, B. and Drewes, J.E. (2016) Co-digestion of food waste in a
municipal wastewater treatment plant: Comparison of batch tests and full-scale experiences. Waste
Management 47, Part A, 28-33.
Kong, D., Shan, J., Iacoboni, M. and Maguin, S.R. (2012) Evaluating greenhouse gas impacts of organic
waste management options using life cycle assessment. Waste management & research,
0734242X12440479.
Kumaraswamy, R., Amha, Y.M., Anwar, M.Z., Henschel, A., Rodríguez, J. and Ahmad, F. (2014) Molecular
analysis for screening human bacterial pathogens in municipal wastewater treatment and reuse.
Environmental science & technology 48(19), 11610-11619.
Lee, S.-H., Park, J.-H., Kim, S.-H., Yu, B.J., Yoon, J.-J. and Park, H.-D. (2015) Evidence of syntrophic acetate
oxidation by Spirochaetes during anaerobic methane production. Bioresource Technology 190, 543-549.
Leibrock, A. (2017) Are Food Waste Bans Working?
Leinonen, R., Sugawara, H. and Shumway, M. (2010) The sequence read archive. Nucleic acids research,
gkq1019.
Li, C., Champagne, P. and Anderson, B.C. (2011) Evaluating and modeling biogas production from
municipal fat, oil, and grease and synthetic kitchen waste in anaerobic co-digestions. Bioresource
Technology 102(20), 9471-9480.
Li, L., He, Q., Ma, Y., Wang, X. and Peng, X. (2016) A mesophilic anaerobic digester for treating food
waste: process stability and microbial community analysis using pyrosequencing. Microbial cell factories
15(1), 65.
Long, J.H., Aziz, T.N., Francis, L. and Ducoste, J.J. (2012) Anaerobic co-digestion of fat, oil, and grease
(FOG): a review of gas production and process limitations. Process Safety and Environmental Protection
90(3), 231-245.
Martín-González, L., Castro, R., Pereira, M.A., Alves, M.M., Font, X. and Vicent, T. (2011) Thermophilic
co-digestion of organic fraction of municipal solid wastes with FOG wastes from a sewage treatment
plant: reactor performance and microbial community monitoring. Bioresource Technology 102(7), 4734-
4741.
Martínez, E.J., Redondas, V., Fierro, J., Gómez, X. and Morán, A. (2011) Anaerobic digestion of high lipid
content wastes: FOG co-digestion and milk processing fat digestion. Journal of Residuals Science &
Technology 8(2).
Matteson, G.C. and Jenkins, B. (2007) Food and processing residues in California: Resource assessment
and potential for power generation. Bioresource Technology 98(16), 3098-3105.

94

Mladenovska, Z. and Ahring, B.K. (1997) Mixotrophic growth of two thermophilic Methanosarcina
strains, Methanosarcina thermophila TM-1 and Methanosarcina sp. SO-2P, on methanol and
hydrogen/carbon dioxide. Applied Microbiology and Biotechnology 48(3), 385-388.
Morris, R., Schauer-Gimenez, A., Bhattad, U., Kearney, C., Struble, C.A., Zitomer, D. and Maki, J.S. (2014)
Methyl coenzyme M reductase (mcrA) gene abundance correlates with activity measurements of
methanogenic H2/CO2-enriched anaerobic biomass. Microbial biotechnology 7(1), 77-84.
Mosbæk, F., Kjeldal, H., Mulat, D.G., Albertsen, M., Ward, A.J., Feilberg, A. and Nielsen, J.L. (2016)
Identification of syntrophic acetate-oxidizing bacteria in anaerobic digesters by combined protein-based
stable isotope probing and metagenomics. The ISME journal.
Narihiro, T., Nobu, M.K., Kim, N.K., Kamagata, Y. and Liu, W.T. (2015) The nexus of syntrophy-associated
microbiota in anaerobic digestion revealed by long-term enrichment and community survey.
Environmental Microbiology 17(5), 1707-1720.
Nobu, M.K., Narihiro, T., Rinke, C., Kamagata, Y., Tringe, S.G., Woyke, T. and Liu, W.-T. (2015) Microbial
dark matter ecogenomics reveals complex synergistic networks in a methanogenic bioreactor. The ISME
journal 9(8), 1710-1722.
Noutsopoulos, C., Mamais, D., Antoniou, K., Avramides, C., Oikonomopoulos, P. and Fountoulakis, I.
(2013) Anaerobic co-digestion of grease sludge and sewage sludge: The effect of organic loading and
grease sludge content. Bioresource Technology 131, 452-459.
Palatsi, J., Illa, J., Prenafeta-Boldú, F., Laureni, M., Fernandez, B., Angelidaki, I. and Flotats, X. (2010)
Long-chain fatty acids inhibition and adaptation process in anaerobic thermophilic digestion: batch tests,
microbial community structure and mathematical modelling. Bioresource Technology 101(7), 2243-
2251.
Platt, B., Goldstein, N., Coker, C. and Brown, S. (2014) State of Composting in the US. Institute for Local
Self-Reliance.
Props, R., Kerckhof, F.-M., Rubbens, P., De Vrieze, J., Sanabria, E.H., Waegeman, W., Monsieurs, P.,
Hammes, F. and Boon, N. (2016) Absolute quantification of microbial taxon abundances. The ISME
journal.
Rasit, N., Idris, A., Harun, R. and Ghani, W.A.W.A.K. (2015) Effects of lipid inhibition on biogas production
of anaerobic digestion from oily effluents and sludges: An overview. Renewable and Sustainable Energy
Reviews 45, 351-358.
Regueiro, L., Lema, J.M. and Carballa, M. (2015) Key microbial communities steering the functioning of
anaerobic digesters during hydraulic and organic overloading shocks. Bioresource Technology 197, 208-
216.
Salter, S.J., Cox, M.J., Turek, E.M., Calus, S.T., Cookson, W.O., Moffatt, M.F., Turner, P., Parkhill, J.,
Loman, N.J. and Walker, A.W. (2014) Reagent and laboratory contamination can critically impact
sequence-based microbiome analyses. BMC biology 12(1), 1.
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley,
B.B., Parks, D.H. and Robinson, C.J. (2009) Introducing mothur: open-source, platform-independent,
community-supported software for describing and comparing microbial communities. Applied and
environmental microbiology 75(23), 7537-7541.
Schnürer, A., Schink, B. and Svensson, B.H. (1996) Clostridium ultunense sp. nov., a mesophilic
bacterium oxidizing acetate in syntrophic association with a hydrogenotrophic methanogenic bacterium.
International Journal of Systematic and Evolutionary Microbiology 46(4), 1145-1152.
Sekiguchi, Y., Kamagata, Y., Nakamura, K., Ohashi, A. and Harada, H. (2000) Syntrophothermus
lipocalidus gen. nov., sp. nov., a novel thermophilic, syntrophic, fatty-acid-oxidizing anaerobe which
utilizes isobutyrate. International Journal of Systematic and Evolutionary Microbiology 50(2), 771-779.
Shin, S.G., Han, G., Lim, J., Lee, C. and Hwang, S. (2010) A comprehensive microbial insight into two-
stage anaerobic digestion of food waste-recycling wastewater. Water research 44(17), 4838-4849.

95

Silva, S.A., Cavaleiro, A.J., Pereira, M.A., Stams, A.J.M., Alves, M.M. and Sousa, D.Z. (2014) Long-term
acclimation of anaerobic sludges for high-rate methanogenesis from LCFA. biomass and bioenergy 67,
297-303.
Silvestre, G., Rodríguez-Abalde, A., Fernández, B., Flotats, X. and Bonmatí, A. (2011) Biomass adaptation
over anaerobic co-digestion of sewage sludge and trapped grease waste. Bioresource Technology
102(13), 6830-6836.
Smith, A.L., Skerlos, S.J. and Raskin, L. (2015) Membrane biofilm development improves COD removal in
anaerobic membrane bioreactor wastewater treatment. Microbial biotechnology 8(5), 883-894.
Sousa, D.Z., Salvador, A.F., Ramos, J., Guedes, A.P., Barbosa, S., Stams, A.J.M., Alves, M.M. and Pereira,
M.A. (2013) Activity and viability of methanogens in anaerobic digestion of unsaturated and saturated
long-chain fatty acids. Applied and environmental microbiology 79(14), 4239-4245.
Sousa, D.Z., Smidt, H., Alves, M.M. and Stams, A.J.M. (2009) Ecophysiology of syntrophic communities
that degrade saturated and unsaturated long-chain fatty acids. FEMS microbiology ecology 68(3), 257-
272.
Sundberg, C., Al-Soud, W.A., Larsson, M., Alm, E., Yekta, S.S., Svensson, B.H., Sørensen, S.J. and Karlsson,
A. (2013) 454 pyrosequencing analyses of bacterial and archaeal richness in 21 full-scale biogas
digesters. FEMS microbiology ecology 85(3), 612-626.
Tang, Y.-Q., Ji, P., Hayashi, J., Koike, Y., Wu, X.-L. and Kida, K. (2011) Characteristic microbial community
of a dry thermophilic methanogenic digester: its long-term stability and change with feeding. Applied
Microbiology and Biotechnology 91(5), 1447-1461.
Wang, L., Aziz, T.N. and Francis, L. (2013) Determining the limits of anaerobic co-digestion of thickened
waste activated sludge with grease interceptor waste. Water research 47(11), 3835-3844.
Welander, P.V. and Metcalf, W.W. (2005) Loss of the mtr operon in Methanosarcina blocks growth on
methanol, but not methanogenesis, and reveals an unknown methanogenic pathway. Proceedings of
the National Academy of Sciences of the United States of America 102(30), 10664-10669.
Werner, J.J., Garcia, M.L., Perkins, S.D., Yarasheski, K.E., Smith, S.R., Muegge, B.D., Stadermann, F.J.,
DeRito, C.M., Floss, C. and Madsen, E.L. (2014) Microbial community dynamics and stability during an
ammonia-induced shift to syntrophic acetate oxidation. Applied and environmental microbiology 80(11),
3375-3383.
Westerholm, M., Dolfing, J., Sherry, A., Gray, N.D., Head, I.M. and Schnürer, A. (2011) Quantification of
syntrophic acetate-oxidizing microbial communities in biogas processes. Environmental microbiology
reports 3(4), 500-505.
Wu, L.-J., Kobayashi, T., Kuramochi, H., Li, Y.-Y. and Xu, K.-Q. (2016) Improved biogas production from
food waste by co-digestion with de-oiled grease trap waste. Bioresource Technology 201, 237-244.
Xu, R., Yang, Z., Chen, T., Zhao, L., Huang, J., Xu, H., Song, P. and Li, M. (2015) Anaerobic co-digestion of
municipal wastewater sludge with food waste with different fat, oil, and grease contents: study of
reactor performance and extracellular polymeric substances. RSC Advances 5(125), 103547-103556.
Yang, Z.-H., Xu, R., Zheng, Y., Chen, T., Zhao, L.-J. and Li, M. (2016) Characterization of extracellular
polymeric substances and microbial diversity in anaerobic co-digestion reactor treated sewage sludge
with fat, oil, grease. Bioresource Technology 212, 164-173.
Yenigün, O. and Demirel, B. (2013) Ammonia inhibition in anaerobic digestion: a review. Process
Biochemistry 48(5), 901-911.
Zhu, X., De Francisci, D., Treu, L., Kougias, P., Campanaro, S. and Angelidaki, I. (2015) Metagenomic
analysis on thermophilic biogas reactors fed with high load of Long Chain Fatty Acids (LCFA). Proceedings
of the 14th World Congress on Anaerobic Digestion.
Ziels, R.M., Beck, D.A.C., Martí, M., Gough, H.L., Stensel, H.D. and Svensson, B.H. (2015) Monitoring the
dynamics of syntrophic β-oxidizing bacteria during anaerobic degradation of oleic acid by quantitative
PCR. FEMS microbiology ecology 91(4), fiv028.

96

Ziels, R.M., Beck, D.A.C. and Stensel, H.D. (2017) Long-chain fatty acid feeding frequency in anaerobic
codigestion impacts syntrophic community structure and biokinetics. Water research 117, 218-229.
Ziels, R.M., Karlsson, A., Beck, D.A.C., Ejlertsson, J., Yekta, S.S., Bjorn, A., Stensel, H.D. and Svensson, B.H.
(2016) Microbial community adaptation influences long-chain fatty acid conversion during anaerobic
codigestion of fats, oils, and grease with municipal sludge. Water research 103, 372-382.

97

CHAPTER 4

4. Two-phase improves performance of anaerobic membrane
bioreactor treatment of food waste at high organic loading rates

4.1 Abstract
Anaerobic membrane bioreactors (AnMBRs) are in use at the full-scale for energy recovery from
food waste (FW). In this study, the potential for two-phase (acid/gas) AnMBR treatment of FW
was investigated as a strategy to increase microbial diversity, thereby improving performance.
Two bench-scale AnMBRs were operated in single-phase (SP) and two-phase (TP) mode across
incremental increases in organic loading rate (OLR) from 2.5 to 15 g total chemical oxygen
demand (COD) L·d
-1
. The TP acid-phase (TP-AP) enriched total VFAs by 3-fold compared to
influent FW and harbored a distinct microbial community enriched in fermenters that thrived in
the low pH environment. The TP methane phase (TP-MP) showed increased methane production
and resilience relative to SP as OLR increased from 3.5 to 10 g COD L·d
-1
. SP showed signs of
inhibition (i.e., rapid decrease in methane production per OLR) at 10 g COD L·d
-1
, whereas both
systems were inhibited at 15 g COD L·d
-1
. At 10 g COD L·d
-1
, where the highest difference in
performance was observed (20.3% increase in methane production), activity of syntrophic

98

bacteria in TP-MP was double that of SP. Our results indicate that AnMBRs in TP mode could
effectively treat FW at OLRs up to 10 g COD∙L day
-1
by improving hydrolysis rates (and VFA
production), microbial diversity, syntroph activity, and enriching resistant communities to high
OLRs relative to AnMBRs in SP mode.
4.2 Introduction
As landfills rapidly reach capacity in the US and elsewhere, diversion of organic wastes is expected
to become the norm. Anaerobic membrane bioreactors (AnMBRs), which combine anaerobic
treatment with membrane separation, have emerged as a sustainable food waste (FW)
management strategy with reduced environmental footprint relative to landfilling and
composting, while also providing energy recovery via biogas production (Becker et al. 2017).
Compared to anaerobic digesters, AnMBRs enable operation at longer sludge retention time
(SRT) and improve effluent quality via membrane separation. Effluent quality is of particular
concern for decentralized FW treatment, as is done by Divert, Inc. at their full-scale facilities in
California and Massachusetts, as it directly impacts quality surcharge fees from utilities receiving
effluent flows. The upper limits of organic loading rate (OLR) is another critical parameter that
dictates system capacity and reactor dimensions. Optimizing performance of decentralized FW
management strategies, such as AnMBRs, is needed to increase treatment capacity, reduce
environmental impacts, and improve economic benefits of resource recovery.
Anaerobic treatment of FW can be accomplished in single-phase (SP) or two-phase (TP)
systems. TP (acid/gas) anaerobic digestion separates the microbial conversion to an acid-phase
(TP-AP) and methane-phase (TP-MP) digester operated in series. The TP-AP provides a unique
biochemical environment supporting the growth of fermenting bacteria, while methanogenic

99

archaea are suppressed by the low pH and short SRT (typically less than 3 days). The TP-AP
digester also acts as an equalization tank, protecting the TP-MP from pulses in organic loading or
potential inhibitors that could result in performance instability (Smith et al. 2017). Relative to
anaerobic digestion of wastewater sludges, decentralized FW management systems experience
greater temporal variability in feedstock quality and strength. Therefore, TP could provide
greater benefit in this waste management scenario relative to SP. The TP-MP digester is operated
near neutral pH at a longer SRT to enrich for methanogenic archaea and other microbial
populations with low growth rates. Overall, TP has been reported to allow for higher OLRs and
can improve biogas production and quality relative to SP(Bhattacharya et al. 1996, Bowles et al.
2012).
To date, very few studies have investigated AnMBR treatment of only FW (Cheng et al. 2018,
Xiao et al. 2015), and no research has evaluated microbial community dynamics within these
systems. It is important to note that co-management of food waste and domestic wastewater via
AnMBRs has received attention in the literature recently(Becker et al. 2017, Cho et al. 2018,
Jeong et al. 2017, Zamorano-López et al. 2018), but is not the focus of this research. Further, few
studies have systematically compared performance of SP versus TP (Ganesh et al. 2014, Jo et al.
2018, Shen et al. 2013, Voelklein et al. 2016) or evaluated the impact of phase separation on the
microbial community (Shin et al. 2010, Wu et al. 2016) during FW treatment via anaerobic
digestion (without membrane separation), and there is no prior work investigating TP AnMBR
treatment of FW. Although a consensus regarding performance benefits of TP versus SP digestion
of FW is lacking (Ganesh et al. 2014, Jo et al. 2018, Sahu et al. 2017, Shen et al. 2013, Voelklein
et al. 2016), studies have attributed optimized hydrolysis rates and minimization of

100

ammoniaaccumulation (Sahu et al. 2017, Voelklein et al. 2016) as key differentiators that
resulted in increased methane production in TP. Although TP has been implemented at the full-
scale for primary sludge and waste activated sludge management (Smith et al. 2017), to our
knowledge, no full-scale TP systems are in operation for FW management. Therefore, in this work
we aimed to systematically compare SP and TP management of FW using bench-scale AnMBRs
equipped with flat-sheet ceramic membranes. Both systems were operated at incremental OLRs
up to 15 g COD L·day
-1
to compare performance and operating limits. DNA- and RNA-based high-
throughput sequencing were used to evaluate differences in microbial community structure and
activity, respectively, between the SP and TP systems across the range of applied OLRs.
4.3 Materials & Methods
4.3.1 Bench-scale AnMBR configuration
Two replicate jacketed 7 L reactors (Chemglass, NJ) with 5.2 L liquid volume were operated for
approximately 300 days (Figure S20). Each system contained a submerged flat-sheet ceramic
microfiltration membrane with pore size of 0.1 µm (Cembrane, Denmark) and a total effective
membrane area of 0.0113 m
2
. Both AnMBRs were mixed continuously at 250 rpm with an
impellor located near the bottom of the reactor vessel. The jacketed reactors were connected to
a recirculating water bath (Fisher Scientific, Hampton, NH) for temperature control to 37
o
C with
reactor temperature monitored via a probe submerged in the biomass. Pressure in the headspace
and permeate lines were monitored using pressure transducers (Transducers Direct, Cincinnati,
OH). Influent and permeate were pumped using peristaltic pumps to maintain a constant liquid
volume (NewEra, Farmingdale, NY and Langer, Boonton, NJ, respectively). Periodic backwashing
(2 minutes for every 10 minutes) and continuous biogas sparging (at a flow rate of 1 m
3
m
-2
·h
-1
)

101

were employed to manage membrane fouling, where flux ranged from 0.8 - 1.8 LMH for the
different feeding rates. For sparging, a mini diaphragm pump (Parker, North Carolina)
recirculated produced biogas through sparging tubes transversely mounted below the
membrane housing. The headspace was connected to a GFM mass flow meter (Aalborg, New
York) that continuously measured biogas production. All data acquisition and permeate pump
control was done via LabVIEW (National Instruments, Austin, TX), with data recorded in minute
intervals.
4.3.2. Inoculation and operational parameters
The bench-scale AnMBRs were inoculated with mixed liquor (ML) collected from a full-scale,
mesophilic (37
o
C) AnMBR treating FW. The inocula had a total solids (TS) concentration of 28 g·L
-
1
and a total volatile solids (TVS) concentration of 19 g·L
-1
. A FW slurry was also collected
periodically from Divert, Inc. and filtered using 1 mm aluminum mesh to remove large particles
and prevent clogging of tubing in the bench-scale systems. The FW slurry was a blend of
processed FW collected from grocery stores in Los Angeles and creamery wastewater, and had
an average chemical oxygen demand (COD) of 122 ± 7 g·L
-1
. The pH of the feed FW was 3.5 and
total volatile fatty acid (VFA) concentration was 5.8 ± 1.2 g·L
-1
(SI Table 2).
Initially, both AnMBRs (denoted SP1 and SP2) were operated in SP mode, using replicate
operational conditions. After similar performance was observed in both reactors (<10%
fluctuations in biogas production and COD removal over 47 days after startup), one AnMBR was
transitioned to TP mode by incorporating a TP-AP upstream of the TP-MP AnMBR. The TP-AP was
continuously stirred at 250 rpm and partially submerged in a recirculating water bath to maintain
a mesophilic temperature of 37
o
C. Gas production in the TP-AP was monitored in 10-minute

102

intervals using an MPA-200 Methane Potential Analyzer (Challenge Technology, Springdale, AR).
The hydraulic retention time (HRT) and SRT was 3 days. Next, OLR was incrementally increased
from 2.5 to 3.5, 5, 10, and 15 g COD L·d
-1
. Divert, Inc. expects to decrease their HRT in the future
as additional FW is collected and provided for treatment. Their full-scale SP AnMBR is currently
operated at OLRs between 1 and 5 g COD L·d
-1
. For the lowest OLRs tested, 2.5 and 3.5 g COD
L·day
-1
, FW was diluted to maintain sufficient influent flow rates to prevent operational issues (SI
Table 1).
4.3.3 Chemical assay and sampling
Reactor performance was monitored by evaluating permeate characteristics, including COD,
VFAs, and ammonia concentration. Permeate samples were filtered with 0.2 µm nylon
membrane filters (Whatman, Pittsburgh, PA) to measure soluble constituents (COD, ammonium,
VFAs, etc.). The Nessler-Method(APHA 2005) was used to determine ammonia concentration.
VFAs (formic acid, acetic acid, propionic acid, butyric acid, and valeric acid) and other ions (e.g.,
nitrate and sulfate) were determined using ion chromatography (ICS-2000, Dionex, Sunnyvale,
CA) equipped with a refrigerated autosampler (Thermo Scientific, NY, USA). Chromatographic
separation was achieved using a 2 mm AS-11HC column (Dionex, Sunnyvale, CA). The
composition of biogas was measured using the Trace 1310 GC system (Thermo Scientific, NY)
equipped with a flame ionization detector (FID) using hydrogen as carrier gas, where a TG-BOND
Q 30m x 0.53mm x 20 µm column was used for chromatographic separation. TS and TVS of the
biomass were determined using procedures outlined in Standard Methods (APHA 2005).

103

4.3.4 Microbial community analysis
Biomass samples were collected weekly from SP, TP-AP, and TP-MP. In addition to bench-scale
AnMBR samples, FW and mixed liquor (ML) samples from the full-scale AnMBR were collected
monthly over 6 months for DNA-based analyses and weekly over 1 month for RNA-based
analyses, to understand microbial community structure and activity profiles at the full-scale,
which is operated in SP mode. All samples were centrifuged at 5,000 x g for 5 min at 4
o
C,
decanted, and preserved at -80
o
C until further processing. Approximately 0.2 g of pelletized
biomass was taken for extraction. DNA extraction, RNA extraction, and sequencing were
conducted as detailed in our previous work (Amha et al. 2017b). Briefly, DNA extraction was
conducted using the Maxwell 16 Blood LEV kit according to manufacturer’s instruction (Promega,
Madison, WI), whereas RNA extraction was conducted using Maxwell 16 simplyRNA blood kit. An
additional DNase treatment was conducted using DNA-free™ DNA Removal Kit (Invitrogen,
Carlsbad, CA) to remove DNA contamination from RNA extracts. DNA and RNA quantity was
measured using the Quant-iT™ PicoGreen® dsDNA Assay (Invitrogen, Carlsbad, CA) and Quant-
iT™ RiboGreen® RNA Assay, respectively. Afterwards, reverse transcription to generate single-
stranded complementary DNA (cDNA) from RNA extracts was performed using the GoScript™
Reverse Transcription System according to manufacturer’s instructions (Promega, Madison, WI).
One hundred ng of RNA was taken from each sample for cDNA synthesis. Library preparation and
sequencing were conducted at the University of Michigan via Illumina MiSeq using the MiSeq
Reagent Kit V2 (2x250 bp reads) and sequencing primers described previously (Kozich et al. 2013).
Sequencing results were analyzed using mothur (Schloss et al. 2009) with Silva 132 (Pruesse et
al. 2007) as a reference database for alignment and classification. “Relative abundance” is the

104

percentage of 16S rRNA gene sequences (DNA-based) for a given population out of total 16S
rRNA gene sequences for archaea and bacteria. Likewise, “relative activity” is used to describe
percentage of 16S rRNA sequences (RNA-based) out of total 16S rRNA sequences for archaea and
bacteria. Statistical analyses were conducted using JMP Pro (SAS Institute, North Carolina) and
LEfSe tools (Segata et al. 2011).
4.4 Results and Discussion
4.4.1. TP-AP effectively increased VFAs concentration and enriched a distinct microbial
community from FW
Stable VFA production was achieved in TP-AP, with acetate and propionate constituting the
majority of VFAs detected (Figure S21). Initially, propionate concentration was higher than
acetate, but acetate later emerged as the dominant VFA after 50 days of operation, which
corresponded with the increase of OLR from 2.5 to 3.5 g COD L·day
-1
. Notably, pH in TP-AP
remained relatively constant throughout all OLRs at 3.70 ± 0.39 (Figure S3). We elected to not
adjust operating pH to a neutral pH as has been done previously to increase VFA production
(Chen et al. 2017, Khan et al. 2019). Nonetheless, the total average VFA concentration in the
effluent of the TP-AP was 19.8 ± 6.7 g HAc
-
eq L
-1
, a substantial increase from the VFA
concentration in the FW, 6.3 ± 1.2 g HAc
-
equivalent L
-1
(Table S6).
The microbial community analysis revealed distinct community structures and activity profiles
between FW and TP-AP (Figure S23). Non-metric multidimensional scaling (NMDS) and
subsequent analysis of molecular variance (AMOVA) were used to determine whether the NMDS
clustering was significant, by pooling only FW and TP-AP samples. The difference in microbial
communities between FW and TP-MP were statistically significant (p<0.001), for both RNA- and

105

DNA-based analyses (i.e., 16S rRNA and 16S rRNA genes), as shown from the ANOVA results.
Lactobacillus was the most active genus in both FW and TP-AP, representing 71-85% relative
activity in FW samples and 64-90 % relative activity in TP-AP samples (Figure S24). The dominance
of Lactobacillus, a genus known for fermentative metabolism (Rault et al. 2009), in TP-AP is not
surprising given that this population has been shown to withstand low pH conditions.
Prominent shifts in the microbial activity profile were observed with increasing OLR in the TP-
AP. Aeriscardovia, a population that was not active in FW (0.23 ± 0.14% relative activity), was
highly active at 2.5 g COD L·day
-1
with an average of 15.2% relative activity, decreased to <1.30%
at 3.5-10 g COD L·day
-1
, and increased to 8.38% at 15 g COD L·day
-1
(Figure S24). A previous study
similarly reported a shift from an Aeriscardovia-dominated community to a Lactobacillus-
dominated community at decreased HRT in a TP digester (Shin et al. 2010). The only known
species in the genus, Aeriscardovia aeriphila, have been described as requiring a minimum pH of
4.5 for initial growth (Simpson et al. 2004). Therefore, it is surprising that this genus dominated
in our system at a lower pH of 3.70 ± 0.39. Coincidently, as Aeriscardovia became non-active at
3.5 g COD L·day
-1
, the TP-AP VFA profile transitioned from being propionate-dominated to
acetate-dominated. The low activity of Aeriscardovia at OLRs between 3.5-10 g COD L·day
-1
corresponded with a substantial increase in activity of acetogenic Acetobacter.(IIZUKA and
KOMAGATA 1963) Although commonly labelled as obligate aerobes (Fukaya et al. 1992),
Acetobacter have been documented at high relative abundance during FW fermentation to VFAs
(Cao et al. 2019, Sträuber et al. 2018). Other populations that were present in high relative
activity in FW, such as Pseudomonas and Flavobacterium, continued to show high relative activity
in most TP-AP samples. Methanogens and syntrophs were both effectively inhibited in the TP-AP

106

at <0.07% and <0.024% of relative activity, respectively (Figure S25). Low pH and high VFA
concentrations >5.8 g L
-1
have been shown to completely inhibit methanogens, eventually also
inhibiting the growth of syntrophs that depend on methanogens to maintain low hydrogen partial
pressure (Xu et al. 2014).
4.4.2. TP AnMBR resulted in improved performance relative to SP AnMBR at high OLRs and
>98% COD removal efficiency was achieved
During the first phase of operation where both AnMBRs were operated in SP mode, methane
production was significantly similar between SP1 and SP2 (two-tailed t-test, p=0.71) (Figure 12).
COD removal efficiency during this initial phase was >99% and methane production for SP1 and
SP2 was 4.24 ± 0.38 and 4.20 ± 0.36 L d
-1
, respectively. After this period, SP2 was converted to TP
mode (consisting of TP-AP and TP-MP), while SP1 was maintained in SP mode.
Significant differences in performance were apparent between SP and TP-MP as OLR was
incrementally increased. At 2.5 g COD L·day
-1
both reactors showed similar performance (p=0.85),
whereas at 3.5 g COD L·day
-1
, TP-MP showed significantly higher methane production
(p=0.00034). Similarly, OLRs of 5 and 10 g COD L·day
-1
resulted in significantly higher methane
production in TP-MP than SP (Figure 12). At 10 g COD L·day
-1
, mean methane production in TP-
MP was 20.3 ± 8.3% greater (95% confidence interval (CI) of mean) than SP. At each OLR, high
variability in daily methane production was apparent for both SP and TP-MP (Figure 12).
However, no significant outliers in daily methane production rate were identified (p<0.05) using
Grubbs' test or the extreme studentized deviate (ESD) method. The highest methane per OLR in
SP was observed at 2.5 g COD L·day
-1
, which was 0.32 ± 0.06 L CH 4 g COD
-1
fed. The highest specific
methane yield per OLR in TP was observed at 3.5 g COD L·day
-1
, 0.33 ± 0.02 L CH 4 g COD
-1
fed

107

(Figure 12).

Figure 12. Methane production rate (primary-axis) and methane per OLR (secondary y-axis).
Solid fill shows SP and pattern fill signifies TP-MP. The number after SP or TP-MP shows the
OLRs for each system, for example TP-MP - 2.5 indicates, the mean methane production for the
TP- MP treatment at 2.5 g COD L·d
-1
. SP1 and SP2 are initial runs, where both AnMBRs were run
as SP. The solid error bar indicates 95% confidence interval of the mean, and the circles indicate
standard deviation of mean daily methane production for each OLRs. The symbol * shows that
mean methane production were significantly different between SP and TP-MP for the specified
feeding OLRs, as indicated by <0.05 p values with the two-tailed t-test.

The COD mass balance indicated that greater COD went to biomass growth and VS
accumulation in SP compared to TP-MP (Figure S26), indicating that TP mode improved hydrolysis
of FW. At an OLR of 10 g COD L·day
-1
, the VS concentration in SP reached 43 g L
-1
, but only 31 g L
-
1
for TP-MP (Figure S27). The high VS in SP negatively impacted membrane performance and SRT
was subsequently decreased from 140 days (at 2.5 g COD L·day
-1
) to 32.5 days for SP, and 43.3
days for TP-MP to maintain similar VS concentrations in both systems. Methane production per

108

OLR data indicated inhibition in SP at 10 g COD L·day
-1
, where an abrupt decrease of 26% relative
to 2.5 g COD L·day
-1
was observed. In TP-MP, only an 11% decrease in methane per OLR was
observed. At 15 g COD L·day
-1
both systems showed severe inhibition, with 27% and 28% decline
in methane per OLR relative to 2.5 g COD L·day
-1
(Figure 12). The HRT in the SP and TP-MP ranged
between 24 to 10 days for the different OLRs. We did not take into account the additional HRT
as a result of the TP-AP treatment in TP. We believe this had a relatively small contribution in
impacting performance because of the long total SRT in both systems (Table S5). We did not see
major differences in effluent COD concentrations in both systems throughout the different OLRs.
Therefore, we believe the performance differences were more likely linked to SRT rather than
HRT. Similarly, a previous study reported that COD removal efficiency in AnMBR during high-
strength slaughterhouse waste treatment was independent of HRT and organic loading rate
(Jensen et al. 2015).
Increased VFA concentration signalled disturbance in SP and TP-MP at high OLRs. At 10 g COD
L·day
-1
, total VFAs increased 33.5 fold in SP and 21.4 fold in TP-MP (Figure S28). Total ammonia
nitrogen (TAN) in both systems approached inhibiting concentrations at 10 and 15 g COD L·day
-1

(Figure S3). At the highest OLR, TAN reached 1820 mg·L
-1
in SP and 1840 mg·L
-1
in TP-MP, with
calculated free ammonia nitrogen (FAN) concentrations reaching 235 mg·L
-1
and 279 mg·L
-1
in SP
and TP-MP, respectively. Although reported inhibitory ammonia concentrations vary greatly
across studies in the literature (Amha et al. 2018), FAN concentrations as low as 80 mg·L
-1
have
been shown to cause stress on anaerobic microbial communities (Werner et al. 2014). It is
important to note that a significant decrease in pH was not observed (Figure S22) in either system
due to the high alkalinity, 7.8 ± 1.1 and 7.6 ± 1.6 mg L
-1
as CaCO 3 in SP and TP-MP, respectively,

109

at the highest OLR. Effluent COD concentrations increased by 27.9% in SP and 27.8% in TP-MP at
15 g COD L·day
-1
compared to 2.5 g COD L·day
-1
(Figure S29). However, COD removal efficiency
remained high throughout operation, remaining ≥99% for both SP and TP.
4.4.3. TP enriched for syntrophic fatty-acid oxidizing bacteria while stable methanogen activity
suggested functional redundancy
Higher relative activity of syntrophic fatty-acid oxidizing bacteria was observed in TP-MP
compared to SP, whereas methanogens showed similar activity in both systems at all OLRs (Figure
13). We primarily relied on RNA-based sequencing data, due to the higher sensitivity of RNA-
based data relative to DNA-based data, particularly when evaluating inhibiting conditions (Amha
et al. 2018). At 10 g COD L·day
-1
, where the highest difference in performance was observed
between SP and TP-MP, potential syntrophs accounted for 5.13 ± 3.17% relative activity in SP and
10.2 ± 3.3% relative activity in TP-MP. Methanogens represented 19.2 ± 2.1% and 20.5 ± 3.4%
relative activity in SP and TP-MP, respectively. These results suggest that methanogens had
redundant functionality and were not directly linked to the reduced relative performance of SP.
Even at 15 g COD L·day
-1
, where both systems showed rapid decline in methane production per
OLR, methanogens remained at high relative activity, 19.7 ± 2.1% in SP and 26.9 ± 1.8% in TP-MP.
Our results are in-line with a review paper (Carballa et al. 2015) that theorized based on response
of the anaerobic microbiome to disturbances that methanogens could be resistant (endure
changes), resilient (rebound after inhibition), and redundant (replace population with similar
functionality). However, acetogens/syntrophs were classified as only resistant and resilient and
are highly ‘function-specialized’ (Carballa et al. 2015), corroborating our observed positive
correlation between syntroph activity and performance. The sustained high activity of

110

methanogens in both SP and TP-MP indicates that inhibition of other key populations led to the
poor performance at 10 g COD L·day
-1
in SP and 15 g COD L·day
-1
in TP-MP. Methanosaeta were
the most active methanogens in both systems, which was also observed in ML samples from the
full-scale AnMBR (Figure S30), where Methanosaeta accounted for 19.5 ± 0.03% relative activity.

Figure 13. (A) Relative activity of syntrophic fatty-acid oxidizers, and (B) relative activity of
methanogens identified at the genus level where possible using 16S rRNA sequencing for SP
and TP-MP, at increasing OLRs. The results shown are average data from three separate
sampling points taken for each OLR. Results are expressed as a percentage normalized using
total of 16S rRNA sequences (Bacteria and Archaea). Truncated y-axes (0 to 11% and 0 to 50%
on figure A and B, respectively) are shown to accentuate differences in activity. The secondary
y-axis and pink diamond shaped data points for figure A and B signify the ratio of TP-MP to SP
for total syntrophs and methanogens, respectively.

A.
B.

111

4.4.4. Microbial activity data revealed increased diversity in TP-MP at high OLRs and distinct
community profile
In evaluating the microbial community data, we focused on three questions: (i) were there
distinct differences in microbial community dynamics between SP and TP-MP; (ii) which microbial
communities in SP and TP-MP were resilient at higher OLRs in the presence of potential inhibitors
(e.g., ammonia, salt, and VFAs); and (iii) which communities were linked with increased
performance at high OLRs? We utilized Inverse Simpson Index, NMDS analysis, genus-level
classification, Linear Discriminant Analysis (LDA), and Spearman Rank analyses to address the
above three questions.

Figure 14. Inverse Simpson Index of 16S rRNA gene sequencing results for SP, TP-MP, and TP-
AP at different Organic Loading Rates (OLRs). Inverse Simpson Index measures richness in
community or alpha-diversity. The error bar indicates standard deviation of triplicate sampling
days for each OLR.

112

Inverse Simpson Index indicated higher diversity in TP-MP for all OLRs ≥ 3.5 g COD L·day
-1
,
relative to SP (Figure 14). A linear decline in diversity with increasing OLR was observed in SP,
whereas the TP-MP community had relatively stable diversity throughout operation. NMDS
analysis showed some spatial segregation between SP and TP communities (Figure S31). ANOVA
analysis of the ordination indicated that the microbial communities in SP and TP-MP were
statistically significantly different for the respective DNA- and RNA-based data (p<0.001). Initially,
when both systems were operated in SP mode, they showed similar ordination (SP1 and SP2;
Figure S31). However, with increasing OLRs, the difference in SP and TP-MP ordination became
apparent. In the TP-AP community, which had relatively low diversity, the DNA- and RNA-based
data clustered together, and ANOVA analysis confirmed that the differences in ordination were
not statistically significant (p=0.19).
RNA-based microbial activity data was first screened to only evaluate genera present in ≥ 3%
relative activity in at least one sample, and the resultant genera subdivided into three distinct
groups according to their average relative activity change from 2.5 g COD L·day
-1
(SP1/SP2) to an
OLR of 10 or 15 g COD L·day
-1
: (1) decreased by ≥ 50%, (2) showed ≤ 50% change, and (3)
increased by ≥ 50% (Figure 15). One of the most notable trends in Group 1 was the increase of
Leptotrichiaceae with increasing OLRs, particularly in SP. Leptotrichiaceae relative activity was
only 2.27 ± 0.48% during the initial phase (SP1 at 2.5 g COD L·day
-1
), but increased to 10.6 ± 1.7%
at 15 g COD L·day
-1
(Figure 15). Although the family Leptotrichiaceae, generally known to
metabolize a wide range of carbohydrates including disaccharides (Lory 2014), have been
identified in other similar studies (Stolze et al. 2016, Ziels et al. 2018), they are poorly

113

characterized in anaerobic digestion systems. Another important trend from Group 1
communities identified in SP was the substantial increase of Pelolinea, which was not observed
in TP-MP. Pelolinea, a recently classified genus from subseafloor sediments (Imachi et al. 2014),
accounted for 2.23 ± 0.85% relative activity initially and increased to 4.73 ± 2.10% and 6.40 ±
0.73% at 10 and 15 g COD L·day
-1
, respectively. Pelolinea are filamentous chemoorganotrophs
that ferment sugars (Imachi et al. 2014). To our knowledge, we are the first study to report the
proliferation of this genus in anaerobic digestion systems. It is possible that salt accumulation at
higher OLRs could have provided a selective pressure resulting in increased activity of Pelolinea
as they are known to be halotolerant (Imachi et al. 2014). In our study, chloride concentrations
in SP biomass at 15 g COD L·day
-1
OLR exceeded 2 g·L
-1
.

114

Figure 15. (A). Relative activity for genera that showed ≥3% relative activity in at least one
sample in SP, and (B). TP-MP samples. Three distinct groups were formed based on relative
activity change in either 10 or 15 g COD L·day
-1
relative to SP1/SP2, with increased OLR: (1)
decreased by ≥ 50% (red-fill) (2) showed ≤ 50% change (blue-fill), and (3) increased by ≥ 50%
(green-fill). Results are expressed as a percentage normalized using total of 16S rRNA sequences
(Bacteria and Archaea). Truncated y-axes (0 to 90%) are shown to accentuate differences in
activity.

A.
B.

115

Figure 16. (A) Taxonomic Cladogram for groups that showed significant differential relative
activity in one of the four categories, low (2.5-3.5 g COD L·day
-1
), medium (5 g COD L·day
-1
), or
high (10 and 15 g COD L·day
-1
) OLRs for SP, and (B) TP-MP samples. All taxa presented here
showed significant differential activity in one of the OLR categories by resulting in Linear
Discriminant Analysis (LDA) score of ≥ 2. The analysis was conducted with LEfSe tool and
relative activity data was used as input for all groups that showed ≥0.5% relative activity in at
least one sample. The highest available classification level for each OTU is used for labelling.

The genera that showed the most substantial increase in relative activity in TP-MP were
Lactobacillus and Treponema (Figure 15). Treponema increased from 0.37 ± 0.06% in SP2 to 3.53
± 3.02% at 15 g COD L·day
-1
in TP-MP. Treponema isolates from termite guts have been
characterized as homoacetogens via the Wood-Ljungdahl pathway, while rumen Treponema
strains have been described as fermenters able to degrade polysaccharides and disaccharides
(Bekele et al. 2011, Paster and Canale-Parola 1985). Treponema have been found to dominate

116

during high acetate concentrations (Xie et al. 2018) and the possibility of this population being
syntrophic acetate oxidizers has previously been theorized (Ahlert et al. 2016). Also noteworthy,
Treponema have been shown to persist in high TAN of up 10 g·L
-1
(Poirier et al. 2016), and this
could have benefited their activity levels in TP-MP. Unclassified bacteria and Proteobacteria also
steeply increased in TP-MP with increasing OLR, indicating that at higher OLR, yet to be described
populations likely play an important role in how these systems adapt to inhibition.
To characterize microbial populations that were resilient to inhibition at high OLRs (10 or 15 g
COD L·day
-1
), we used non-parametric analysis with the LEfSe (Segata et al. 2011) tool for all
genera ≥ 0.5% relative activity in at least one sample. The LEfSe tool applies an LDA analysis,
where a score of ≥ 2 is deemed significant (Figure 16). Our results indicated that three
methanogens in TP-MP, Methanoculleus, Methanospirillum, and Methanomassiliicoccus, showed
differential activity at 15 g COD L·day
-1
. Conversely, Methanoculleus and Methanospirillum were
more prominent in low OLRs (2.5-3.5 g COD L·day
-1
) in SP. In TP-MP, potential syntrophs,
unclassified Syntrophaceae and Candidatus Cloacimonas, showed high activity at 15 g COD L·day
-
1
. A genomic reconstruction study (Pelletier et al. 2008) on Candidatus Cloacimonas indicated
that this population is a hydrogen producing syntroph. Significant LDA scores at high OLRs in TP-
MP were also observed for two populations with limited characterization in anaerobic digestion,
Pedosphaeraceae and Phaselicystis. A single genome study (Martinez-Garcia et al. 2012) reported
similar sequences of Pedosphaeraceae and found that phylotypes of Verrucomicrobia within this
family significantly contributed to polysaccharide degradation, and we speculate that a similar
phenomenon could explain their high activity in TP-MP.
For SP, only five genera, Anaerolineaceae, Methanomicrobiales, Pelolinea, Syntrophaceae, and

117

Treponema, showed significant LDA scores indicating differential grouping at high OLRs (10 or 15
g COD L·day
-1
), suggesting that there was more widespread inhibition of active microbial
populations in SP (Figure 16). Candidatus Caldatribacterium, Methanobacterium and
Methanoculleus were found to be prominent taxa at low OLRs, with the highest LDA scores. In
contrast, Methanobacterium and Methanoculleus were among taxa that exhibited differential
grouping at high OLRs in TP-MP. Candidatus Caldatribacterium and unclassified Atribacteria JS1
were present at significant LDA at low OLRs in both SP and TP-MP, suggesting that these
populations may be sensitive to high OLR regardless of SP or TP mode. Candidatus
Caldatribacterium belongs in the recently proposed phylum Atribacteria JS1 (Dodsworth et al.
2013). These populations are linked with possible syntrophic propionate metabolism (Nobu et al.
2016) and their disappearance in TP-MP could be associated with the observed transition from
propionate- to acetate-dominated TP-AP (Figure S21). Although these populations have also
been shown to metabolize acetate (Nobu et al. 2016), they may have been outcompeted by other
acetate scavengers.
Spearman Rank analysis revealed that several of the populations that correlated with methane
production were methanogens or confirmed/suspected acetogens/syntrophs (Figure 17).
Populations that were present prominently in both SP and TP-MP and showed significant (p<0.05)
positive correlation with methane production were Spirochaetaceae M2PT2-76 termite group,
Blvii28 wastewater-sludge group unclassified Proteobacteria, and Syntrophaceae. The only
described species in the genus Blvii28_wastewater-sludge_group, Acetobacteroides
hydrogenigenes, has been described as a fermenter able to utilize a range of carbohydrates (Su
et al. 2014).

118

Figure 17. (A) Average relative activity of communities that showed significant (p<0.05)
correlation with methane production in SP, and (B). TP-MP, identified at the genus level where
possible based with 16S rRNA sequencing. The relative activity results shown are average data
from three separate sampling points taken for each OLR. Results are expressed as a percentage
normalized using total of 16S rRNA sequences (Bacteria and Archaea). Truncated y-axes (0 to
50%) are shown to accentuate differences in activity. The Spearman correlation coefficient
values (ρ) are shown as x-axis on the legend. The secondary y-axis and blue diamond shaped
data points for both figure A and B signify the mean methane production per day for each OLR.
For each OLR, samples from three time points were sequenced, thus, the Spearman rank
analysis was conducted using this triplicate data points for each OLR, i.e. relative activity data
for all groups that showed ≥ 0.5% relative activity in at least one sample with methane
production data for that specific sampling date.

Overall, the comparative microbial activity results shed light on which populations were
A.
B.

119

resistant to high OLRs in SP and TP-MP. We were also able to link the improved performance of
TP relative to SP for OLRs 3.5-10 g COD L·day
-1
to changes in microbial community dynamics, i.e.,
increased and stable diversity (inverse Simpson index), higher number of taxa that were resistant
to increased OLR conditions (LDA scores), and increased activity of specific populations, such as
Lactobacillus and Treponema (genera that showed ≥50% increase at high OLRs). Further, it is
important to note that several populations highly correlated with methane production and also
displaying significant differential grouping at high OLRs, have not been well characterized in
literature on anaerobic systems. For example, Spirochaetaceae M2PT2-76_termite_group have
only been described in the termite gut microbiome, yet showed significant correlation with
methane production in both SP and TP-MP. Other populations with similar trends were
Phaselicystis, Pedosphaeraceae, Paludibacteraceae, and Leptotrichiaceae. This indicates that
future research should be dedicated to understanding the underlying competitive advantage
these populations may have during inhibiting conditions. Our results are somewhat limited by
our reliance on relative activity and abundance data, as opposed to absolute data. In addition,
more advanced approaches that link specific substrate uptake or metabolism with detected
microbial communities, such as fluorescent reporters and labelled substrates (Amha et al. 2018),
could better elucidate how these systems adjust to operating changes. Future studies should
apply more advanced molecular tools to provide a mechanistic understanding of community
resistance to operating parameters such as OLR.
4.4.5. Fouling was severe, but reversible at increased OLRs, with SP and TP-MP biofilms
exhibiting similar community structure and activity profiles
Transmembrane pressure (TMP) remained low, between 2 and 10 kPa for both SP and TP-MP,

120

during the majority of operation. TMP >10 kPa was first observed after 224 and 193 days of
operation in SP and TP-MP, respectively, with maximum TMP recorded at 48 kPa for SP and 45
kPa for TP-MP. Fouling was found to be reversible with chemical cleaning, using hydrochloric acid
(0.1 M) and sodium hypochlorite (5% W/V). We observed that there was significantly increased
reversible fouling with increased TVS and TS, requiring more frequent chemical cleaning at OLRs
>10 g COD L·day
-1
, which necessitated the reduction of SRT to maintain operable TS
concentrations. Six membrane cleaning cycles were conducted throughout the study. Notably,
long-term operation and chemical cleaning did not compromise ceramic membrane integrity
based on visual observations and effluent quality, which has been an operational challenge in
full-scale AnMBRs equipped with polymeric membranes. Biomass samples taken from the
biofilm/membrane foulant layer (10 g COD L·day
-1
OLR) indicated similar community structure
and activity profiles in SP and TP-MP, but a distinct community compared to the respective
biomass samples. NMDS revealed similar ordination for the biofilm DNA- and RNA-based analyses
for both SP and TP-MP (Figure S31). Therefore, we grouped these samples together to conduct a
non-parametric t-test to identify communities that showed significant differential
representation. Comparing SP and TP-MP biofilm communities indicated that Geobacter had
significantly differential representation in the SP biofilm, whereas Treponema was more present
the TP-MP biofilm. Biofilm methanogenic relative activity was 40% and 56% greater in biofilm
samples relative to biomass samples for SP and TP-MP, respectively. This correlates with
observations on similar dates where biomass samples showed higher VFA concentrations relative
to effluent samples (Figure S28). Thus, the biofilm community aided in increasing permeate
quality, as also indicated by the high COD removal efficiency, even at high OLRs (Figure S29). A

121

similar observation has been reported during low-strength wastewater treatment, where biofilm
development significantly improved permeate quality by reducing effluent acetate and
propionate concentrations (Smith et al. 2015).
Prominent communities that showed significant differential presence in the SP biofilm relative
to biomass were Anaerolineaceae and Synergistaceae, with 78% and 342% increase, respectively
(Figure S34). Synergistaceae (243% increase) and Syntrophomonas (212% increase) were
significantly more present in the TP-MP biofilm relative to biomass. Various studies have
reported co-occurrence of Anaerolineaceae alongside methanogens (McIlroy et al. 2017, Xu et
al. 2016), prompting discussion of this genus as potential exoelectrogens that co-exist with
Methanosaeta (McIlroy et al. 2017, Xu et al. 2016). In addition, an omics study showed
populations within Anaerolineae having substantially active type VI pili, which enables cellular
attachment, and provide a competitive advantage during attached-growth mode. OTUs within
this class had relative activity of 18% and 7.7% in SP and TP-MP biofilm samples, respectively.
Although speculative, the overall higher activity of Geobacter and members of Anaerolineae in
SP compared to TP-MP in both biofilm and biomass samples suggests that direct interspecies
electron transfer (DIET) could be an adaptive mechanism during SP treatment at higher OLRs,
whereas TP-MP systems enrich activity of traditional syntrophic populations.
Compared to conventional ADs used for FW treatment, AnMBRs offer various benefits, such
as operation at higher OLRs and improved effluent quality (Cheng et al. 2018). A study that
compared SP and TP ADs for FW treatment showed that OLR exceeding 3.55 g COD L·day
-1

resulted in severe instability in both reactors (Voelklein et al. 2017). The highest performance
reported in this study was 0.22 L CH 4 g COD
-1
for SP and 0.295 L CH 4 g COD
-1
in TP, which were

122

observed at 2.84 g COD L·day
-1
. Another study (Voelklein et al. 2016) operated TP-MP reactors
up to 7.1 g COD L·day
-1
and reported peak performance at 3.3 g COD L·day
-1
, which was 0.284 L
CH 4 g COD
-1
. Both aforementioned studies (Voelklein et al. 2016, Voelklein et al. 2017) showed
lower specific methane production rates and lower maximum OLRs for both SP and TP relative
to our study. In terms of microbial community dynamics, our study reported significantly higher
syntrophic community (Amha et al. 2017a) and methanogenic community (Fontana et al. 2018)
compared to other studies that applied ADs for FW treatment. Notably, the high activity of
Methanosaeta in our study at all OLRs, particularly in biofilm samples, indicated additional
advantage of AnMBRs, by promoting the growth of these methanogenic groups known for
affinity for surface attachment (Harb et al. 2015), even at operating conditions reported to cause
inhibition of these groups in ADs (Fontana et al. 2018). Since, there are no prior studies that have
studied microbial dynamics in AnMBRs treating FW, our study provides an original contribution
in describing both microbial structure and activity changes with FW treatment at high OLRs.
Overall, this study showed how microbial community dynamics can be manipulated using phase-
separation and increasing OLRs. It also highlighted the advantages of AnMBR treatment of FW,
by increasing effluent quality and performance. There are important tradeoffs to consider for
full-scale applications when comparing AnMBRs and ADs, particularly in terms of the added costs
for incorporating membrane separation. Future research should investigate how AnMBRs
compare in cost and environmental impact with conventional ADs for FW treatment to provide
a systems-level comparison of these management options.
We have demonstrated that TP relative to SP mode improves AnMBR treatment of FW,
particularly at high OLRs. Our study demonstrated (1) increased methane recovery of up to 20%

123

at high OLRs for TP, (2) significant enrichment of syntrophic bacteria, (3) increased diversity at
high OLRs likely resulting in a more resilient community in the presence of microbial inhibitors,
and (4) improved hydrolysis (and VFA production) by enriching for fermentative bacteria in the
AP at low pH, while effectively inhibiting methanogens and syntrophs. In addition, we
demonstrated that ceramic membranes can be applied during long-term FW treatment without
non-reversible fouling. The results indicate that the full-scale AnMBR operated by Divert Inc. can
be safely operated at 4 times the current average OLR (2.5 g COD L·day
-1
) by incorporating TP
treatment. Identifying higher operational limits has important implications because of the
expected increase of FW recycling in California and elsewhere as new regulations mandate FW
landfill diversion. Higher OLR increases the economic favourability of anaerobic treatment
systems, by increasing energy production and minimizing reactor footprint and capital costs. TP
AnMBRs are worth considering to increase efficiency of decentralized anaerobic treatment of
FW.

4.5 Literature Cited
Ahlert, S., Zimmermann, R., Ebling, J. and König, H. (2016) Analysis of propionate-degrading consortia
from agricultural biogas plants. Microbiologyopen 5(6), 1027-1037.
Amha, Y.M., Anwar, M.Z., Brower, A., Jacobsen, C.S., Stadler, L.B., Webster, T.M. and Smith, A.L. (2018)
Inhibition of anaerobic digestion processes: applications of molecular tools. Bioresource Technology
247, 999-1014.
Amha, Y.M., Sinha, P., Lagman, J., Gregori, M. and Smith, A.L. (2017a) Elucidating microbial community
adaptation to anaerobic co-digestion of fats, oils, and grease and food waste. Water research 123, 277-
289.
Amha, Y.M., Sinha, P., Lagman, J., Gregori, M. and Smith, A.L. (2017b) Elucidating microbial community
adaptation to anaerobic co-digestion of fats, oils, and grease and food waste. Water research.
APHA (2005) Standard Methods for the Examination of Water and Wastewater, American Public Health
Association, Washington, D.C.
Becker, A.M., Yu, K., Stadler, L.B. and Smith, A.L. (2017) Co-management of domestic wastewater and
food waste: A life cycle comparison of alternative food waste diversion strategies. Bioresource
Technology 223, 131-140.

124

Bekele, A.Z., Kobayashi, Y. and Koike, S. (2011) Phylogenetic diversity and dietary association of rumen
Treponema revealed using group-specific 16S rRNA gene-based analysis. FEMS Microbiology Letters
316(1), 51-60.
Bhattacharya, S.K., Madura, R.L., Walling, D.A. and Farrell, J.B. (1996) Volatile solids reduction in two-
phase and conventional anaerobic sludge digestion. Water research 30(5), 1041-1048.
Bowles, E.C., Novak, J.T., Bott, C.B. and Pruden, A. (2012) Performance of Acid-Gas Anaerobic Digesters
for Minimization of Siloxane and Hydrogen Sulfide in the Biogas. Proceedings of the Water Environment
Federation 2012(2), 639-653.
Cao, J., Zhang, Q., Wu, S., Luo, J., Wu, Y., Zhang, L., Feng, Q., Fang, F. and Xue, Z. (2019) Enhancing the
anaerobic bioconversion of complex organics in food wastes for volatile fatty acids production by zero-
valent iron and persulfate stimulation. Science of The Total Environment 669, 540-546.
Carballa, M., Regueiro, L. and Lema, J.M. (2015) Microbial management of anaerobic digestion:
exploiting the microbiome-functionality nexus. Current opinion in biotechnology 33, 103-111.
Chen, Y., Jiang, X., Xiao, K., Shen, N., Zeng, R.J. and Zhou, Y. (2017) Enhanced volatile fatty acids (VFAs)
production in a thermophilic fermenter with stepwise pH increase – Investigation on dissolved organic
matter transformation and microbial community shift. Water research 112, 261-268.
Cheng, H., Hiro, Y., Hojo, T. and Li, Y.-Y. (2018) Upgrading methane fermentation of food waste by using
a hollow fiber type anaerobic membrane bioreactor. Bioresource Technology 267, 386-394.
Cho, K., Jeong, Y., Seo, K.W., Lee, S., Smith, A.L., Shin, S.G., Cho, S.-K. and Park, C. (2018) Effects of
changes in temperature on treatment performance and energy recovery at mainstream anaerobic
ceramic membrane bioreactor for food waste recycling wastewater treatment. Bioresource Technology
256, 137-144.
Dodsworth, J.A., Blainey, P.C., Murugapiran, S.K., Swingley, W.D., Ross, C.A., Tringe, S.G., Chain, P.S.,
Scholz, M.B., Lo, C.-C. and Raymond, J. (2013) Single-cell and metagenomic analyses indicate a
fermentative and saccharolytic lifestyle for members of the OP9 lineage. Nature communications 4,
1854.
Fontana, A., Campanaro, S., Treu, L., Kougias, P.G., Cappa, F., Morelli, L. and Angelidaki, I. (2018)
Performance and genome-centric metagenomics of thermophilic single and two-stage anaerobic
digesters treating cheese wastes. Water research 134, 181-191.
Fukaya, M., Park, Y. and Toda, K. (1992) Improvement of acetic acid fermentation by molecular breeding
and process development. Journal of applied bacteriology 73(6), 447-454.
Ganesh, R., Torrijos, M., Sousbie, P., Lugardon, A., Steyer, J.P. and Delgenes, J.P. (2014) Single-phase and
two-phase anaerobic digestion of fruit and vegetable waste: Comparison of start-up, reactor stability
and process performance. Waste Management 34(5), 875-885.
Harb, M., Xiong, Y., Guest, J., Amy, G. and Hong, P.-Y. (2015) Differences in microbial communities and
performance between suspended and attached growth anaerobic membrane bioreactors treating
synthetic municipal wastewater. Environmental Science: Water Research & Technology 1(6), 800-813.
IIZUKA, H. and KOMAGATA, K. (1963) Taxonomy of genus Pseudomonas with special reference to their
modes of metabolism of carbon compounds. The Journal of General and Applied Microbiology 9(1), 83-
95.
Imachi, H., Sakai, S., Lipp, J.S., Miyazaki, M., Saito, Y., Yamanaka, Y., Hinrichs, K.-U., Inagaki, F. and Takai,
K. (2014) Pelolinea submarina gen. nov., sp. nov., an anaerobic, filamentous bacterium of the phylum
Chloroflexi isolated from subseafloor sediment. International Journal of Systematic and Evolutionary
Microbiology 64(3), 812-818.
Jensen, P., Yap, S., Boyle-Gotla, A., Janoschka, J., Carney, C., Pidou, M. and Batstone, D. (2015) Anaerobic
membrane bioreactors enable high rate treatment of slaughterhouse wastewater. Biochemical
engineering journal 97, 132-141.

125

Jeong, Y., Hermanowicz, S.W. and Park, C. (2017) Treatment of food waste recycling wastewater using
anaerobic ceramic membrane bioreactor for biogas production in mainstream treatment process of
domestic wastewater. Water research 123, 86-95.
Jo, Y., Kim, J., Hwang, K. and Lee, C. (2018) A comparative study of single-and two-phase anaerobic
digestion of food waste under uncontrolled pH conditions. Waste Management 78, 509-520.
Khan, M.A., Ngo, H.H., Guo, W., Chang, S.W., Nguyen, D.D., Varjani, S., Liu, Y., Deng, L. and Cheng, C.
(2019) Selective production of volatile fatty acids at different pH in an anaerobic membrane bioreactor.
Bioresource Technology 283, 120-128.
Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K. and Schloss, P.D. (2013) Development of a dual-
index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq
Illumina sequencing platform. Applied and environmental microbiology 79(17), 5112-5120.
Lory, S. (2014) The family Leptotrichiaceae. The Prokaryotes: Firmicutes and Tenericutes, 213-214.
Martinez-Garcia, M., Brazel, D.M., Swan, B.K., Arnosti, C., Chain, P.S., Reitenga, K.G., Xie, G., Poulton,
N.J., Gomez, M.L. and Masland, D.E. (2012) Capturing single cell genomes of active polysaccharide
degraders: an unexpected contribution of Verrucomicrobia. PloS one 7(4), e35314.
McIlroy, S.J., Kirkegaard, R.H., Dueholm, M.S., Fernando, E., Karst, S.M., Albertsen, M. and Nielsen, P.H.
(2017) Culture-Independent Analyses Reveal Novel Anaerolineaceae as Abundant Primary Fermenters in
Anaerobic Digesters Treating Waste Activated Sludge. Frontiers in microbiology 8(1134).
Nobu, M.K., Dodsworth, J.A., Murugapiran, S.K., Rinke, C., Gies, E.A., Webster, G., Schwientek, P., Kille,
P., Parkes, R.J. and Sass, H. (2016) Phylogeny and physiology of candidate phylum ‘Atribacteria’(OP9/JS1)
inferred from cultivation-independent genomics. The ISME journal 10(2), 273.
Paster, B.J. and Canale-Parola, E. (1985) Treponema saccharophilum sp. nov., a large pectinolytic
spirochete from the bovine rumen. Appl. Environ. Microbiol. 50(2), 212-219.
Pelletier, E., Kreimeyer, A., Bocs, S., Rouy, Z., Gyapay, G., Chouari, R., Rivière, D., Ganesan, A., Daegelen,
P., Sghir, A., Cohen, G.N., Médigue, C., Weissenbach, J. and Le Paslier, D. (2008) "Candidatus
Cloacamonas acidaminovorans": genome sequence reconstruction provides a first glimpse of a new
bacterial division. Journal of bacteriology 190(7), 2572-2579.
Poirier, S., Desmond-Le Quéméner, E., Madigou, C., Bouchez, T. and Chapleur, O. (2016) Anaerobic
digestion of biowaste under extreme ammonia concentration: identification of key microbial phylotypes.
Bioresource Technology 207, 92-101.
Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J. and Glöckner, F.O. (2007) SILVA: a
comprehensive online resource for quality checked and aligned ribosomal RNA sequence data
compatible with ARB. Nucleic acids research 35(21), 7188-7196.
Rault, A., Bouix, M. and Béal, C. (2009) Fermentation pH influences the physiological-state dynamics of
Lactobacillus bulgaricus CFL1 during pH-controlled culture. Appl. Environ. Microbiol. 75(13), 4374-4381.
Sahu, N., Deshmukh, S., Chandrashekhar, B., Sharma, G., Kapley, A. and Pandey, R. (2017) Optimization
of hydrolysis conditions for minimizing ammonia accumulation in two-stage biogas production process
using kitchen waste for sustainable process development. Journal of environmental chemical
engineering 5(3), 2378-2387.
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley,
B.B., Parks, D.H. and Robinson, C.J. (2009) Introducing mothur: open-source, platform-independent,
community-supported software for describing and comparing microbial communities. Applied and
environmental microbiology 75(23), 7537-7541.
Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W.S. and Huttenhower, C. (2011)
Metagenomic biomarker discovery and explanation. Genome Biology 12(6), R60.
Shen, F., Yuan, H., Pang, Y., Chen, S., Zhu, B., Zou, D., Liu, Y., Ma, J., Yu, L. and Li, X. (2013) Performances
of anaerobic co-digestion of fruit & vegetable waste (FVW) and food waste (FW): single-phase vs. two-
phase. Bioresource Technology 144, 80-85.

126

Shin, S.G., Han, G., Lim, J., Lee, C. and Hwang, S. (2010) A comprehensive microbial insight into two-
stage anaerobic digestion of food waste-recycling wastewater. Water research 44(17), 4838-4849.
Simpson, P.J., Ross, R.P., Fitzgerald, G.F. and Stanton, C. (2004) Bifidobacterium psychraerophilum sp.
nov. and Aeriscardovia aeriphila gen. nov., sp. nov., isolated from a porcine caecum. International
Journal of Systematic and Evolutionary Microbiology 54(2), 401-406.
Smith, A., Shimada, T. and Raskin, L. (2017) A comparative evaluation of community structure in full-
scale digesters indicates that two-phase digesters exhibit greater microbial diversity than single-phase
digesters. Environmental Science: Water Research & Technology 3(2), 304-311.
Smith, A.L., Skerlos, S.J. and Raskin, L. (2015) Membrane biofilm development improves COD removal in
anaerobic membrane bioreactor wastewater treatment. Microbial biotechnology 8(5), 883-894.
Stolze, Y., Bremges, A., Rumming, M., Henke, C., Maus, I., Pühler, A., Sczyrba, A. and Schlüter, A. (2016)
Identification and genome reconstruction of abundant distinct taxa in microbiomes from one
thermophilic and three mesophilic production-scale biogas plants. Biotechnology for biofuels 9(1), 156.
Sträuber, H., Bühligen, F., Kleinsteuber, S. and Dittrich-Zechendorf, M. (2018) Carboxylic acid production
from ensiled crops in anaerobic solid-state fermentation-trace elements as pH controlling agents
support microbial chain elongation with lactic acid. Engineering in Life Sciences 18(7), 447-458.
Su, X.-L., Tian, Q., Zhang, J., Yuan, X.-Z., Shi, X.-S., Guo, R.-B. and Qiu, Y.-L. (2014) Acetobacteroides
hydrogenigenes gen. nov., sp nov., an anaerobic hydrogen-producing bacterium in the family
Rikenellaceae isolated from a reed swamp.
Voelklein, M., Jacob, A., O’Shea, R. and Murphy, J. (2016) Assessment of increasing loading rate on two-
stage digestion of food waste. Bioresource Technology 202, 172-180.
Voelklein, M.A., O'Shea, R., Jacob, A. and Murphy, J.D. (2017) Role of trace elements in single and two-
stage digestion of food waste at high organic loading rates. Energy 121, 185-192.
Werner, J.J., Garcia, M.L., Perkins, S.D., Yarasheski, K.E., Smith, S.R., Muegge, B.D., Stadermann, F.J.,
DeRito, C.M., Floss, C. and Madsen, E.L. (2014) Microbial community dynamics and stability during an
ammonia-induced shift to syntrophic acetate oxidation. Applied and environmental microbiology 80(11),
3375-3383.
Wu, Y., Wang, C., Liu, X., Ma, H., Wu, J., Zuo, J. and Wang, K. (2016) A new method of two-phase
anaerobic digestion for fruit and vegetable waste treatment. Bioresource Technology 211, 16-23.
Xiao, X., Huang, Z., Ruan, W., Yan, L., Miao, H., Ren, H. and Zhao, M. (2015) Evaluation and
characterization during the anaerobic digestion of high-strength kitchen waste slurry via a pilot-scale
anaerobic membrane bioreactor. Bioresource Technology 193, 234-242.
Xie, X., Yang, C., Guan, L.L., Wang, J., Xue, M. and Liu, J.X. (2018) Persistence of cellulolytic bacteria
Fibrobacter and Treponema after short-term corn stover-based dietary intervention reveals the
potential to improve rumen fibrolytic function. Frontiers in microbiology 9.
Xu, H., Wang, C., Yan, K., Wu, J., Zuo, J. and Wang, K. (2016) Anaerobic granule-based biofilms formation
reduces propionate accumulation under high H2 partial pressure using conductive carbon felt particles.
Bioresource Technology 216, 677-683.
Xu, Z., Zhao, M., Miao, H., Huang, Z., Gao, S. and Ruan, W. (2014) In situ volatile fatty acids influence
biogas generation from kitchen wastes by anaerobic digestion. Bioresource Technology 163, 186-192.
Zamorano-López, N., Moñino, P., Borrás, L., Aguado, D., Barat, R., Ferrer, J. and Seco, A. (2018) Influence
of food waste addition over microbial communities in an Anaerobic Membrane Bioreactor plant treating
urban wastewater. Journal of environmental management 217, 788-796.
Ziels, R.M., Svensson, B.H., Sundberg, C., Larsson, M., Karlsson, A. and Yekta, S.S. (2018) Microbial rRNA
gene expression and co-occurrence profiles associate with biokinetics and elemental composition in full-
scale anaerobic digesters. Microbial biotechnology 11(4), 694-709.

127

CHAPTER 5

5. Co-digestion of FOG improves performance of anaerobic membrane
bioreactor treatment of food waste

5.1 Abstract
Anaerobic membrane bioreactors (AnMBRs) are an attractive technology that can increase
energy recovery during food waste (FW) treatment while producing a high effluent quality. Co-
digestion of fats, oils, and grease (FOG) alongside FW was evaluated in this study, as an
operational strategy to improve energy recovery. The upper limits of FOG addition without
microbial inhibition and severe membrane fouling was evaluated by increasing fats addition with
FOG incrementally (0.5 - 1 kg m
3
·day
-1
), where single-phase (SP) and two-phase, with acid-phase
(TP-AP) and methane-phase (TP-MP), AnMBRs were used for treatment. The results show that
the sludge acclimated to FOG addition after 82 days of operation, with fats loading of 0.5 - 0.75
kg m
3
·day
-1
. Higher methane production, ranging from 4 - 12% was observed in TP-MP compared
to SP, but the difference in means were not found statistically significant due to high variability
in daily methane production. However, the results showed that co-digestion with FOG improved
daily methane recovery significantly (p<0.05) in both SP and TP at fats loading of 1 kg m
3
·day
-1
.

128

In SP, the increase between mono-digestion and co-digestion was 11%, while an increase of 13%
was observed in TP. In addition, microbial community analyses from weekly biomass samples
taken from each reactor are expected to show enrichment of long chain fatty acid (LCFA)
degrading communities. This study shows, for the first time, that FW and FOG can be co-digested
effectively in AnMBR and is expected to advise full-scale decisions on optimum fats loading and
impact of reactor design.
5.2 Introduction
Fats, oils, and grease (FOG) originate from cooking and food processing industries, and are
collected in grease traps and interceptors that prevent damage to sewage collection systems
(Alqaralleh et al. 2016, Amha et al. 2017). The treatment of FOG in anaerobic digesters can reduce
environmental impact by diverting waste from landfills (the conventional method of treating
FOG), while enabling energy recovery (Alqaralleh et al. 2016). Various studies have reported
increased energy recovery by co-digesting FOG with other waste streams (Amha et al. 2017, Long
et al. 2012). Anaerobic membrane bioreactors (AnMBRs) are an attractive option in organic waste
management, due to the potential to implement energy-positive treatment, while enabling high
reduction of chemical oxygen demand (COD) in permeate, as demonstrated in a study that used
AnMBR to treat food processing wastewater (Galib et al. 2016). In addition, the legislative push
in California and elsewhere to divert organic waste from landfills makes food waste (FW) and
FOG treatment in AnMBRs an attractive management strategy. The addition of FOG during
AnMBR treatment can enhance energy recovery by increasing the organic loading and potentially
compensating for temporal fluctuations in FW characteristics.

129

We demonstrated in a previous study (Chapter 3) that addition of FOG with FW during anaerobic
digestion led to increased activity of key syntrophic fatty acid degrading populations that were
directly correlated with improved performance, compared to digestion without FOG addition
(Amha et al. 2017). However, addition of FOG can also increase the concentration of long chain
fatty acids (LCFAs) that can negatively impact the microbial community by causing damage to cell
membranes, reducing nutrient transport, and decreasing cell permeability (Long et al. 2012,
Palatsi et al. 2010, Sousa et al. 2013). A full-scale AnMBR operated by Divert Inc., has set the
internal limit for maximum fats loading rate at 0.5 kg m
3
·day
-1
and have witnessed stress
condition, i.e. erratic gas production, fat floatation, and formation of calcium oleate fat balls,
when fats addition exceeds this value. Another study has also reported a similar formation of fat
balls at fat loading rates up to 1.16 kg m
3
·day
-1
during high-lipid ethanol thin stillage wastewater
treatment (Dereli et al. 2014). Therefore, in this study we aimed to evaluate two-phase (TP)
treatment with acid-phase (TP-AP) and methane-phase (TP-MP), to evaluate a design strategy
that can improve degradation of fats in FW and FOG co-digestion in AnMBRs. In TP-AP, substrate
hydrolysis is optimized by maintaining short hydraulic retention time (HRT) and solids retention
time (SRT) and low pH (3.5). In addition, TP-MP is expected to enrich for a more diverse microbial
community with higher activity of syntrophic fatty acid oxidizers compared to SP, thereby,
increasing methane production. Although, there are limited studies that have investigated the
performance of AnMBRs used for high lipid wastewater, such as from slaughterhouse (Jensen et
al. 2015) and ethanol thin stillage (Dereli et al. 2012, Dereli et al. 2015, Dereli et al. 2014), FW
and FOG treatment in AnMBRs has yet to be evaluated. Further, there are no studies that have

130

characterized the microbial community changes during high-lipid wastewater treatment in
AnMBRs.
Another major concern with incorporating FOG in a membrane bioreactor is potential reversible
and irreversible fouling. A study (Ramos et al. 2014a) that used AnMBR to treat high oils and
grease (O&G) wastewater from a food processing facility reported that the system was stable at
an organic loading rate (OLR) of 2 kg COD m
3
d
-1
and O&G concentration of 4.6-36 g O&G L
-1
, while
critical flux decreased from 11.1 to 9.7 L m
2
h
-1
after 40 days of operation. Fouling rate ranged
between 0.96 and 3.95 mbar d
-1
during operation. A prior study by the same authors showed that
high concentration of lipids were present in the organic foulants removed after chemical cleaning
(Ramos et al. 2014b). On the contrary, a study that also used lipid-rich wastewater during
digestion of corn-to-ethanol thin stillage (Dereli et al. 2015) reported decreased fouling
propensity, due to an increase in hydrophobicity of the sludge with accumulation of LCFA.
The objective of this study was to evaluate co-digestion of FW and FOG during bench-scale SP
and TP AnMBR treatment. We increased FOG addition incrementally to investigate the optimum
and maximum FOG addition without inhibiting the microbial community. Ceramic membrane
performance was characterized over long-term operation to understand the impact of high fat
concentrations on fouling. Further, the microbial community response to FOG addition was
investigated using DNA- and RNA-based sequencing, to elucidate change in microbial structure
and activity, respectively. The results of this study are expected to advice full-scale AnMBRs that
plan to increase energy recovery with co-digestion.
5.3 Materials and Methods

131

5.3.1 Bench-scale AnMBR configurations
Two identical jacketed 7 L reactors (Chemglass, NJ) with 5 L liquid volume were used (Figure 18).
Each system (Reactor 1 and 2) contained a submerged flat-sheet ceramic microfiltration
membrane with pore size of 0.2 µm and a total effective membrane area of 0.011 m
2
. Both
reactors were mixed continuously at 250 rpm, with an impellor located near the bottom of the
reactor vessel. The reactors’ jacket was connected to a recirculating water bath (Fisher Scientific,
Hampton, NH) for temperature control with reactor temperature monitored via a probe
submerged in the biomass. Pressure in the headspace and permeate lines were controlled using
pressure transducers (Transducers Direct, Cincinnati, OH). Influent and permeate were pumped
using peristaltic pumps to maintain a constant liquid volume (NewEra, Farmingdale, NY and
Langer, Boonton, NJ). Membrane backwashing was conducted for 3 minutes during 10 minute
operational intervals for fouling control. The headspace was connected to a GFM mass flow
meter (Aalborg, New York) that measured the biogas produced and a mini diaphragm pump
(Parker, North Carolina) that recirculated produced biogas through sparging tubes mounted
below the membrane housing. All data acquisition and permeate pump controls were conducted
with LabVIEW (National Instruments, Austin, TX) data acquisition software that recorded the data
in minute intervals.

132

Figure 18. Schematic of bench-scale AnMBR.

5.3.2 Inoculation and operational parameters
The bench-scale AnMBRs were inoculated with mixed liquor from our previous experimental
period (Chapter 4), which was sludge originally collected from a mesophilic (37
o
C) AnMBR
treating FW. The reactor setup from the previous experiment was maintained for both SP and
TP, i.e., the sludge from SP was used as inocula for the SP system in this study, and likewise for
the TP systems for the methane-phase (TP-MP) and acid-phase (TP-AP). The inocula had a total
solids (TS) concentration of 49 g L
-1
and a total volatile solids (TVS) concentration of 39.5 g L
-1
for
the SP, and 46 g L
-1
TS and 36.6 g L
-1
TVS for the TP-MP. The TP-AP had a TS and TVS concentration
of 76.3 g L
-1
and 69.2 g L
-1
, respectively. The feed FW was collected from Divert, Inc. and filtered
using 1 mm aluminum mesh to remove large particles and prevent clogging of bench-scale
reactor tubing. The feed FW had an average chemical oxygen demand (COD) of 123 ± 7 g COD L
-

133

1
and TS and TVS concentrations of 65.8 g L
-1
and 60.6 g L
-1
, respectively. The pH of the feed FW
was 3.5 and ammonia concentration was 1.92 ± 0.19 g NH 4-N L
-1
. To prevent biodegradation, the
feed was kept at 4
o
C inside a refrigerator which was directly connected to the feed line of the
AnMBRs. The FOG samples were collected from Baker Commodities (Vernon, CA), who operate
a facility that collects and treats FOG retrieved from restaurants’ grease traps. Two types of FOG
samples were collected, one dilute stream and one scum layer from the top of a storage tank.
Initially, to acclimate the reactors, the dilute FOG wastewater was used during co-digestion with
FW, at a total fats concentration of 0.5 kg m
3
·day
-1
. Subsequently, the fats concentration was
increased to 0.75 kg m
3
·day
-1
, using the same dilute FOG wastewater. After running in the
aforementioned conditions for a total 82 days, we switched to using the FOG retrieved from the
scum layer, and increased the fats loading to 1 kg m
3
·day
-1
. The operational parameters during
this phase are summarized in (Table 10). Fats concentration was increased incrementally from 1-
3 kg m
3
·day
-1
or until severe signs of inhibition were observed. During 0.5-0.75 kg m
3
·day
-1
, only
one TP-AP was operated where mixed FW and FOG were fed with HRT of 3 days. However, at
higher fats addition, mixing and floatation of a fat layer became increasingly difficult for normal
operations in both SP and TP-MP. Therefore, we divided the feeding to two streams, one with
concentrated FW and FOG mixture (28% fats) fed as a daily pulse, and the remaining FW stream
fed in-line by a pump connected to the AnMBRs (Figure 18). We also divided the TP-AP to two
digesters accordingly, referred hereafter as TP-AP1 and TP-AP2, with one operated on the
concentrated FOG and FW mixture and one operated only on FW. Both TP-AP digesters were
monitored independently to elucidate any difference in performance and microbial community.

134

Table 10. Operating conditions for SP and TP-MP.
Operating Conditions Value units
OLR 5 gCOD LD
-1

Temperature 37
o
C
HRT 40 day
-1

SRT 100 days
Membrane Area 0.011 m
2

Flux 0.5 L (m
2
h)
-1

Reactor Volume 5.2 L
MLVSS 19.11 g L
-1

MLTSS 28.6 g L
-1

Feed (FW) COD 102.7 gCOD L
-1

Sparging 1 m
3
m
2
·h
Backwash 3 min every 10 min

5.3.3 Chemical assays and sampling
Reactor performance was monitored by evaluating permeate characteristics, including COD,
volatile fatty acids (VFAs), sulfate, ammonia, phosphate, biogas production, biogas methane and
carbon dioxide concentration, total solids (TS) and total volatile solids (TVS) concentration. TS
and TVS were determined using the procedures outlined in Standard Methods (APHA 2005).
Permeate samples were filtered with 0.2 µm nylon membrane filters (Whatman, Pittsburgh, PA)
to measure soluble constituents (chemical oxygen demand (COD), ammonium, volatile fatty acids
(VFAs), etc.). The Nessler Method (APHA 2005) was used to determine ammonia concentrations,
whereas VFAs (formic acid, acetic acid, propionic acid, butyric acid, and valeric acid) and other
ions (nitrate and sulfate) were determined using ion chromatography (ICS-2000, Dionex,
Sunnyvale, CA) equipped with a refrigerated auto-sampler (Thermo Scientific, NY, USA). The
column flow was set at 0.50 mL/min and KOH was used as eluent. Chromatographic separation
was achieved using a 2 mm AS-11HC (Dionex, Sunnyvale, CA). The composition of biogas was
measured using the Trace 1310 GC system (Thermo Scientific, NY) equipped with a flame

135

ionization detector using hydrogen as carrier gas, where a TG-BOND Q 30m x 0.53mm x 20 µm
column was used for chromatographic separation.
Fats were analyzed using the acid hydrolysis method that targets all fatty acids, triglycerides,
esters, long chain alcohols, hydrocarbons, and other glycol esters and sterols. Hydrochloric acid
was added and samples were incubated at 70°C. After, we extracted crude fats manually using
diethyl and petroleum ether as solvents. This method was optimized for our samples, by
repeating the solvent addition in 4 cycles (with 40 mL diethyl and petroleum ether used each
cycle), to ensure complete extraction of fats. We also quantified five LCFAs (palmitic, stearic,
myristic, linoleic, and oleic acids) using the Trace 1310 GC system (Thermo Scientific, NY)
equipped with a flame ionization detector (FID) using hydrogen as carrier gas, where a TG-
WAXMS A 30m x 0.32mm x 0.25μm column was used for chromatographic separation. Calibration
standards for each LCFA were run at 25 - 100 mg L
-1
and C-14 labelled oleic acid was used as an
internal standard.
5.4 Results and Discussion
We acclimated the sludge in both SP and TP-MP using dilute FOG wastewater with total fat
addition of 0.5 - 0.75 kg m
-3
·day
-1
(Figure 19). High methane content was observed for all fats
additions, ranging from 64-69% in SP and 66-70% in TP-MP. The results showed high variability
in methane production per day during this phase. After 82 days of acclimation, we increased the
fats addition to 1 kg m
-3
·day
-1
by blending FW and FOG from scum layer samples. Biogas
production during this phase showed high variability. This could be caused by LCFA adsorption
onto biomass and degrading after a lag-phase, as has been reported by another study that

136

evaluated AnMBRs in treatment of ethanol thin stillage wastewater (Dereli et al. 2014). This
variability could also be due to inadequate mixing in the feeding bottles due to floatation of fats.
Therefore, we elected to modify our feeding regime thereafter by applying a daily pulse feed of
FOG and remaining FW fed continuously. The TP-AP was also divided to two digesters TP-AP1
(FW) and TP-AP2 (FW+FOG) to accommodate this change in feeding.
The performance results at 1 kg m
-3
·day
-1
fats loading showed a 12% increase in methane
production in TP-MP relative to SP (Figure 20), but the difference in means were not statistically
significant using a two-tailed t-test (p>0.05) due to high variability in daily methane production
in both reactors. The average daily methane production in SP and TP-MP for this condition were
7.3 ± 0.8 (95% Confidence interval (C.I)), and 8.2 ± 0.8 (95% C.I.) L CH 4 day
-1
, respectively.
Conversely, statically significant difference in means was found when comparing FW mono-
digestion vs. co-digestion of FW and FOG, for both SP and TP-MP at 1 kg m
-3
·day
-1
(Figure 21). In
SP, significant difference in means was also observed for the 0.75 kg m
-3
·day
-1
fats feeding
condition, where average daily methane production increased by 6.4% with co-digestion relative
to mono-digestion. At 1 kg m
-3
·day
-1
fats feeding condition, this increase was even higher, with
11.3% increase in methane production in co-digestion. In TP-MP, there was a decline in methane
production at 0.5 kg m
-3
·day
-1
and similar production at 0.75 kg m
-3
·day
-1
fats feeding conditions,
however, these difference in means for both feeding conditions were not found statistically
significant compared to the mono-digestion results. At 1 kg m
-3
·day
-1
fats feeding condition,
where significant difference in means was observed, the increase in methane production in co-
digestion was 13.2% compared to mono-digestion in TP-MP. These results indicate that in TP-
MP, where methane recovery was already high with mono-digestion, fats addition did not alter

137

the recovery potential significantly for the two initial feeding conditions, but in SP treatment,
where the energy recovery with mono-digestion was lower, FOG addition was more impactful,
even at lower concentrations.

Figure 19. Daily methane production (primary y-axis) and biogas composition (secondary y-axis)
for at (A) 0.5 kg m
3
·day
-1
,

(B) 0.75 kg m
3
·day
-1
,

and (C) 1 kg m
3
·day
-1
fats loading rates.

A.
B.
C.

138

Figure 20. Methane production rate (primary y-axis) and methane per OLR (secondary y-axis).
Solid fill shows SP and pattern fill signifies TP-MP. The numerical value after SP or TP-MP shows
the fats loading for each system, for example TP-MP - 0.5 indicates, the mean methane
production for the TP-MP treatment at 0.5 kg m
3
·day
-1
. The solid error bar indicates 95%
confidence interval of the mean, and the circles indicate standard deviation of mean daily
methane production for each OLRs.
The mass balance analysis indicated that more COD went to biomass growth and biomass
wasted at 0.5 kg m
3
·day
-1
fats feeding condition in TP-MP (Figure 22). This could also be observed
in the TVS concentration results in TP-MP, which were higher compared to SP in this initial phase
(Figure 23). However, with increasing fats addition, TVS accumulation decreased in both AnMBRs.
The pH remained stable throughout this study in SP, TP-MP, TP-AP1 and TP-AP2, with an average
of 7.9 ± 0.2, 8.0 ± 0.2, 3.8 ± 0.2, and 3.3 ± 0.1, respectively (Figure 24). Ammonia concentrations
also remained stable throughout operation, with concentration of 1270 ± 170 in SP and 1620 ±
160 in TP-MP. Effluent COD concentration was low at all feeding condition, with removal
efficiency >99% (Figure 25).

139

Figure 21. Methane production rate (primary y-axis) and methane per OLR (secondary y-axis).
Solid dark blue fill shows SP and solid dark red fill signifies TP-MP. The clear fill show mono-
digestion of FW only for SP (blue) and TP-MP (red). The number after SP or TP-MP shows the
fats loading for each system, for example TP-MP - 0.5 indicates, the mean methane production
for the TP-MP treatment at 0.5 kg m
3
·day
-1
fats addition. The solid error bar indicates 95%
confidence interval of the mean, and the circles indicate standard deviation of mean daily
methane production for each OLRs. The symbol * shows that mean methane production were
statistically significantly different between FW mono-digestion vs. FW and FOG co-digestion for
the specified reactor, as indicated by <0.05 p values with two-tailed t-test.

Figure 22. (A) Mass balance analysis for SP based on COD allocation of output relative to input
(%), at different fats addition. (B) Mass balance analysis for TP-MP based on COD allocation of
output relative to input (%), at different fats addition. Complete sulfate reduction was assumed
based on influent sulfate concentration.

140

Figure 23. Total volatile solids (TVS) concentration (primary y-axis), signified by solid line, and
total solids (TS) concentration (secondary y-axis) signified by broken line for single-phase (SP)
and two-phase methane-phase (TP-MP), at different fats loading. Error bars for TVS and TS
concentrations represent the standard deviation for duplicate samples.

Figure 24. pH in SP, TP-MP, TP-AP1, and TP-AP2 at different fats loading rates.

141

Figure 25. Chemical oxygen demand (COD) concentration in effluent (primary y-axis) and COD
removal efficiency (secondary y-axis) for different FOG addition for single-phase (SP) and two-
phase methane-phase (TP-MP).

One of the main concerns in high FOG addition is microbial inhibition caused by LCFA
accumulation. Therefore, we plan to monitor LCFAs in sludge and effluent samples. This will
also aid to determine whether LCFAs are being degraded through microbial process or are being
precipitated with cations (such as Ca
2+
and Mg
2+
), similar to reports from a previous study
(Dereli et al. 2014). In addition, future analyses of microbial community data will help elucidate
the changes due to FOG addition, where we expect to see increase in communities that can
degrade LCFAs, i.e., syntrophic fatty acid β-oxidizing bacteria.
Overall, the results show that FOG and FW co-digestion can increase performance compared to
mono-digestion of FW. Although, we observed 4 - 12% increase in performance in TP-MP

142

compared to SP at FOG addition of 0.5 - 1 kg m
-3
·day
-1
fats., the difference in means were not
found statistically significant, due to high variability in daily methane production data.
Accidentally air exposure in SP and TP-MP during 1 kg m
-3
·day
-1
fats feeding stage led to reactor
failure after 60 and 36 days, respectively, and reactors were not able to be recovered. This
indicates that both reactors showed susceptibility to shock condition at 1 kg m
-3
·day
-1
fats
feeding rate. Overall, this study showed that FOG addition could be an effective operational
strategy that can increase performance in AnMBRs treating FW.

5.5 Literature Cited
Alqaralleh, R.M., Kennedy, K., Delatolla, R. and Sartaj, M. (2016) Thermophilic and hyper-thermophilic
co-digestion of waste activated sludge and fat, oil and grease: Evaluating and modeling methane
production. Journal of environmental management 183, 551-561.
Amha, Y.M., Sinha, P., Lagman, J., Gregori, M. and Smith, A.L. (2017) Elucidating microbial community
adaptation to anaerobic co-digestion of fats, oils, and grease and food waste. Water research.
APHA (2005) Standard Methods for the Examination of Water and Wastewater, American Public Health
Association, Washington, D.C.
Dereli, R., Urban, D., Heffernan, B., Jordan, J., Ewing, J., Rosenberger, G. and Dunaev, T. (2012)
Performance evaluation of a pilot-scale anaerobic membrane bioreactor (AnMBR) treating ethanol thin
stillage. Environmental technology 33(13), 1511-1516.
Dereli, R.K., Heffernan, B., Grelot, A., van der Zee, F.P. and van Lier, J.B. (2015) Influence of high lipid
containing wastewater on filtration performance and fouling in AnMBRs operated at different solids
retention times. Separation and Purification Technology 139, 43-52.
Dereli, R.K., van der Zee, F.P., Heffernan, B., Grelot, A. and van Lier, J.B. (2014) Effect of sludge retention
time on the biological performance of anaerobic membrane bioreactors treating corn-to-ethanol thin
stillage with high lipid content. Water research 49, 453-464.
Galib, M., Elbeshbishy, E., Reid, R., Hussain, A. and Lee, H.-S. (2016) Energy-positive food wastewater
treatment using an anaerobic membrane bioreactor (AnMBR). Journal of environmental management
182, 477-485.
Jensen, P., Yap, S., Boyle-Gotla, A., Janoschka, J., Carney, C., Pidou, M. and Batstone, D. (2015) Anaerobic
membrane bioreactors enable high rate treatment of slaughterhouse wastewater. Biochemical
engineering journal 97, 132-141.
Long, J.H., Aziz, T.N., Francis, L. and Ducoste, J.J. (2012) Anaerobic co-digestion of fat, oil, and grease
(FOG): a review of gas production and process limitations. Process Safety and Environmental Protection
90(3), 231-245.
Palatsi, J., Illa, J., Prenafeta-Boldú, F., Laureni, M., Fernandez, B., Angelidaki, I. and Flotats, X. (2010)
Long-chain fatty acids inhibition and adaptation process in anaerobic thermophilic digestion: batch tests,

143

microbial community structure and mathematical modelling. Bioresource Technology 101(7), 2243-
2251.
Ramos, C., García, A. and Diez, V. (2014a) Performance of an AnMBR pilot plant treating high-strength
lipid wastewater: Biological and filtration processes. Water research 67, 203-215.
Ramos, C., Zecchino, F., Ezquerra, D. and Diez, V. (2014b) Chemical cleaning of membranes from an
anaerobic membrane bioreactor treating food industry wastewater. Journal of Membrane Science
458(Supplement C), 179-188.
Sousa, D.Z., Salvador, A.F., Ramos, J., Guedes, A.P., Barbosa, S., Stams, A.J.M., Alves, M.M. and Pereira,
M.A. (2013) Activity and viability of methanogens in anaerobic digestion of unsaturated and saturated
long-chain fatty acids. Applied and environmental microbiology 79(14), 4239-4245.

144

CHAPTER 6

6. Conclusions

6.1 Overview
The primary objective of this dissertation was to advance energy recovery from food waste (FW)
using anaerobic biotechnologies. Two barriers that impact widespread application of anaerobic
treatment of FW are low energy recovery that make it difficult to justify the high initial
investment, and susceptibility to inhibition due to the variable nature of FW as substrate. Thus,
the approach of the dissertation was to assess operating and design strategies that can increase
performance and resilience of anaerobic treatment of FW. The first work focused on compiling
recent literature on inhibition in anaerobic systems, with emphasis on the use of advanced
molecular tools to elucidate microbial community changes during inhibition (Chapter 2; (Amha
et al. 2018)). Next, co-digestion of FW and fats, oils, and grease (FOG) with wastewater treatment
plant (WWTP) sludges was evaluated as an operational strategy to increase energy recovery
(Chapter 3; (Amha et al. 2017)). The following work looked at a design approach, i.e., two-phase
treatment, to increase energy recovery in anaerobic membrane bioreactors (AnMBRs) solely
treating FW (Chapter 4; (Amha et al. 2019)). Lastly, co-digestion of FOG in AnMBRs designed to

145

treat FW was investigated as an operational strategy to increase performance in single- and two-
phase AnMBRs (Chapter 5).
6.2 Diverse molecular tools are elucidating inhibitor impact on the AD microbiome
Inhibition of anaerobic digestion (AD) could be caused by changes in substrate addition and
operating conditions. Early detection of inhibitory conditions in AD systems could reduce risk of
reactor failure. To this end, advanced molecular tools enable characterization of community
changes, which in turn could be used as early warning signs of inhibitory conditions. The review
of recent literature in inhibition studies indicated that reported inhibitory concentrations are
highly variable among different studies for the same inhibitor, likely stemming from differences
in inoculum and microbial community adaptation. We compiled data from studies that
investigated inhibition caused by ammonia, long chain fatty acids (LCFAs), volatile fatty acids
(VFAs), and other inhibitors (such as humic acid, OLR, etc.). Metadata analysis comparing
microbial community structure across various studies that investigated inhibition caused by
accumulation of LCFAs and VFAs identified communities that showed similar trends with
increasing concentration. Further, the core microbiome was identified for these two inhibitor
groups based on communities that showed persistent abundance at high inhibitor concentrations
across different studies. We reported that most inhibition studies were published without
adequate metadata information, and we made recommendations on minimum information that
should be included in future publications. Further, we critically reviewed how recent
advancement in molecular tools, such as such as ‘omics’ tools, substrate mapping, and real-time
sequencing, can enhance our understanding of how microbial communities respond to
perturbations. We also emphasized that future advances in molecular tools should focus on

146

developing predictive abilities to avoid system failure. Thus, understanding shifts in key
populations during stress conditions will be important information in developing real-time
monitoring tools.
6.3 Co-digestion of FW and FOG can increase energy recovery in WWTPs
Co-digestion is an operational strategy that can increase energy recovery in WWTPs by increasing
organic loading rate. Few studies have evaluated how co-digestion impacts the microbial
community of AD. In this work, we ran three sequential bench-scale respirometry experiments
to evaluate FW and FOG co-digestion. First, various combinations of primary sludge (PS),
thickened waste activated sludge (TWAS), FOG, and FW were run. Simultaneous co-digestion of
FOG and FW gave the highest improvement in methane production compared to digestion of PS
and TWAS, which was 26% increase in methane production at run 3. The results also showed that
FOG co-digestion resulted in significant increase in lag-phase, which was reduced by more than
half in run 3. FOG addition showed higher impact in improving performance than FW in all three
runs, where methane production increased by up to 21% compared to digestion of PS and TWAS
in run 3. Co-digestion of FW showed no impact in run 1, decline in run 2, and increase of 18%
methane production in run 3, relative to digestion of PS and TWAS, suggesting that the microbial
communities needed to adapt to FW addition. To test the optimum loading of FOG, two
additional runs were performed where FOG addition was sequentially increased from 30% to 60%
in volatile solid loading (VSL; v/v). The results showed that 30% FOG addition resulted in the
highest methane production increase (53%) compared to digestion of only PS and TWAS. FOG
addition above 50% VSL resulted in more than 90% reduction in methane production, signaling
severe inhibition.

147

The microbial community shifts with substrate addition were characterized using DNA- and RNA-
based Illumina sequencing. In addition, a targeted approach was used to quantify methyl
coenzyme M reductase (mcrA) genes, a functional gene of methanogens. The overall microbial
community analyses showed that hydrogenotrophic methanogens dominated in all substrate
feeding conditions. In the high FOG addition runs, syntrophic fatty acid oxidizers showed high
increase in relative activity in the 30% and 40% FOG conditions. This increase was highest in the
30% FOG condition, where syntrophic fatty acid oxidizers increase in run end by 34-fold
compared to run initial. The most active group in 30% FOG conditions was Syntrophomonas, a
syntrophic fatty acid β-oxidizer that can degrade LCFAs. Spearman rank correlation showed that
relative activity of Syntrophomonas was significantly correlated with methane production, and
total activity of syntrophic fatty acid oxidizers showed stronger correlation with methane than
total activity of methanogens. The 50% and 60% FOG conditions resulted in severe inhibition in
activity of syntrophs and methanogens. The absence in activity of syntrophic β-oxidizers at >50%
FOG conditions indicate that LCFA accumulation led to inhibition of syntrophs and methanogens,
ensuing reactor failure. The targeted analysis of mcrA genes also confirmed the observations with
16S rRNA sequencing, showing severe inhibition of methanogens at >50% FOG feeding
conditions. The study demonstrated the benefits of co-digestion in terms of performance
enhancement and enrichment of key active microbial populations.
6.4 Two-phase AnMBRs increased energy recovery and microbial community resiliance at
high organic loading rates
FW treatment with AnMBRs increases energy recovery from FW by decoupling sludge retention
time and hyrulic retention time, while producing superior effluent quality relative to non-

148

membrane based anaerobic processes. This is particularly interesting for full-scale plants that
have both a FW and wastewater stream, which is the case with our collaborator for this work,
Divert Inc. All the FW that was not able to be sold or donated from grocery marts, Ralph’s and
Food4Less, is brought to a distrbution center located in Compton. The company is able to offset
25-30% of the distribution center’s overall energy demand and save money on surchage fees for
wastewater discharge to the sewer line.
In this study, we investegated phase-separtion as a strategy to increase microbial diversity,
thereby improving performance. Initially, two identical bench-scale systems (SP1 and SP2) were
operated at identical conditions until stable performance was achieved. After one month start-
up period, methane production in the two reactors was significantly similar (two tail t-test,
p=0.71), with average daily methane production for SP1 and SP2 at 4.24 ± 0.38 and 4.20 ± 0.36 L
d
-1
, respectively. After, SP2 was converted from single- to two-phase AnMBR (with acid-phase
(TP-AP) and methane-phase (TP-MP)), whereas SP1 was kept as single-phase AnMBR. Organic
loading rate (OLR) was then incrementally increased from 2.5 to 15 g COD∙L day
-1
. Volatile fatty
acids (VFAs) accumulated in TP-AP, with acetate and propionate constituting the majority of VFAs
detected. Initially, propionate showed higher concentration than acetate, but acetate later
emerged as the dominant VFA at 3.5 g COD∙L day
-1
OLR and above. The microbial community
analsysis revealed a unique microbial community in TP-AP compared to FW samples, where
communities that did not show high activity in FW samples, such as Aeriscardovia, showed
dominance in the TP-AP reactor. VFA concentration was enriched in TP-AP by up to four fold
compared to FW samples.

149

The two-phase system showed increased methane production relative to the single-phase
system as OLR increased. At 10 gCOD∙L day
-1
methane production for two-phase was 20.3%
higher than single-phase. Effluent COD concentration remained low throughout the experiments,
with removal efficiency >98% throught the different feeding stages in both SP and TP-MP. The
mass balance analysis indicated that more COD went to biomass growth and volatile solids
accumulation in the single-phase AnMBR compared to the two-phase system, signalling that
hydrolysis was not being effectively done in SP. In addition, the performance results indicated
that there was inhibition in SP at 10 gCOD∙L day
-1
OLR rate, where methane per OLR fell below
0.25 L CH 4 gCOD
-1
. At 15 gCOD∙L day
-1
, both SP and TP-MP showed sever inhibitory conditions,
with ammonia and free ammonia concentrations exceeding 2000 mg L
-1
and 200 mg L
-1
,
respectively. In addition, both reactors showed accumlation of volatile solids and methane per
OLR below 0.24 L CH 4 gCOD
-1
at this feeding stage.
The community analysis showed that micribial diversity remained stable in TP-MP at all feeding
stages, while it showed steep decline with increasing OLR in SP. This indicated that the distinct
biochemical environments provided via TP treatment increased microbial diversity, resulting in
improved system performance and resilence to increased OLR. Non-meteric multi-dimensional
analysis (NMDS) showed that there was statistically distinct clustering between SP and TP-MP
microbial community profile. Further, syntrophic fatty acid oxidizers showed higher activity in TP-
MP than SP, while methanogens showed similar activity in both reactors. For example, at 10 g
COD∙L day
-1
, where the highest differences in performance between SP and TP were observed,
syntrophs showed two-fold greater activity in TP-MP compared to SP. However, methanogens
showed similar activity in both systems at 10 g COD∙L day
-1
, indicating that methaogens had

150

redundant functionality in our system and changes in methanogenic activity were not a major
factor in the performance difference between the two reactors. As inhibitory conditions
increased in both systems with increasing OLRs, the communities that were phased out were
more effectively replaced in TP-MP compared to SP, showing that adaptation played a factor in
the relativiely stable microbial diversity and higher resilence observed in TP-MP. As there are no
preceding studies evaluating microbial dynamics of AnMBR systems treating food waste in high
resolution, this study will be helpful to inform decisions at the full-scale on design and operational
parameters that enhance performance.
6.5 Co-digestion of FOG improves performance of anaerobic membrane bioreactor treatment
of food waste
In this study, we evaluated the use of FOG co-digestion as an operational strategy to increase
performance in AnMBRs treating FW. High FOG addition could also cause accumulation of LCFAs
that could be inhibitory. We investigated the upper limits of FOG addition without sever microbial
inhibition by increasing FOG incrementally, where feeding stages were defined by the total fats
added as substrate, which were 0.5, 0.75, and 1 kg fats m
-3
·day
-1
. SP and TP reactors were used
to evaluate the impact of reactor design. Although TP showed higher methane production than
SP (4-12%), the difference between the mean daily methane production in SP and TP was not
statistically significant for all three feeding stages (two tailed t-test: p>0.05). This was due to the
high variability in daily methane production for both systems. We theorized that LCFA adsorption
onto biomass and long lag-phase in degradation could have caused the high variability in the
systems. Comparing performance of SP with only FW (mono-digestion) vs. with FOG co-digestion,
we observed a statistically significant increase in mean daily methane production at 0.75 and 1

151

kg fats m
-3
·day
-1
feeding stages, which was 6.4% and 11.3% improvement, respectively. TP
systems only showed statistically significant increase with co-digestion at 1 kg fats m
-3
·day
-1

(p=0.0243), which was 13.2% increase compared to the mono-digestion of FW. SP and TP-MP
showed complete and irreversible inhibition after 60 and 36 days of operation at 1 kg fats m
3
·day
-
1
after an accidental air exposure, respectively. This indicates that both systems are vulnerable
to operational changes at these feeding conditions. Notably, we did not observe severe
membrane fouling throughout the experiment, showing that addition of FOG did not negatively
impact membrane performance. Overall, the study showed that FOG co-digestion can be an
operational strategy that can increase methane production in AnMBRs treating FW.
6.6 Future research
This dissertation work looked at various topics that can increase performance and resilence of
anaerobic systems designed to treat FW. The work done here is important and timely, given the
increasing legislative push in California to divert FW from landfills. It is expected that the number
of new systems in California designed to treat FW will increase, therefore, the findings of this
work can guide future and current projects on how they can maximize both profitability and
environmental sustainablity of such systems.
This dissertation work focused on the use AD and AnMBRs to treat FW and recover energy by
producing methane. However, there are also other high value products that can be recovered
from FW via other anaerobic processess, such as medium chain fatty acids (Spirito et al. 2014),
VFAs (Wang et al. 2014), and biohydrogen (Han and Shin 2004). Therefore, future work should
utilze system-level analysis to understand how ADs/AnMBRs compare with other alternative

152

treatment systems and what products will maximize the value recovery from FW. This kind of
comparion should include economic analysis and environmental impact assessment over the full
life-cycle of these different products.
Further, the FW used here is unique in that it is grocery FW mixed with wastewater from a
creamery, which impacted the microbial community profile (Chapter 4). Monitoring the feed and
mixed liquor community over 6 months at the full-scale plant did not show major temporal shifts.
However, future studies should compare FW from various sources, to understand how the type
of FW source might impact results. Notably, the FW used in this study does not include meat
products, which are typically used for animal feed. This could be a factor that could impact the
results reported in this study, particularly in determining the maximum OLR (Chapter 4).
The study on FOG addition for enhanced energy recovery showed reactor failure during 1 kg
fats∙m
-3
day
-1
, both for SP and TP AnMBRs, due to short-term accidental air exposure. This
indicates that at this fats loading rate, both reactor are highly vulnerable to operational
disturbances. Future studies should investigate how to increase the resilence of AnMBRs at high
fats loading rates. An example of an intervention could be regular alkanity dosing to delay shock
responses of pH drop in the system.
The study on phase-seperation in AnMBRs (Chapter 4) indicated that biofilm developed on the
membrane could have resulted in higher effluent quailty. VFAs concentration in effluent were
lower than VFAs in biomass at high OLRs. Further, the COD removal efficiency was >98% even at
elevated OLRs. Therefore, future studies should systematically characterize the role of biofilm

153

treatment in AnMBRs treating FW, by comparing different levels of biofilm development and the
impact on effluent quality.
More advanced approaches in upgrading methane production should be further studied both in
AD and AnMBRs. One strategy for this could be promoting direct interspecies electron transfer
(DIET), an electron exchange mechanism that does not require diffusive molecules (H 2 and
formate) for electron transfer (Summers et al. 2010). DIET has the potential to enhance
performance because traditional electron transfer between syntrophic bacteria and
methanogens can be rate limiting in anaerobic communities (Stams 1994). Although there are
very limited studies that have demonstrated a combined AnMBR-bioelectrochemical system
(BES) in wastewater treatment (Ge et al. 2013, Tian et al. 2014), there are no studies that have
applied an anaerobic membrane bio-electrochemical reactor (AnMBER) to treat high strength
substrates, such as food waste. Therefore, future studies should demonstrate the use of AnMBER
systems to increase energy reovery from FW.
Lastly, with the ambitious goal for Los Angeles to reach >90% reduction in landfilled solid waste
by 2025, as part of RENEW LA goals (Murphy and Pincetl 2013), more policy changes are expected
to be introduced. Therefore, it is important to understand at the system-level what is the best
approach to realize reduction and complete diversion of FW. Particularly, collection and
treatment of residential FW will be challenging, in terms of enforcing source seperation and
disposal. Further, efforts to reduce FW production by improving the food supply-chain system,
and the impact on existing or planned treatment plants should also be analyzed, as it will impact
the total FW that needs to treated.

154

6.7 Literature Cited
Amha, Y.M., Anwar, M.Z., Brower, A., Jacobsen, C.S., Stadler, L.B., Webster, T.M. and Smith, A.L. (2018)
Inhibition of anaerobic digestion processes: applications of molecular tools. Bioresource Technology
247, 999-1014.
Amha, Y.M., Corbett, M. and Smith, A.L. (2019) Two-phase improves performance of anaerobic
membrane bioreactor treatment of food waste at high organic loading rates. Environmental science &
technology (Submitted).
Amha, Y.M., Sinha, P., Lagman, J., Gregori, M. and Smith, A.L. (2017) Elucidating microbial community
adaptation to anaerobic co-digestion of fats, oils, and grease and food waste. Water research.
Ge, Z., Ping, Q. and He, Z. (2013) Hollow-fiber membrane bioelectrochemical reactor for domestic
wastewater treatment. Journal of Chemical Technology & Biotechnology 88(8), 1584-1590.
Han, S.-K. and Shin, H.-S. (2004) Biohydrogen production by anaerobic fermentation of food waste.
International journal of hydrogen energy 29(6), 569-577.
Murphy, S. and Pincetl, S. (2013) Zero waste in Los Angeles: Is the emperor wearing any clothes?
Resources, Conservation and Recycling 81, 40-51.
Spirito, C.M., Richter, H., Rabaey, K., Stams, A.J. and Angenent, L.T. (2014) Chain elongation in anaerobic
reactor microbiomes to recover resources from waste. Current opinion in biotechnology 27, 115-122.
Stams, A.J. (1994) Metabolic interactions between anaerobic bacteria in methanogenic environments.
Antonie van Leeuwenhoek 66(1), 271-294.
Summers, Z.M., Fogarty, H.E., Leang, C., Franks, A.E., Malvankar, N.S. and Lovley, D.R. (2010) Direct
exchange of electrons within aggregates of an evolved syntrophic coculture of anaerobic bacteria.
Science 330(6009), 1413-1415.
Tian, Y., Ji, C., Wang, K. and Le-Clech, P. (2014) Assessment of an anaerobic membrane bio-
electrochemical reactor (AnMBER) for wastewater treatment and energy recovery. Journal of
Membrane Science 450, 242-248.
Wang, K., Yin, J., Shen, D. and Li, N. (2014) Anaerobic digestion of food waste for volatile fatty acids
(VFAs) production with different types of inoculum: effect of pH. Bioresource Technology 161, 395-401.

155

APPENDIX A

A. Supplementary Information

SI 1.0 Chapter 2
SI 1.1 Tables

Table S1. Concentration ranges used to define inhibitory levels of LCFAs and VFAs in metadata
analysis.

Category Concentration
LCFA None Inoculum/Control
Low < 3000 mg VS/L
Medium 3000 - 4000 mg VS/L
High > 4000 mg VS/L
VFA None to Low Inoculum/Control
Low to Medium < 100 mg/L
High 100-150 mg/L
Inhibitory > 150 mg/L

156

Table S2. Advantages and limitations of different molecular method.
Advantages Disadvantages/Limitations
Amplicon based 1. Shifts in microbial
diversity over
time/concentration or
specific factor e.g.
LCFA/VFA
2. Reference based and
Denovo comparison both
possible
1. PCR/Primer Bias
2. Cannot predict specific metabolic
function
3. 16s copy number variation doesn’t allow
absolute abundance and thus relative
abundance is used which can be
misleading in certain scenarios where
the community dynamics are changing
Metagenomics 1. No PCR/Primer bias
2. Sufficient Sequencing
depth
3. Functional potential of the
community
1. Characterizes communities based on
phylogeny and functional potential, not
gene expression
2. Low signal/abundance or rare
populations can be challenging to
estimate
Metatranscriptomics 1. Profiling of community-
wide gene expression
2. Gene expression
abundance
3. Differential gene
expression analysis
1. Low amount of mRNA quantification
results in less than 5% of mRNA
2. Cannot predict post translational
regulation
Metaproteomics 1. Expressed proteins
characterized thus
enabling prediction of
Post-translational
regulation
2. Identifying novel
functional systems and
obtaining direct functional
insights
3. Metaproteomics coupled
with metabolomics can
help in measuring novel
information on microbial
activity and system
functioning
1. Challenging to extract sufficient
amounts high quality protein
2. Mostly relied on metagenomics unless
denovo peptide sequencing is available

157

SI 1.2 Figures

Figure S1. Community structure by class over increasing inhibitory concentration of volatile fatty acids and long chain fatty acids (left
to right). The y-axis represents relative abundance of OTUS that are 0.1% or greater in the community. Stacked bars within each class
(same color) represent orders.

158

SI 2.0 Chapter 3 1
SI 2.1 Materials & Methods 2
Chemical analyses and biogas production 3
Total solids (TS) and total volatile solids (TVS) were determined in the seed, PS, TWAS, FW, and 4
FOG samples using the procedures outlined in Standard Methods (Rice et al. 2012). Samples were 5
filtered with 0.2 µm nylon membrane filters (Whatman, Pittsburgh, PA) to measure soluble 6
constituents (chemical oxygen demand (COD), ammonium, volatile fatty acids (VFAs), etc.). 7
Ammonium concentration was determined using the phenate method (Searle, 1984). VFAs 8
(formic acid, acetic acid, propionic acid, butyric acid, and valeric acid) and other ions (nitrate and 9
sulfate) were determined using ion chromatography (ICS-2000, Dionex, Sunnyvale, CA) equipped 10
with a refrigerated auto-sampler (Thermo Scientific, NY, USA). The column flow was set at 0.50 11
mL/min and KOH was used as eluent. Chromatographic separation was achieved using a 2 mm 12
AS-11HC (Dionex, Sunnyvale, CA). 13
Biogas production was monitored in-line and recorded at 10-minute intervals. The composition 14
of produced biogas was analyzed periodically, using the Trace 1310 GC system (Thermo Scientific, 15
NY) equipped with a thermal conductivity detector (TCD). Biogas samples were taken from the 16
vessel headspace using a syringe. A TG-BOND Q 30m x 0.53mm x 20 µm column was used for 17
chromatographic separation. The injection and oven temperatures were set at 250
o
C and 30
o
C, 18
respectively. The vessels were run until the cumulative biogas production plateaued. At the end 19
of each run, biomass samples were taken for analysis of TS, TVS, ammonium, VFAs, and sulfate. 20
Free ammonia concentration was calculated using equation (1) (Hansen et al. 1998). 21
[𝑁 𝐻 3
]
[𝑇𝑁 𝐻 3
]
= (1 +
10
−𝑝𝐻
10
−(0.09018+
2729.92
𝑇 ( 𝑘 )
)
)
−1
(1) 22
where [𝑁 𝐻 3
] is the concentration of free ammonia, [𝑇𝑁 𝐻 3
] is total ammonia concentration, 23
T(k) is temperature in K. 24

159

The use of non-linear regression models has been shown to be more suitable to accurately 25
describe and predict cumulative methane production than linear regression methods (Li et al. 26
2011). Therefore, the modified Gompertz equation was used as follows (Lo et al. 2010), 27
Where 𝑦 is the biogas accumulation (L/kg) at time t (d). µ
𝑚 is the maximal biogas production 28
rate (L/kg.d) (Lo et al. 2010). A is the biogas production potential (L/kg), and λ is the lag phase 29
(d). 30
Nucleic acids extraction and cDNA synthesis 31
Biomass samples were taken at the beginning and end of each run, centrifuged at 5,000 x g for 5 32
min at 4
o
C, decanted, and preserved at -80
o
C until further processing. DNA extraction from 33
preserved biomass was performed by three 2-min bead beating steps (Mini-Beadbeater-96, 34
BioSpec Products, Bartlesville, OK) with 0.1 mm diameter zirconium beads in lysis buffer, 35
proteinase K digestion, and automated extraction using the Maxwell 16 Blood LEV kit according 36
to manufacturer’s instruction (Promega, Madison, WI). RNA extraction was performed by three 37
1-min bead beating steps with 0.1 mm diameter silicon beads in 1-thiolyglycerol homogenization 38
buffer and automated extraction using the Maxwell 16 simplyRNA blood kit according to 39
manufacturer’s instructions except that 10 µL of DNase 1 (instead of 5 µL) was used to ensure 40
complete DNA degradation. For both assays, approximately 0.2 g of the pelletized biomass was 41
taken for extraction. DNA quality and quantity was assessed via spectrophotometry 42
(BioSpectrometer Fluorescence, Eppendorf, Hamburg, Germany) and Quant-iT™ PicoGreen
®
43
dsDNA Assay (Invitrogen, Carlsbad, CA), respectively. For the RNA extracts, an additional DNase 44
treatment was conducted using, DNA-free™ DNA Removal Kit (Invitrogen, Carlsbad, CA), to 45
remove DNA contamination. DNA contamination was subsequently checked via quantitative 46
polymerase chain reaction (qPCR) targeting the 16S rRNA gene, which confirmed that no 47
amplification occurred up to a quantification cycle (Cq) of 40 cycles. The DNase treated RNA 48
extracts were quantified using Quant-iT™ RiboGreen
®
RNA Assay. Reverse transcription to 49
generate single-stranded complementary DNA (cDNA) from RNA extracts was performed using 50
the GoScript™ Reverse Transcription System according to manufacturer’s instructions (Promega, 51
Madison, WI). 100 ng of RNA was taken from each sample for cDNA synthesis. 52

160

PCR and sequencing 53
Due to the potential for differences in microbial community structure and activity within replicate 54
vessels under the same substrate mixtures, triplicate DNA and RNA samples for Run 1, and 55
triplicate DNA and duplicate RNA samples for Run 3 were taken for analysis (Table S3). PCR, 16S 56
rRNA gene sequencing, and 16S rRNA (cDNA transcript) sequencing were conducted using a 57
universal 16S rRNA gene primer set targeting the V4 region (Caporaso et al. 2011) barcoded and 58
using sequencing primers described in Kozich et al. (Kozich et al. 2013). PCR reactions were 20 µL 59
and included 100 nM of each primer, 17 µL AccuPrime™ Pfx SuperMix (Invitrogen, Carlsbad, CA), 60
0.5 ng template, and nuclease-free water. 61
Thermocycling conditions consisted of an initial 2 min denaturation at 95°C, followed by 30 cycles 62
of denaturing at 95°C for 20 s, annealing at 55°C for 15 s, and extension at 72°C for 5 min, followed 63
by a final extension at 72°C for 5 min. Amplicons were pooled by equal mass using the 64
SequalPrep
TM
Normalization Plate Kit (Life Technologies, Grand Island, NY). Multiplexed 65
amplicons were sequenced via Illumina MiSeq using the MiSeq Reagent Kit V2 (2x250 bp reads) 66
and sequencing primers described in Kozich et al. (Kozich et al. 2013) at the University of 67
Michigan. The sequencing results were analyzed using mothur (Schloss et al. 2009), according to 68
the Schloss MiSeq SOP. The results from the first phase of the study and the FOG inhibition study 69
were analyzed separately. Quality filtering and chimera removal using the UCHIME algorithm was 70
conducted to generate high quality reads. After quality filtering, an average of 13,200 ± 3,690 71
paired-end sequences per sample were obtained for the first three runs of the study, with 72
minimum and maximum sequences of 7,560 and 24,400. For the FOG inhibition study (Run 4 and 73
5), the average number of sequences was 12,200 ± 2750 per sample, with minimum of 5,880 and 74
maximum of 21,200. 7,560 and 5,880 paired-end reads per sample were generated after 75
subsampling for the first phase of the study and the FOG inhibition study, respectively. The SILVA 76
reference database (Pruesse et al. 2007) was used to align sequences, and classification was 77
conducted using Ribosomal Database Project (Maidak et al. 2001). Operational taxonomic unit 78
(OTU)-based clustering was conducted with average neighbor algorithm at a 3% cutoff. Non- 79
dimensional multidimensional scaling (NMDS) was used to characterize similarity/dissimilarity 80

161

within the different samples. In addition, Spearman rank was used for non-parametric analysis 81
to correlate sequence data with methane production and VFAconcentration with MaxStat 3.6 82
(Germany). For the Spearman rank correlation analysis, only vessels with substrate addition were 83
considered (excluding controls) and the analyses for the first phase of the study and FOG 84
inhibition study were conducted separately. 85
Reverse transcription-quantitative PCR (RT-qPCR) 86
The 16S rRNA and mcrA transcripts were quantified using RT-qPCR. Positive controls for mcrA 87
and 16S rRNA were isolated from a pool of 10 cDNA samples used in the study and pooled by 88
equal mass. PCR to prepare the RT-qPCR standards was performed as previously described (Smith 89
et al. 2015). Forward and reverse primers for mcrA quantification included mlas and mcrA-rev 90
primers (Steinberg and Regan 2008). The 16S rRNA was quantified via the 515F and 806R 91
targeting the V4 region (Caporaso et al. 2011). Isolated PCR products were run on an agarose gel, 92
purified with the Wizard® SV Gel and PCR Clean-Up System (Promega, Madison, WI), and 93
quantified with Quant-iT™ PicoGreen (Invitrogen, Carlsbad, CA), after which serial dilutions of 94
10
8
to 10
1
copies of each gene were prepared. RT-qPCR of mcrA transcripts was performed using 95
conditions described in (Smith et al. 2015). Genomic DNA of Thermus Thermophilus isolated from 96
pure culture (American Type Culture Collection (ATCC) 27634) was used as a negative control, in 97
addition to a no-template control. Melt curve analysis was conducted to check for specificity of 98
amplifications. The R
2
value for the mcrA standard curve was 1.00 and the average efficiency was 99
84.0%. The R
2
value for the 16S rRNA standard curve was also 1.00 with an average efficiency of 100
78.0%. 101
Reagent Control 102
A freeze dried culture of Thermus thermophilus (ATCC 27634) was purchased from ATCC and 103
propagated on LB broth according to ATCC’s recommendations. After incubation for 48 h at 70
o
C, 104
a confirmed positive culture was used to make six, ten-fold serial dilutions from the starter 105
culture in fresh LB. For the highest dilution sample (denoted as 0), 1 mL of the culture was taken 106

162

directly, whereas the subsequent dilution aliquots (denoted serial dilutions 1-6) were made by 107
diluting 1:10 the preceding aliquot with pure LB broth, for each dilution level. Subsequently, DNA 108
extractions were conducted on the dilution aliquots with the aforementioned extraction 109
methods. In addition to the dilution reagent controls, blank LB and ultra-pure DNase/RNase free 110
water were also taken for sequencing. Three of the six dilutions, serial dilutions 1-3, had to be 111
excluded from sequencing analysis, as only 12, 9 and 8 paired-end reads were generated, 112
respectively. This could be due to underrepresentation of sequences from these samples in the 113
library pool during the normalization step. Further, due to high variability in reads per sample, 114
5,412 ± 8,049 sequences, subsampling was not performed. 115
The resulting 16S rRNA gene sequences for the various dilutions analyzed showed that DNA 116
contamination was ubiquitous, being present in extraction kits, laboratory reagents, and the 117
laboratory environment, and the impact of contamination was higher in samples with lower DNA 118
concentration (Figure S1), confirming similar observations made recently (Salter et al. 2014). 119
Pelomonas, Methylobacterium, Sphingomonas, Bradyrhizobium, and unclassified 120
Enterobacteriaceae were commonly detected in multiple samples. These populations were also 121
detected as contaminant genera by Salter et al. (Salter et al. 2014). Diversity, as measured by the 122
inverse Simpson metric, was highest in the fifth and sixth serial dilutions. In addition, qPCR was 123
conducted on DNA and reverse transcribed cDNA from RNA extracts of the different dilutions of 124
Thermus thermophilus to quantify the observed contamination. The DNA results showed 292 and 125
595 16S rRNA gene copies/µL for LB broth and the ultrapure water used in this study, 126
respectively, whereas 2,930 and 960 16S rRNA copies/µL were quantified for LB broth and 127
ultrapure water, respectively. Therefore, contamination was more severe with RNA than DNA 128

163

extracts. In addition, the dilution analysis using Thermus thermophilus showed that log quantity 129
decreased linearly until the fourth serial dilution for DNA and the third serial dilution for RNA 130
samples (Figure S3). 131
Reference 132
Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-Lyons, D., Lozupone, C.A., Turnbaugh, P.J., Fierer, N. 133
and Knight, R. (2011) Global patterns of 16S rRNA diversity at a depth of millions of sequences per 134
sample. Proceedings of the National Academy of Sciences 108(Supplement 1), 4516-4522. 135
Hansen, K.H., Angelidaki, I. and Ahring, B.K. (1998) Anaerobic digestion of swine manure: inhibition by 136
ammonia. Water research 32(1), 5-12. 137
Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K. and Schloss, P.D. (2013) Development of a dual- 138
index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq 139
Illumina sequencing platform. Applied and environmental microbiology 79(17), 5112-5120. 140
Li, C., Champagne, P. and Anderson, B.C. (2011) Evaluating and modeling biogas production from 141
municipal fat, oil, and grease and synthetic kitchen waste in anaerobic co-digestions. Bioresource 142
Technology 102(20), 9471-9480. 143
Lo, H., Kurniawan, T., Sillanpää, M., Pai, T., Chiang, C., Chao, K., Liu, M., Chuang, S., Banks, C. and Wang, 144
S. (2010) Modeling biogas production from organic fraction of MSW co-digested with MSWI ashes in 145
anaerobic bioreactors. Bioresource Technology 101(16), 6329-6335. 146
Maidak, B.L., Cole, J.R., Lilburn, T.G., Parker Jr, C.T., Saxman, P.R., Farris, R.J., Garrity, G.M., Olsen, G.J., 147
Schmidt, T.M. and Tiedje, J.M. (2001) The RDP-II (ribosomal database project). Nucleic acids research 148
29(1), 173-174. 149
Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J. and Glöckner, F.O. (2007) SILVA: a 150
comprehensive online resource for quality checked and aligned ribosomal RNA sequence data 151
compatible with ARB. Nucleic acids research 35(21), 7188-7196. 152
Rice, E.W., Bridgewater, L., Association, A.P.H., Association, A.W.W. and Federation, W.E. (2012) 153
Standard Methods for the Examination of Water and Wastewater, American Public Health Association. 154
Salter, S.J., Cox, M.J., Turek, E.M., Calus, S.T., Cookson, W.O., Moffatt, M.F., Turner, P., Parkhill, J., 155
Loman, N.J. and Walker, A.W. (2014) Reagent and laboratory contamination can critically impact 156
sequence-based microbiome analyses. BMC biology 12(1), 1. 157
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley, 158
B.B., Parks, D.H. and Robinson, C.J. (2009) Introducing mothur: open-source, platform-independent, 159
community-supported software for describing and comparing microbial communities. Applied and 160
environmental microbiology 75(23), 7537-7541. 161
Smith, A.L., Skerlos, S.J. and Raskin, L. (2015) Membrane biofilm development improves COD removal in 162
anaerobic membrane bioreactor wastewater treatment. Microbial biotechnology 8(5), 883-894. 163
Steinberg, L.M. and Regan, J.M. (2008) Phylogenetic comparison of the methanogenic communities from 164
an acidic, oligotrophic fen and an anaerobic digester treating municipal wastewater sludge. Applied and 165
environmental microbiology 74(21), 6663-6671. 166
167

164

SI 2.2 Figures

Figure S2. Relative abundance of genera identified with 16S rRNA gene sequencing for reagent control samples that were serially
diluted using genomic DNA of Thermus thermophilus. CD0, CD4, CD5, and CD6 indicate undiluted genomic DNA of Thermus
thermophilus, 10
-4
dilution with ultrapure reagent water, 10
-5
dilution, and 10
-6
dilution, respectively. CLB and CW are control LB and
ultrapure reagent water, respectively. CD1, CD2, and CD3 are not shown due to low sequencing depth. All sequencing results are
shown in percentage normalized to the total 16S rRNA gene sequencing (Bacteria and Archaea).

165

Figure S3. 16S rRNA gene and 16S rRNA quantification using quantitative PCR (qPCR) and reverse-transcription RT-qPCR, respectively.
The error bar shows the standard deviation for triplicate runs for each sample. For the highest dilution sample (denoted as 0), 1 mL
of Thermus thermophilus culture in Lysogeny broth (LB) was taken directly, whereas the subsequent dilution aliquots (denoted
dilution-level 1-6) were made by diluting 1:10 the preceding aliquot with pure LB broth, for each dilution level. In addition, LB and
ultrapure DNase/RNase free water samples were taken for qPCR and RT-qPCR analyses.

166

Figure S4. (A) Non-linear regression fitting for PS+TWAS, (B) PS+TWAS+FW, (C) PS+TWAS+FW,
and (D) PS+TWAS+FOG+FW substrate mixtures for Run 2. Vessel 1-3 represent the triplicate
substrate mixtures for the run. A is the biogas production potential (L/kg), λ is the lag phase (d),
and the R
2
value shows the non-linear regression coefficient.
C.
D.
B. A.

167

Figure S5. Genus-level classification of samples collected from Hyperion Wastewater Treatment Plant (seed, primary sludge (PS),
thickened waste activated sludge (TWAS), and fats, oils, and grease (FOG)), and from Divert Inc. (food waste (FW)). The DNA-based
and RNA-based sequencing results are shown for all samples, expect for FOG samples, where only results with RNA-based sequencing
are shown.

168

Figure S6. Relative activity based on 16S rRNA sequencing classified at the phyla level for Run
1-3 end of run samples with various substrate mixture. All data are expressed as a percentage
normalized using total 16S rRNA sequences (Bacteria and Archaea). A y-axis break was used to
accentuate differences in lower activity populations.

169

Figure S7. Relative abundance based on 16S rRNA gene sequencing classified at the phyla level
for Run 1-3 end of run samples with various substrate mixture. All data are expressed as a
percentage normalized using total 16S rRNA sequences (Bacteria and Archaea). A y-axis break
was used to accentuate differences in lower activity populations.

170

Figure S8. (A) Relative abundance based on 16S rRNA gene sequencing identified at the genus level where possible at the end of Run
3. (B) Relative abundance in PS+TWAS+FOG+FW for Run 1, Run 2, and Run 3. For Run 1 and Run 3, triplicate and duplicate samples
are shown, respectively, to represent methodological precision. All data are expressed as a percentage normalized using total 16S
rRNA gene sequences (Bacteria and Archaea). A y-axis break was used to accentuate differences in lower activity populations.
A. B.

171

Figure S9. (A) Relative abundance of methanogens identified at the genus level where possible based on 16S rRNA gene sequencing
and (B) Relative abundance of syntrophic fatty-acid oxidizers identified at the genus level where possible using 16S rDNA sequencing.
Results are expressed as a percentage normalized using total of 16S rRNA sequences (Bacteria and Archaea). Truncated y-axes (0 to
0.25% and 0 to 4% on figure A and B, respectively) are shown to accentuate differences in abundance.
A.
B.

172

Figure S10. Relative activity of syntrophic acetate oxidizers for Run 1-3, using 16S rRNA
sequencing. Results are expressed as percentages normalized using the total number of 16S
rRNA sequencing (including Bacteria and Archaea). A truncated y-axis (0 to 1.6%) is shown to
accentuate differences in abundance.

173

Figure S11. Methane production for all substrate mixtures in Run 1-3 vs. relative activity of
methanogens and syntrophic fatty-acid oxidizers using 16S rRNA sequencing. The relative
activity data were normalized using total 16S rRNA sequences (including Archaea and Bacteria).
The regression lines are only to aid in visualization of trend and not to suggest goodness of fit.

174

Figure S12. Non-metric multidimensional scaling (NMDS) plot for end of Run 1 samples using
(A) 16S rRNA sequencing and (B) 16S rRNA gene sequencing. The error bars for both plots show
the standard deviation for triplicate samples for each substrate mixture.

A.
B.

175

Figure S13. Inverse Simpson diversity metric for the total relative activity (Bacteria and Archaea)
based on 16S rRNA sequencing. Error bars are shown only for Run 1 results and represent the
standard deviation of the inverse Simpson metric for triplicate substrate mixtures (vessels).

176

Figure S14. Cumulative biogas production normalized to initial organic loading in mL/gTVS for
Run 5. The error bars indicate the standard deviation every 40 h for triplicate vessels. One of
the triplicate vessels for the 40% FOG+FW condition was excluded due to an outlier biogas
production compared to the remaining replicate substrate mixtures.

177

Figure S15. (A) Relative activity and (B) relative abundance of top 30 most abundant OTUs identified to the genus level for beginning
and end of Run 4. Results are expressed as percentages normalized using the total number of 16S rRNA sequences and 16S rRNA
gene sequences for A and B, respectively (including Bacteria and Archaea).

B.
A.

178

Figure S16. Relative abundance of top 30 most abundant OTUs identified to the genus-level for beginning and end of Run 5 using
16S rRNA gene sequencing. Results are expressed as percentages normalized using the total number of 16S rRNA gene sequencing
(including Bacteria and Archaea). End of run results shown are from triplicate vessels for each substrate mixture.

179

Figure S17. (A) Relative abundance of methanogens identified at the genus level where possible based on 16S rRNA gene sequencing.
(B) Relative abundance of syntrophic fatty-acid oxidizers identified at the genus level where possible using 16S rRNA gene
sequencing. Results are expressed as a percentage normalized using total 16S rRNA sequences (Bacteria and Archaea). Truncated y-
axes (0 to 0.3% and 0 to 9% on A and B, respectively) are shown to accentuate differences in abundance.
B.
A.

180

Figure S18. Non-metric multidimensional scaling (NMDS) plot for initial and end of
Run 5 samples using 16S rRNA gene sequencing. The error bars show the standard
deviation for triplicate samples for end of run results for each substrate mixture.

181

Figure S19. Comparison of relative activity of methanogens and syntrophic fatty-acid
oxidizers detected in Run 5 with methane production. The total relative activity of
methanogens and syntrophs identified using 16S rRNA sequencing were normalized
using the total number of 16S rRNA sequencing (including Bacteria and Archaea).
Duplicate results for each substrate mixture is shown. The error bar indicates the
standard deviation for methane production where all triplicate vessels for each
substrate mixture were considered (including the outlier biogas production for the
40% FOG substrate mixture).

182

SI 2.3 Tables
Table S3. Biomass samples taken for microbial community analyses. The raw samples
were seed, primary sludge (PS), thickened waste activated sludge (TWAS), food waste
(FW) and fats, oils, and grease (FOG). DNA samples are shown in blue and RNA samples
are shown in orange.
Environmental samples
DNA RNA
Raw
First Study

Beginning of Run End of Run

DNA RNA DNA RNA
Run 1

( )x3

Run 2

Run 3

( )x3

FOG inhibition study
Run 4

Run 5

( )x3

PS TWAS
Seed
FW FOG

PS TWAS Seed FW FOG

183

Table S4. Spearman rank correlation analysis for Run 1 – 3, showing significant
correlation (defined as p<0.05) between methane production and microbial community
activity.
Genera positively
correlated
R
2
p Genera negatively
correlated
R p
Methanoculleus 0.749 <
0.0001
Dictyoglomus -0.673 0.0003
Clostridium_XI 0.655 0.0005 unclassified Bacteria -0.610 0.002
Clostridium_sensu_stricto 0.571 0.004

Clostridiales 0.547 0.006

Soehngenia 0.537 0.007

Petrimonas 0.531 0.008

Methanosarcina 0.521 0.009

Lactobacillus 0.509 0.011

Clostridium_IV 0.498 0.013

Turicibacter 0.493 0.014

Porphyromonadaceae 0.492 0.015

Syntrophomonadaceae 0.453 0.026

Syntrophaceticus 0.424 0.039

Tepidanaerobacter 0.419 0.042

Syntrophomonas 0.406 0.050

Total methanogens 0.475 0.019

Total syntrophs 0.475 0.019

184

SI 3.0 Chapter 4
SI 3.1 Tables

Table S5. Experimental conditions for different organic loading rates (OLRs).

+
initial identical AnMBRs also run at same conditions
FW= Food Waste; SRT=Sludge Retention Time; HRT= Hydraulic Retention Time; SP= Single-Phase; TP-AP=
Two-phase acid-phase; TP-MP= Two-phase methane-phase

Table S6. FW characterization.
Item Value units
pH 3.5

Acetate 3980 ± 550 mg L
-1

Propionate 1810 ± 680 mg L
-1

Formate 5.5 ± 0.9 mg L
-1

Phosphate 426 ± 12 mg L
-1

Sulfate 85.8 ± 15.7 mg L
-1

Chloride 956 ± 47 mg L
-1

Nitrate 50.5 ± 13.9 mg L
-1

TS 65.8 ± 1.0 g L
-1

TVS 60.1 ± 0.9 g L
-1

COD 123 ± 6.6 g L
-1

Protein 1
Wt%

Fat 1.5
Wt%

Carbohydrates 3.5
Wt%

185

Moisture 94
%

C:N Ratio 16

SI 3.2 Figures

Figure S20. Schematic of TP system consisting of TP-AP and methane-phase (TP-MP). In
single-phase (SP), FW was directly fed (no acid-phase) and it consisted of the
components to the right of the vertical dashed line.

186

Figure S21. Volatile fatty acids (VFAs) concentrations in Two-Phase Acid-Phase (TP-AP)
at different Organic Loading Rates (OLRs). The solid line represents VFA concentration
(primary y-axis) and the dashed line represents ratio of VFAs to Chemical Oxygen
Demand (COD) in the FW (secondary y-axis). Error bars for VFA concentrations represent
the standard deviation for triplicate samples.

187

Figure S22 (A). pH in Single-Phase (SP), Two-Phase Methane-Phase (TP-MP), and Two-
Phase Acid-Phase (TP-AP) at different Organic Loading Rates (OLRs). (B). Total Ammonia-
Nitrogen concentration in Single-Phase (SP), Two-Phase Methane-Phase (TP-MP), and
Two-Phase Acid-Phase (TP-AP) at different Organic Loading Rates (OLRs).

188

Figure S23. Non-metric multi-dimensional analysis (NMDS) plot showing ordination of
Food Waste (FW) and Two-Phase Acid-Phase (TP-AP) samples (at different Organic
Loading Rates (OLRs)) analyzed using DNA- and RNA-based sequencing.

189

Figure S24. Relative activity of microbial communities in Food Waste (FW) and Two-
Phase Acid-Phase (TP-AP) based on 16S rRNA sequencing, identified to the genus level
where possible. FW samples were retrieved from the full-scale plant, Divert Inc., weekly
in December, 2017 (denoted as Sample 1-4). TP-AP samples were from the different
Organic Loading Rates (OLRs) at bench-scale and the numbers on the x-axis represent
days after startup of the AnMBR. All data are expressed as a percentage normalized
using total 16S rRNA sequences (Bacteria and Archaea). A y-axis break (at 63%) was used
to accentuate differences in lower activity populations.

190

Figure S25. (A) Relative activity of methanogens identified at the genus level where
possible based on 16S rRNA sequencing and (B) relative activity of syntrophic fatty-acid
oxidizers identified at the genus level where possible using 16S rRNA sequencing.
Results are expressed as a percentage normalized using total of 16S rRNA sequences
(Bacteria and Archaea). Truncated y-axes (0 to 0.07% and 0 to 0.03% on figure A and B,
respectively) are shown to accentuate differences in abundance.

191

Figure S26. Mass-balance analysis for SP based on CID allocation of output relative to
input (%), at different OLRs. (B). Mass-balance analysis for TP-MP based on COD
allocation of output relative to input (%), at different Organic Loading Rates (OLRs).
Complete sulfate reduction was assumed based on influent sulfate concentration.

192

Figure S27. Total Volatile Solids (TVS) concentration (primary y-axis), signified by solid
line, and Total Solids (TS) concentration (secondary-axis) signified by broken line for
Single-Phase (SP) and Two-Phase Methane-Phase (TP-MP), at different Organic Loading
Rates (OLRs). Error bars for TVS and TS concentrations represent the standard deviation
for duplicate samples.

193

Figure S28. (A) Volatile Fatty Acids (VFAs) concentration in Single-Phase (SP) for effluent
(solid-line) and biomass (dashed-line) samples at different Organic Loading Rates (OLRs).
(B). VFAs concentration in Two-Phase Methane-Phase (TP-MP) for effluent (solid-line)
and biomass (dashed-line) samples at different Organic Loading Rates (OLRs). Error bars
for VFAs concentrations represent the standard deviation for triplicate samples.

194

Figure S29. Chemical oxygen demand (COD) concentration in effluent (primary axis) and
COD removal efficiency (secondary axis) for different Organic Loading Rates (OLRs) for
Single-Phase (SP) and Two-Phase Methane-Phase (TP-MP).

195

Figure S30. Relative abundance (primary y-axis) and relative activity (secondary y-axis)
of microbial communities in mixed liquor (ML) samples form full-scale plant (Divert Inc.)
based on 16S rRNA gene and 16S rRNA sequencing, respectively. The microbial
communities were identified to the genus level where possible. ML samples were
retrieved from Divert monthly (June-November, 2017) and weekly on the month of
(denoted as Sample 1-4). Monthly relative abundance data (June - November) are
average values for triplicate samples. All data are expressed as a percentage normalized
using total 16S rRNA gene sequences (Bacteria and Archaea) and 16S rRNA sequences
(Bacteria and Archaea) for relative abundance and relative activity data, respectively.

196

Figure S31. Non-metric multi-dimensional analysis (NMDS) plot showing ordination of
microbial community structure (DNA-based) or activity (RNA-based) for SP, TP-MP, and
TP-AP samples. The different fills for SP and TP-MP samples indicate different Organic
Loading Rates (OLRs). The star symbols signify Biofilm samples from SP and TP-MP. The
table shows two-tailed t-test to test the hypothesis that the clustering of ordination in
between different groups is significant, where the p-value<0.05 indicates significant
differential clustering.

197

Figure S32. Heat-map showing log Relative Activity (%) of microbial communities in SP
based on 16S rRNA sequencing, identified to the genus level where possible. All data are
expressed as a percentage normalized using total 16S rRNA sequences (Bacteria and
Archaea).

198

Figure S33. Heat-map showing log Relative Activity (%) of microbial communities in TP-
MP based on 16S rRNA sequencing, identified to the genus level where possible. All data
are expressed as a percentage normalized using total 16S rRNA sequences (Bacteria and
Archaea).

199

Figure S34. Relative abundance (for DNA) and relative activity (for RNA) of biofilm
communities in SP and TP-MP at an OLR of 10 g COD L·day
-1
based on 16S rRNA
sequencing, identified to the genus level where possible. All data are expressed as a
percentage normalized using total 16S rRNA sequences (Bacteria and Archaea).

Abstract (if available)

Abstract Food waste (FW) is one of the largest components of municipal solid waste (MSW), with an estimated 40% of all food produced in the U.S. being wasted from farm to fork to landfill. Anaerobic biotechnologies offer an opportunity to recover energy via production of methane-rich biogas, while also enabling reuse of the embedded nutrients in FW in the form of biosolids. However, low energy production and vulnerability of these systems to disturbances are two barriers in their application for FW management at the full scale. In this dissertation, operating and design strategies were investigated to increase energy recovery and resilience of anaerobic systems to potentially inhibiting conditions. First, co-digestion of FW and fats, oils, and grease (FOG), which also originate from food-processing activities, was evaluated as a strategy to increase energy recovery at wastewater treatment plants. A synergetic increase in methane production of 26% was observed when FW and FOG were co-digested with wastewater sludges. RNA-based sequencing indicated that syntrophic fatty-acid oxidizers had the greatest influence on system performance. Next, the potential for two-phase (acid/methane) anaerobic membrane bioreactor (AnMBR) treatment of FW was investigated. We systematically compared single-phase (SP) and two-phase (TP) AnMBR treatment of FW and characterized the impact of phase separation on microbial community structure and activity profiles at incrementally increasing organic loading rates (OLRs). The TP system increased methane production (up to 20.3%) relative to the SP system as OLR increased from 3.5 to 10 g COD L·d-1. At high OLR, activity of syntrophic bacteria in TP was double that of SP. Our results indicated that AnMBRs in TP mode can effectively treat FW at OLRs up to 10 g COD∙L day-1 by improving hydrolysis rates, microbial diversity, syntrophic activity, and enriching more resistant communities to high OLRs relative to AnMBRs in SP mode. Last, FOG co-digestion was evaluated to improve energy recovery during AnMBR treatment of FW in SP and TP systems, where the upper limits of FOG addition without microbial inhibition and severe membrane fouling were determined. Anaerobic biotechnologies are expected to play a major role in FW management in the near future and thus, this body of work provides timely guidance on operational strategies and design for full-scale anaerobic systems.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Bioelectrochemical treatment of anaerobic process effluents: mitigation of dissolved methane and sulfide

PDF

Fate of antibiotic resistance in anaerobic membrane bioreactors

PDF

Wastewater reclamation and potable reuse with novel processes: membrane performance and system integration

PDF

Thermally driven water treatment with membrane distillation: membrane performance, waste heat integration, and cooling analysis

PDF

Optimizing biomembrane reactor systems for water reclamation and reuse applications

Asset Metadata

Creator Amha, Yamrot Mulugeta (author)

Core Title Advancing energy recovery from food waste using anaerobic biotechnologies: performance and microbial ecology

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Engineering (Environmental Engineering)

Publication Date 07/31/2019

Defense Date 05/07/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag anaerobic biotechnologies,anaerobic digestion,anaerobic membrane bioreactor,ceramic membrane,co-digestion,DNA-based methods,energy recovery,fats, oils, and grease,food waste,Illumina sequencing,inhibition,methane,methanogens,microbial community,molecular methods,OAI-PMH Harvest,RNA-based methods,syntrophs,two-phase,wastewater treatment

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Smith, Adam L. (committee chair), Childress, Amy E. (committee member), McCurry, Daniel L. (committee member), Moffett, James (committee member), Sanders, Kelly T. (committee member)

Creator Email amha@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-202730

Unique identifier UC11662630

Identifier etd-AmhaYamrot-7707.pdf (filename),usctheses-c89-202730 (legacy record id)

Legacy Identifier etd-AmhaYamrot-7707.pdf

Dmrecord 202730

Document Type Dissertation

Format application/pdf (imt)

Rights Amha, Yamrot Mulugeta

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA