Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Biological interactions on the behavioral, genomic, and ecological scale: investigating patterns in Drosophila melanogaster of the southeast United States and Caribbean islands
(USC Thesis Other)
Biological interactions on the behavioral, genomic, and ecological scale: investigating patterns in Drosophila melanogaster of the southeast United States and Caribbean islands
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
BIOLOGICAL INTERACTIONS ON THE BEHAVIORAL, GENOMIC, AND
ECOLOGICAL SCALE:
INVESTIGATING PATTERNS IN DROSOPHILA MELANOGASTER OF THE
SOUTHEAST UNITED STATES AND CARIBBEAN ISLANDS
by
Joyce Y Kao
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(COMPUTATIONAL BIOLOGY AND BIOINFORMATICS)
August 2014
Copyright 2014 Joyce Y Kao
ii
Dedication
This dissertation is dedicated
to the memory of my grandmother,
Ann Yang,
to my loving parents,
Jem and Jing Kao,
and finally to my extraordinary husband,
Kjong Lehmann.
iii
Acknowledgements
First and foremost, I would like to thank my advisor, Sergey Nuzhdin, who valued
and supported (both intellectually and financially) my academic freedom
throughout my doctoral studies. I am grateful for his patience and infectious
enthusiasm for science, while I blundered through many not-so-successful
endeavors to eventually find my identity in the scientific community. His
generosity allowed me to not only broaden my horizons in a wide-range of
scientific topics, but also gave me the opportunity to experience many unique
adventures such as sailing on the Enigma (where the real enigma was why
everyone was always vomiting), death-defying barnacle collecting, mushroom
hunting, and hacking through cacti-ridden fields of Catalina with a machete
chasing after flies.
I would also like to thank the other members of my committee, Peter Ralph and
Kim Siegmund, as well as other faculty at the University of Southern California,
namely, Fengzhu Sun, Ting Chen, John Tower, and Michael Waterman for
donating their time and guidance throughout this journey to completing a Ph.D.
A special thanks goes to Mark Siegal and members of his lab at New York
University for allowing me to squat in their lab and keeping me intellectually
iv
stimulated in my last year of graduate school while I remotely finished my
doctoral studies from New York.
I would like to thank the various collaborators that made the content of this
dissertation possible. There are many people here to acknowledge:
The data collection in chapter 2 was a gargantuan joint effort from an army of
high school summer interns, and undergraduate student researchers-- Gagan
Kaur, Marianne Hom, Milana Grozdanich, Sartaaj Walia, Sumaiya Ahmed, Sean
Kang, Alex Hwang, Lauren Ishida, Brett Mathews, Albert Sung, Chris Kim, David
Kuo, Grace Liang, William Liao, Joel Ong, Tiffany Young, and Forrest Zhang.
The circadian assays in chapter 3 were performed at New York University in
Justin Blau’s lab with much instruction from Matthieu Cavey.
The analysis in chapter 4 was a team effort with help from Asif Zubair, Matt
Salomon, and Daniel Campo.
For the data collection in chapter 5, I would like to thank the Wrigley Marine
Science Center for the use of their facilities and equipment on the island as well
as the Catalina Island Conservancy for the training and use of a 4WD vehicle to
access remote parts of the island. I would also like to thank the
v
lab of Mariana Mateos at Texas A&M University, in particular, Lauryn and Caitlyn
Winter, for assistance in field collections and species identification of some
collected specimens. Thanks to Maxi Polihronakis Richmond at the UC San
Diego Drosophila species stock center for instruction on genitalia dissections as
well as consulting on identifying Drosophila species.
Finally, many, many, many thanks to the past and present members of the
Nuzhdin lab for their support and encouragement, but especially to Julia Saltz
and Brad Foley, who patiently taught me the essentials of fly wrangling in and out
of the lab. Many bottles of scotch were consumed on Fridays and at conferences
over very useful discussions on all things Drosophila, graduate school, and
beyond. I guess I learned a lot about scotch too. Thank you for those important
lessons. I would also like to specifically thank Wendy Vu for her friendship and
tolerance/patience towards me for listening to countless hours of complaining
about science when things did not go as planned, which was pretty much always.
A giant thank you to the amazing people too numerous to all name here, who
have had an impact on me during these formative years.
vi
Table of Contents
Dedication.. ........................................................................................................... ii
Acknowledgements ............................................................................................... iii
List of Tables ....................................................................................................... viii
List of Figures ....................................................................................................... xi
Abstract…… ........................................................................................................ xvi
Chapter 1: Introduction ......................................................................................... 1
1.1 Speciation ................................................................................................ 1
1.2 Reproductive barriers ............................................................................... 2
1.3 Secondary Contact and Hybrid Zones ..................................................... 3
1.4 A Secondary Hybrid Zone of Drosophila melanogaster ........................... 6
1.5 Organization of Body of Work .................................................................. 8
1.6 Chapter References ................................................................................. 9
Chapter 2: The Influence of Postmating Characters in the Incipient
Sexual Isolation of Cosmopolitan Drosophila melanogaster ........... 11
2.1 Introduction ............................................................................................ 11
2.2 Materials and Methodology .................................................................... 15
2.3 Results ................................................................................................... 25
2.4 Discussion .............................................................................................. 37
2.5 Chapter References ............................................................................... 39
Chapter 3: Circadian Rhythms of American and Caribbean Drosophila
melanogaster .................................................................................... 41
3.1 Introduction ............................................................................................ 41
3.2 Materials and Methods ........................................................................... 43
3.3 Results ................................................................................................... 46
3.4 Discussion .............................................................................................. 50
3.5 Chapter References ............................................................................... 52
vii
Chapter 4: Investigating the Sources and Extent of European and
African Admixture in North American Populations of
Drosophila melanogaster .................................................................. 54
4.1 Introduction ............................................................................................ 54
4.2 Materials and Methods ........................................................................... 57
4.3 Results ................................................................................................... 62
4.4 Discussion .............................................................................................. 79
4.5 Chapter References ............................................................................... 82
Chapter 5: Santa Catalina Island as a study system for interspecies
interactions: Island-wide survey of Drosophila species .................... 86
5.1 Introduction ............................................................................................ 86
5.2 Materials and Methods ........................................................................... 88
5.3 Results ................................................................................................... 91
5.4 Discussion .............................................................................................. 98
5.5 Chapter References ............................................................................... 99
References. ....................................................................................................... 101
Appendices. ....................................................................................................... 108
Appendix A: Supplemental for Postmating Project .................................... 108
Appendix B: Supplemental for Admixture Project ...................................... 118
viii
List of Tables
TABLE 2.1: List of locations and isofemale lines from those locations as
well as corresponding map numbers in FIGURE 2.1 .......................................... 16
TABLE 2.2: ANOVA Table for Full Model of Hatchability .................................... 30
TABLE 2.3: ANOVA Table for Reduced Model of Hatchability with no
Latitude ................................................................................................................ 31
TABLE 2.4: ANOVA table of model comparisons for hatchability ....................... 31
TABLE 2.5: Improved Log-rank Test Results ...................................................... 36
TABLE 4.1: Average F
ST
values between populations for chromosome 2
divided by regions 2L (below diagonal) and 2R (above diagonal) ....................... 69
TABLE 4.2: Average F
ST
values between populations for chromosome 3
divided by regions 3L (below diagonal) and 3R (above diagonal) ....................... 70
TABLE 4.3: Average F
ST
values between populations for chromosome X
(lower diagonal) and all autosomes (upper diagonal) ......................................... 70
TABLE 5.1: Species compositions at each sampling site from summer
2012 collections. .................................................................................................. 92
TABLE 5.2: Correlations and associated p-values of Drosophila species
on Catalina Island. Pearson’s r is in the top half of the table and
associated p-values are in the lower half of the table. Significant p-
values and associated correlations are highlighted in light grey. ........................ 94
TABLE A.1: Full model for egg counts with Latitude and Longitude ................. 108
TABLE A.2: Reduced model for egg counts without Latitude and
Longitude ........................................................................................................... 108
ix
TABLE A.3: ANOVA table of model comparisons for egg counts ..................... 108
TABLE A.4: Analysis of Deviance Table for Full Model of Short-term
Remating Rates. ................................................................................................ 111
TABLE A.5: Analysis of Deviance Table for Reduced Model of Short-
term Remating without Longitude or Latitude .................................................... 111
TABLE A.6: Analysis of Deviance Table for Short-term Remating Model
Comparison ....................................................................................................... 111
TABLE A.7: Analysis of Deviance Table for Full Model of Long-term
Remating Rates. ................................................................................................ 111
TABLE A.8: Analysis of Deviance Table for Reduced Model of Long-term
Remating without Longitude or Latitude ............................................................ 112
TABLE A.9: Analysis of Deviance Table for Long-term Remating Model
Comparison ....................................................................................................... 112
TABLE A.10: Analysis of Deviance Table for Reduced Model with
Female x Male Interaction Term ........................................................................ 112
TABLE A.11: Analysis of Deviance Table for Reduced Long-term
Remating Model Comparison with and without Female x Male Interaction
Term .................................................................................................................. 112
TABLE B.1: Average nucleotide diversity (π) and Tajima’s D for
autosomes and X chromosomes by population ................................................ 118
TABLE B.2: Average D' of chromosomal region 2L for each population .......... 123
TABLE B.3: Average D' of chromosomal region 3L for each population ........... 124
TABLE B.4: Average D' of chromosomal region 3L for each population .......... 124
TABLE B.5: Average D' of chromosomal region 3R for each population ......... 124
x
TABLE B.6: Average D' of chromosomal region 3R for each population ......... 125
TABLE B.7: Average D' of chromosome X for each population ....................... 125
xi
List of Figures
FIGURE 2.1: Map of locations used in postmating assays ................................. 17
FIGURE 2.2: Egg counts of females mated with American males and
Caribbean males. ................................................................................................ 26
FIGURE 2.3: Hatchability of females mated with American males and
Caribbean males. ................................................................................................ 29
FIGURE 2.4: Survival curves of females from line 20,17 .................................... 33
FIGURE 2.5: Survival curves of females from line H,25 ..................................... 33
FIGURE 2.6: Survival curves of females from line 13,34 .................................... 34
FIGURE 2.7: Survival curves of females from line 33,11 .................................... 35
FIGURE 2.8: Survival curves of females from line 40,23 .................................... 35
FIGURE 3.1: Boxplots of period of all isofemale lines arranged from
northernmost location to the southernmost location.. .......................................... 46
FIGURE 3.2: Boxplots of power of all isofemale lines arranged from
northernmost location to the southernmost location s. ........................................ 47
FIGURE 3.3: Activity of American and Caribbean grouped over a 24
hour period broken up by 5 minute bins. ............................................................. 48
FIGURE 3.4: Boxplots of total sleep of American and Caribbean during
lights on and lights off .......................................................................................... 49
FIGURE 3.5: Average sleep of American flies and Caribbean flies over a
24 hour period by 5 minute increments. .............................................................. 49
xii
FIGURE 4.1: Map of populations with number of whole genome
sequences and model of currently accepted migration history of D.
melanogaster. ...................................................................................................... 56
FIGURE 4.2: First and second principal components (eigenvectors) of
PCA with populations from Cameroon (CAM), Caribbean Islands (CAR),
France (FRA), Raleigh (RAL), southeast US (SEU), and Winters (WIN) ............ 64
FIGURE 4.3: Zoomed in view of North American plus French
populations, first and second principal components of PCA with
populations from Cameroon (CAM-not pictured), Caribbean Islands
(CAR), France (FRA), Raleigh (RAL), southeast US (SEU), and Winters
(WIN) ................................................................................................................... 65
FIGURE 4.4: First four principal components of PCA including
populations from Cameroon, Caribbean Islands, France, Raleigh,
southeast US, and Winters reveal that most variation explained is within
the Cameroon population .................................................................................... 66
FIGURE 4.5: First and second principal components (eigenvectors) of
PCA with the Cameroon population removed, but including populations
from Caribbean Islands (CAR), France (FRA), Raleigh (RAL), southeast
US (SEU), and Winters (WIN) ............................................................................. 67
FIGURE 4.6: First four principal components of PCA with the Cameroon
population removed, but including populations from Caribbean Islands,
France, Raleigh, southeast US, and Winters ...................................................... 68
FIGURE 4.7: ADMIXTURE results of chromosomal region 2L for K = 2
and K = 3 number of distinct populations. ........................................................... 72
FIGURE 4.8: ADMIXTURE results of chromosomal region 2R for K = 2
and K = 3 number of distinct populations. ........................................................... 72
FIGURE 4.9: ADMIXTURE results of chromosomal region 3L for K = 2
and K = 3 number of distinct populations. ........................................................... 73
FIGURE 4.10: ADMIXTURE results of chromosomal region 3R for K = 2
and K = 3 number of distinct populations. ........................................................... 73
xiii
FIGURE 4.11: ADMIXTURE results of chromosome X for K = 2 and K =
3 number of distinct populations. ......................................................................... 74
FIGURE 4.12: Painted 2L chromosomal region heatmap with
hierarchical clustering of individuals with telomeric region on the left and
centromeric region on right. ................................................................................. 75
FIGURE 4.13: Painted 2R chromosomal region heatmap with
hierarchical clustering of individuals with telomeric region on the right
and centromeric region on the left. ...................................................................... 76
FIGURE 4.14: Painted 3L chromosomal region heatmap with
hierarchical clustering of individuals with telomeric region on the left and
centromeric region on the right. ........................................................................... 76
FIGURE 4.15: Painted 3R chromosomal region heatmap with
hierarchical clustering of individuals with telomeric region on the right
and centromeric region on the right ..................................................................... 76
FIGURE 4.16: Painted X chromosomes heatmap with hierarchical
clustering of individuals with telomeric region on the right and
centromeric region on the left. ............................................................................. 77
FIGURE 4.17: Expected proportion of African ancestry for each
population by chromosomal region ..................................................................... 78
FIGURE 4.18: Average D' as a measure of linkage disequilibrium by
population and chromosome ............................................................................... 79
FIGURE 5.1: Map of collection sites on Santa Catalina Island. .......................... 89
FIGURE 5.2: Species composition at the Wrigley Marine Science Center
between summers. .............................................................................................. 95
FIGURE 5.3: Photos of unknown specimen ........................................................ 97
FIGURE A.1: Short-term (top) and long-term (bottom) remating rates of
females plotted against longitude ...................................................................... 109
xiv
FIGURE A.0.2: Short-term (top) and long-term (bottom) remating rates of
females plotted against latitude ......................................................................... 110
FIGURE A.3: Hazard curves of females from line 13,34 after
experiencing homotypic (solid line) or heterotypic (dashed line) matings ......... 113
FIGURE A.4: Hazard curves of females from line 20,17 after
experiencing homotypic (solid line) or heterotypic (dashed line) matings ......... 114
FIGURE A.5: Hazard curves of females from line 33,11 after
experiencing homotypic (solid line) or heterotypic (dashed line) matings ......... 115
FIGURE A.6: Hazard curves of females from line 40,23 after
experiencing homotypic (solid line) or heterotypic (dashed line) matings ......... 116
FIGURE A.7: Hazard curves of females from line H,25 after experiencing
homotypic (solid line) or heterotypic (dashed line) matings .............................. 117
FIGURE B.1: F
ST
between Caribbean and Cameroon populations across
each chromosomal region ................................................................................. 118
FIGURE B.2: F
ST
between Caribbean and Raleigh populations across
each chromosomal region ................................................................................. 119
FIGURE B.3: F
ST
between Caribbean and southeast United States
populations across each chromosomal region .................................................. 119
FIGURE B.4: F
ST
between Caribbean and Winters populations across
each chromosomal region ................................................................................. 119
FIGURE B.5: F
ST
between France and Cameroon populations across
each chromosomal region ................................................................................. 120
FIGURE B.6: F
ST
between France and Caribbean populations across
each chromosomal region ................................................................................. 120
FIGURE B.7: F
ST
between France and Raleigh populations across each
chromosomal region .......................................................................................... 120
xv
FIGURE B.8: F
ST
between France and southeast United States
populations across each chromosomal region .................................................. 121
FIGURE B.9: F
ST
between France and Winters populations across each
chromosomal region .......................................................................................... 121
FIGURE B.10: F
ST
between Raleigh and Cameroon populations across
each chromosomal region ................................................................................. 121
FIGURE B.11: F
ST
between Raleigh and Winters populations across
each chromosomal region ................................................................................. 122
FIGURE B.12: F
ST
between Caribbean and Cameroon populations
across each chromosomal region ..................................................................... 122
FIGURE B.13: F
ST
between southeast United States and Raleigh
populations across each chromosomal region .................................................. 122
FIGURE B.14: F
ST
between southeast United States and Winters
populations across each chromosomal region .................................................. 123
FIGURE B.15: F
ST
between Winters and Cameroon populations across
each chromosomal region ................................................................................. 123
xvi
Abstract
Biological systems are complicated webs of interactions at many different levels.
It is through the ebb and flow of these interactions do we get processes such as
molecular pathways or social behavioral systems, which can rapidly change
large populations of organisms over time. Speciation is the process by which
new species are formed and underlying this process are these intricate
interactions between cellular, behavioral, and ecological levels. To understand
the fundamentals of speciation, we study these interactions on multiple levels
and at different time points of the ‘speciation continuum’ where at one end we
have freely mating organisms and the other end we have reproductively blocked
separate species.
We present here a study of Drosophila melanogaster populations from the
southeast United States and Caribbean islands. These populations represent a
hybrid zone of secondary contact between cosmopolitan flies from Europe and
African-like flies from West Africa, which diverged over 10,000 years ago. With
the presence of previously established clines in premating reproductive barriers
(i.e. male courtship behavior, cuticular hydrocarbons, etc.), it was proposed that
these flies were undergoing incipient sexual isolation, which is the start of the
speciation process. We investigated putative postmating reproductive barriers of
remating rates, egg hatchability, and female longevity after
xvii
mating to assess the extent and influence of these barriers. We found remating
rates had no effect and female longevity after mating had varied and patchy
effects on several different female lines from our study area. Interestingly
enough, there existed a hatchability ‘dip’ where there was an area of lower
hatchability around the border of the southern US and the Caribbean islands
indicating possible presence of Dobzhansky-Muller incompatibilities. In addition
to these postmating barriers, we also investigated the circadian rhythms of the
southeast US and Caribbean flies and found that these populations are
consistent with an existing sleep cline on the east coast of the US with
increasing sleep with decreasing latitude. The differences of sleep are small, but
statistically significant and could potentially impact gene flow between the D.
melanogaster in the US and Caribbean islands.
To investigate African and European admixture, we sequenced the genomes of
our 23 isofemale fly lines from 12 locations in the southeast US and Caribbean
islands. We compared these genomes to previously sequenced genomes from
Raleigh, NC and Winters, CA from North America as well as to sequenced
genomes of D. melanogaster from Montpellier, France and Oku, Cameroon. The
genome sequences revealed up to 25% African ancestry present in the
Caribbean population and decreasing percentages of African ancestry with
xviii
increasing latitudes into the southeast United States. Given our results, we also
propose a westward expansion model of D. melanogaster in the United States.
Finally, in the spirit of studying speciation on the other end of the continuum, we
establish a new study system of many closely related interacting cactophilic
Drosophila species on Santa Catalina Island off the coast of southern California.
We report species found on the island as well as abundance and seasonality of
species compositions.
1
Chapter 1
Introduction
1.1 Speciation
Evolutionary change in a community of organisms starts at the most basic level,
which is at the level of DNA. Some alterations to a genome have no effect and
are neutral, some are beneficial, and some are deleterious. It is through these
genomic changes that organisms gain or lose advantage in the environment in
which they are living. At some point in evolutionary time, certain genomic
changes cause groups of organisms or rather populations, which once freely
mated with one another, to no longer be able to interbreed. The non-
interbreeding populations would be deemed new species (Coyne and Orr, 2004).
The process by which these new species arise is the very definition of speciation
(Coyne and Orr, 2004). This natural process is helped along with the formation of
reproductive barriers that reduce gene flow within or between populations.
Three main modes of speciation are currently considered in the field: allopatric,
parapatric, and sympatric (Coyne and Orr, 2004). These three
2
types of speciation are defined by the location of speciating populations in
relation to one another as well as different types of isolating barriers playing a
role in the process of divergences. In allopatric speciation, one population is
initially divided by geographic isolation. The subdivided isolated populations
undergo genotypic as well as phenotypic divergence due to different selective
pressures in their new environments and influences from genetic drift plus new
mutations. After independently evolving over many generations, when these two
populations unite again, they are no longer able to produced viable offspring
together as they once did before. Parapatric speciation occurs between
neighboring populations with initial moderate gene flow between them. In this
situation, there is no extrinsic barrier in terms of geography as is in the case of
allopatry. Gene flow is reduced between these by the development of intrinsic
isolating mechanisms such as behavioral differences in the example of mate
preference. When new species form from an ancestral population inhabiting the
same geographic region, this is called sympatric speciation.
1.2 Reproductive barriers
Regardless of the mode of speciation, all of these processes occur with the
evolution of reproductive barriers that reduce gene flow. These barriers in the
simplest form can be geographical isolation such as a
3
mountain range or highway. Reproductive barriers can also be mechanical such
as in the case of Japanese carabid beetles where diversifying selection of male
and female genitalia can result in death after mating due to injuries sustained
from ill-fitting types of genitalia (Sota and Kubota, 1998). Behavioral reproductive
barriers such as those presented with mate preference is often found in many
speciation processes.
Canonically, reproductive barriers have been classified with regards to when they
occur. Those barriers that play a role in sexually reproducing organisms before
the zygote is formed are referred to as pre-zygotic whereas those that occur after
the zygote is formed are termed post-zygotic. These barriers can furthermore be
subdivided before or after mating (i.e. the transfer of sperm) referred to as pre-
and post-mating barriers respectively. Pre- and post-zygotic barriers evolve at
similar rates in both allopatric and sympatric speciation events (Coyne & Orr,
1989).
1.3 Secondary Contact and Hybrid Zones
Harrison (1990, 1993) defines a hybrid zone as a region where genetically
distinct populations meet and mate, resulting in at least some offspring of mixed
ancestry. These hybrid zones are often identified in nature by
4
the presence of concurrent clines (i.e. many clines overlapping in the same
place). A cline is a character gradient (Huxley, 1938) meaning a feature (e.g.
allele locus, courtship behavior, etc.) that changes over some geographical area.
Hybrid zones can originate from one of two situations. The first of which is in situ
via natural selection on the alleles changing their frequencies in continuous
islands of populations. This kind of hybrid zone is usually positioned to
correspond to a sharp change in environmental factors and are usually termed
‘primary hybrid zones’. The second type of hybrid zone comes from two
allopatric populations which have been genetically differentiated from living in two
different environments coming together to interbreed by expansion or migration.
This type of hybrid zone is called a ‘secondary hybrid zone’ and when these two
allopatric populations come together to interbreed is termed ‘secondary contact’.
To distinguish a primary from a secondary hybrid zone is not a trivial task (Endler
1977), but most investigators regard secondary contact as the most probable
explanation for hybrid zones especially if there are multiple clines in several
characteristics such as in morphology, behavior, and allozymes or other genetic
data.
5
Generally, there are four major outcomes from secondary hybrid zones centering
around the fate of hybrids created by the mixture of two allopatric populations:
1) Hybrids can exist indefinitely if the hybrid zone is stable.
2) Hybrids may have lower fitness (hybrid dysgenesis) than their parental
populations and natural selection may favor alleles for postzygotic isolation,
which eventually would lead to the reduction of hybrids and complete
reproductive isolation of the two parental populations (i.e. allopatric populations
would become separate ‘species’ themselves).
3) Hybrids are healthy and have similar fitness as parental populations whereby
the two allopatric populations may freely interbreed and merge into one
population.
4) Hybrids develop reproductive isolation from the parental populations and
become a third new population. Usually this scenario only happens in a portion
of the hybrid zone.
These scenarios have been studied extensively (Servedio and Noor, 2003;
Harrison, 1993) with models. However, how these models fit reality in natural
situations is not fully understood. This may be attributed to a
6
few shortcomings such as lack of population genetic models and insufficient
amount of data to test the models. However, with the advent of next generation
sequencing (NGS) and advancing computational power, we are coming to an
age in science where these problems can be studied extensively using whole
genome data of many individuals of several populations (Seehausen et al.,
2014).
1.4 A Secondary Hybrid Zone of Drosophila melanogaster
An ideal model to study these genomic dynamics is the populations of D.
melanogaster living in the southeast United States as well as in the Caribbean
islands. These populations have been previously studied in terms of their pre-
zygotic reproductive isolating barriers (Yukilevich and True, 2008a; Yukilevich
and True, 2008b). About 10,000 years ago, a subpopulation of African flies
migrated into Europe, which nowadays are regarded as cosmopolitan flies (David
and Capy, 1988). These flies over these 10,000 years developed adaptations to
the European climate, which exhibits much harsher winters than in Africa (David
and Capy, 1988). The colonization of American is suggested to have occurred
in two stages, the first of which tropical African flies followed the trade of slaves
into tropical America (i.e. the Caribbean). The second event occurred in the 19
th
century via the European settlers (David and Capy, 1988).
7
Since then, these two diverged populations of flies have been admixing in this
area of the world. Yukilevich and True (2008a) found that there are mate
preferences in these populations with homotypic matings being more common
than heterotypic matings (i.e. American flies prefer American mates and
Caribbean flies prefer Caribbean mates). In addition, they also found that West
African flies freely mated with Caribbean flies, but West African flies
discriminated mating with American flies. In a subsequent study Yukilevich and
True (2008b), found that morphological characteristics were distributed in a cline.
US flies were generally larger and lighter in pigmentation than Caribbean flies,
which were an intermediate size between US and African flies. In addition to
morphology, the intensity of courtship behavior between US and Caribbean
males was also clinally distributed with American males displaying more wing
song and ovipositor licking acts than the Caribbean males. The desaturase-2
locus was also assayed and it was found that the ancestral African insertion
allele increased in frequency southward in the cline. Desaturase-2 is important
for the variation in female cuticular hydrocarbons (CH) (Dallerac et al., 2000;
Takahasi et al., 2001), which play a role in mate recognition. These concurrent
premating clines in morphology, mate choice, male courtship behavior,
allozymes, and chemical signatures suggest incipient sexual isolation and are
also consistent with the definition of a hybrid zone. Given the historical aspects
8
of human migration, this area may very well be a secondary contact zone for
African and cosmopolitan flies thus making it a secondary hybrid zone.
1.5 Organization of Body of Work
This dissertation is a continuation into the investigation of the secondary hybrid
zone of cosmopolitan and African-like Drosophila melanogaster inhabiting the
southeast United States and Caribbean Islands. We begin by studying putative
reproductive barriers that are post-mating to understand to what extent incipient
speciation is occurring if at all. We have also included a study about circadian
rhythms in this group of flies to investigate whether there is temporal isolation. In
addition to investigating phenotypes that are putative reproductive barriers, we
have also included a study on the genomic sequences of American and
Caribbean flies to evaluate the extent of admixture between these flies to
understand at the sequence level if and how incipient speciation is progressing.
These three chapters are essentially manuscripts submitted or to be submitted
for publication. We end this body of work with a chapter describing a new study
system with potential to explore reproductive barriers and population genomics
between species.
9
1.6 Chapter References
Coyne J, Orr H. Patterns of speciation in Drosophila. Evolution. 1989;43(2):362–
81.
Coyne JA, Orr HA. Speciation. Sinauer Associates Incorporated; 2004.
Dallerac R, Labeur C, Jallon J-M, Knipple DC, Roelofs WL, Wicker-Thomas C. A
Δ9 desaturase gene with a different substrate specificity is responsible for the
cuticular diene hydrocarbon polymorphism in Drosophila melanogaster.
Proceedings of the National Academy of Sciences. National Acad Sciences;
2000 Aug 15;97(17):9449–54.
David J, Capy P. Genetic variation of Drosophila melanogaster natural
populations. Trends Genet. 1988;4(4):106–11.
Endler JA. Geographic Variation, Speciation, and Clines. Princeton, NJ:
Princeton University Press; 1977.
Harrison RG. Hybrid zones: windows on evolutionary process. Oxford Surveys in
Evolutionary Biology. 1990;7:69–128.
Harrison RG. Hybrid Zones and the Evolutionary Process. New York, NY: Oxford
University Press; 1993.
Huxley JS. Clines: an auxiliary taxonomic principle. Nature. 1938.
Seehausen O, Butlin RK, Keller I, Wagner CE. Genomics and the origin of
species. Nature Reviews …. 2014.
Servedio M, Noor M. The role of reinforcement in speciation: theory and data.
Annual Review of Ecology, Evolution, and Systematics. 2003;:339–64.
Sota T, Kubota K. Genital Lock-and-Key as a Selective Agent against
Hybridization. Evolution. 1998 Oct;52(5):1507.
Takahashi A, Tsaur S-C, Coyne JA, Wu C-I. The nucleotide changes governing
cuticular hydrocarbon variation and their evolution in Drosophila melanogaster.
Proceedings of the National Academy of Sciences. National Acad Sciences;
2001 Mar 27;98(7):3920–5.
10
Yukilevich R, True JR. African morphology, behavior and phermones underlie
incipient sexual isolation between US and Caribbean Drosophila melanogaster.
Evolution. 2008a Nov 1;62(11):2807–28.
Yukilevich R, True JR. Incipient sexual isolation among cosmopolitan Drosophila
melanogaster populations. Evolution. 2008b Aug;62(8):2112–21.
11
Chapter 2
The Influence of Post-mating Characters in the Incipient
Sexual Isolation of Cosmopolitan Drosophila melanogaster
2.1 Introduction
The process by which new species are formed is influenced by many natural
forces acting on the evolution of reproductive barriers. When these reproductive
barriers arise, gene flow is reduced between or within populations: giving rise to
the opportunity of speciation. These barriers are generally classified as pre-
zygotic or post-zygotic with further specification of whether they act before
mating or after mating, termed pre- and post-mating respectively (Coyne and Orr,
2004). Additionally, barriers can be extrinsic or intrinsic depending on if it
interacts with the environment or if it influences genetic incompatibilities
respectively (Seehausen et al., 2014). The process of speciation almost always
involves multiple diverse reproductive barriers with varying effect sizes (Coyne
and Orr, 2004).
The order in which barriers arise can depend on many complicating factors.
Theory suggests that pre-mating reproductive barriers may, on
12
average, exert greater isolating effects, but do not necessarily evolve first
(Cozzolino and Scopece, 2008; Coyne and Orr, 2004). Previous findings
suggests that pre- and post-zygotic barriers both evolve at similar rates in both
allopatric and sympatric speciation events (Coyne & Orr, 1989). Identifying the
presence and strength of reproductive barriers is vital to understanding the
process of speciation.
The development of reproductive barriers is a result of sexual selection action on
male and female conflict. Sexual conflict, broadly defined as ‘differences in the
reproductive interests between males and females’ (Chapman et al. 2003), drives
a perpetual cycle whereby males demonstrate traits to stimulate females into
suboptimally mating. Studies show that males and females often have divergent
interests within their environment, which can generate a sexual conflict that
influences the evolution of reproductive traits. When evolutionary interests are
incompatible like this, it results in sex-specific adaptations that can be harmful to
the opposite sex (Arnqvist and Rowe, 2005). One proposed model of how sexual
conflict drives evolutionary processes is through ‘chase away’ selection where
females consequently evolve resistance to male traits, and males in response
evolve novel or exaggerated traits to counteract this evolved resistance (Holland
and Rice, 1998).
13
Males may develop traits that make mating somewhat detrimental for females if it
is advantageous to prevent females from remating with other males (Johnstone
and Keller, 2000, Civetta and Clark, 2000). In Drosophila melanogaster, for
example, male sperm transferred after matting contain accessory gland proteins
reduce female remating rates and increase egg laying (Chapman et al., 2003;
Wolfner, 1997). Reduced receptivity to remating will also decrease the female
opportunity to mate with another male that may result in more fit progeny.
Increased egg laying and the trauma from mating also reduced female lifespan
(Fowler and Partridge, 1989). It is even suggested that male sperm is toxic to
females (Rice 1996). As a result, females develop resistance to these harmful
male traits, and males subsequently evolve new methods to discourage females
from mating with other males (Arnqvist and Rowe, 2005). It has been suggested
that females should be more resistant to males they have co-evolved in with
compared to males they have not coevolved with. These effects vary across
populations, and ecological context appears to be a factor (Arbuthnott et al.,
2014). This rapid, cyclical process termed sexually antagonistic coevolution has
been demonstrated in not only Drosophila species (Knowles and Markow, 2001)
and water striders (Rowe et al., 2002), but also in many other animals.
Coevolution by sexual conflict is a strong force behind reproductive isolation,
which may lead to speciation in specific circumstances (Martin and Hosken,
2003).
14
Studies investigating sexual isolation between populations have generally
focused on premating reproductive barriers, such as mate choice, male
morphology, and courtship behavior (Yukilevich and True, 2008a,b; Hollocher et
al., 1997). Postmating barriers are frequently ignored due to difficulties in
phenotyping and assumptions that premating barriers are of greater strength
(Coyne and Orr, 2004). We present our study as an effort to better understand
the role of postmating reproductive barriers in a system of Drosophila
melangoaster experiencing incipient speciation.
We have investigated the role of female egg laying, remating rates, hatchability,
and female longevity after mating with different males as putative postmating
reproductive barriers. These phenotypes are good candidates to investigate the
scope of extrinsic and intrinsic postmating reproductive barriers. We measure
each of these phenotypes and examine them for geographical patterns, which
may reveal if and how these barriers affect this system of southeast United
States and Caribbean Island Drosophila melanogaster.
15
2.2 Materials and Methodology
Fly Lines and Rearing Conditions
For our phenotypic assays, we used 23 isofemale lines of Drosophila
melanogaster collected in the summer of 2004 and 2005 (Yukilevich and True
2008). Origins are as following (TABLE 2.1; FIGURE 2.1): Selba, AL (ID#: 20, 28
and 20, 17); Thomasville, GA (ID#: 13, 34 and 13, 29); Tampa Bay, FL (ID#: 4,
12 and 4, 27); Birmingham, AL (ID#: 21, 39 and 21, 36); Meridian, MS (ID#: 24, 2
and 24, 9); Sebastian, FL (ID#: 28, 8); Freeport, Grand Bahamas-west (ID#: 33,
16 and 33, 11); George Town, Exumas (ID#: 36, 9 and 36, 12); Bullock’s Harbor,
Berry Islands (ID#: 40, 23 and 40, 10); Cockburn Town, San Salvador (ID#: 42,
23 and 42, 20); Mayaguana, Mayaguana (ID#: 43, 19 and 43, 18); Port Au
Prince, Haiti (ID#: H, 29 and H, 25). All flies were maintained at 25 °C in vials on
a standard cornmeal diet (recipe available upon request) and entrained under a
12hr light:12hr dark regime.
16
Map Number Location Line ID#’s
1 Birmingham, AL 21, 39 and 21, 36
2 Selba, AL 20, 28 and 20, 17
3 Thomasville, GA 13, 34 and 13, 29
4 Meridian, MS 24, 2 and 24, 9
5 Tampa Bay, FL 4, 12 and 4, 27
6 Sebastian, FL 28, 8
7 Freeport, Grand Bahamas-west 33, 16 and 33, 11
8 Bullock’s Harbor, Berry Islands 40, 23 and 40, 10
9 Cockburn Town, San Salvador 42, 23 and 42, 20
10 George Town, Exumas 36, 9 and 36, 12
11 Mayaguana, Mayaguana 43, 19 and 43, 18
12 Port Au Prince, Haiti H, 29 and H, 25
TABLE 2.1: List of locations and isofemale lines from those locations as well as corresponding
map numbers in FIGURE 2.1
17
FIGURE 2.1: Map of locations used in postmating assays
Egg laying, Hatchability, and Remating Rate Assays
Virgin females were collected from all 23 isofemale lines. Male flies up to one day
old were collected from two lines (ID#: 21,36 and 43,19) located at polar ends of
our geographical study region. We chose these two lines as sources for male
flies based on distance as well as maximal differences between
18
courtship profiles and physical characteristics (Yukilevich and True 2008b) to
account female mate preference, which has been previously established
(Yukilevich and True 2008a). All flies were collected on light CO2 anesthesia
and aged for three to four days before entering our assays. We set up a full
factorial experiment where females from each of the isofemale lines were
crossed with the two lines from which males were collected. Each cross was
replicated 15 times.
All flies were live manipulated using aspirators to avoid any behavioral effects
from CO2 anesthesia. Assays lasted 24 days and were in two stages. The first
stage which lasted 10 days measured remating rates and egg laying rates; the
second stage measured hatchability rates and lasted 14 days. In the first stage
(i.e. first 10 days), females were transferred daily by aspirator into new vials of
standard cornmeal fly food with blue food coloring. The blue food coloring helped
visualize eggs laid by females. The vials also had 20 uL of a 10% diluted active
yeast mixture to stimulate females’ reproductive activity. At lights on (ie. dawn)
on the initial day of the first stage, individual females were aspirated into a vial
with two males from either one of the two selected male lines were aspirated into
the vial for initial mating. Approximately 90 minutes was allocated for copulation
to occur, and all males were discarded immediately afterward this time period
using an aspirator. Females that did not mate on the first day did not continue in
the assay. Eggs were counted daily after the females were
19
transferred into a new vial. To assess short-term and long-term receptivity to
remating effects, each female was introduced to two new males of the same
genotype from her initial mating on the fourth and eighth day of the assay (i.e.
three and seven days after initial mating). We allowed 90 minutes on each
remating day for copulations to occur and all males were discarded via aspirator
thereafter. On the first day of the second stage of the assay, female identities
were checked to confirm correct sexing from when males were discarded from
remating days. Incorrectly sexed vials where the female was accidentally
discarded were not included in later analysis. Vials from the first stage of the
postmating assay were monitored daily for fly eclosion. Flies that eclosed were
counted and discarded . Vials were monitored daily for fly eclosion until there
were either three consecutive days where there were no fly eclosions or up to 14
days, whichever came first. All phenotyping assays during the first and second
stages were conducted within the first three hours of lights on (i.e dawn). All flies
from the first stage and eclosing vials in the second stage were kept in a
temperature-controlled incubator (25 °C) with a light timer set for a 12hr light:
12hr dark cycle.
Longevity Assays
For our longevity assays, we phenotyped a subset of lines from the 23 isofemale
lines that spanned the southeast United States and Caribbean
20
Islands. Female flies used in our longevity assays originated (arranged from
north to south) from Selba, Alabama, USA (ID#: 20,17), Thomasville, Georgia,
USA (ID#: 13, 34), Freeport, Grand Bahamas-west (ID#: 33, 11), Bullock’s
Harbor, Berry Islands (ID#: 40,23), and Port Au Prince, Haiti (ID#: H, 25).
Representative ‘American’ and ‘Caribbean’ males were derived lines originating
from the same male collection lines used in the egg laying, hatchability, and
remating assays, i.e. Birmingham, Alabama, USA (ID#: 21, 36) and Mayaguana,
Mayaguana (ID#: 43, 19). “Homotypic” crosses were where the male and female
were both of American origin or Caribbean origin. “Heterotypic” crosses were
defined as those where the male and female were from different origins (i.e.
American x Caribbean). Male and females from the same origin were assumed to
be more related or genetically similar to each other than male and females from
different origins based on some genetic evidence (Yukilevich and True 2008b).
Virgins females were collected on light CO2 anesthesia and aged singly in vials
for four days. Males were collected in the same fashion and aged in groups of
five per vial. We performed crosses in two rounds lasting approximately 70 days
for the first round and 80 days in the second round. In the first round, we crossed
female flies from Selba, Alabama, USA and Port Au Prince, Haiti to either our
representative ‘American’ male or ‘Caribbean’ male. There were 50 replicates for
each type of cross. Because of the large effect size from our initial round, in the
rest of our lines we had 25 replicates for each type of cross. In
21
each round, aged female flies were placed with five male flies for 48 hours to
ensure mating occurred. Male flies were removed using an aspirator after the
mating period. Female flies were then observed on a regular basis five days per
week. Dates of deaths were recorded until the end of the 70 or 80 day
observational period. The females were transferred to fresh vials every seven
days.
Post-mating behavior data analysis
We examined the effects of geographic location on the total number of eggs laid
by females over the first stage of the phenotypic assays, the total hatchability of
those egg laid, and the propensity of females to remate three and seven days
after initial mating. For egg laying and hatchability, we used a linear regression
model with latitude and longitudinal coordinates as predictors as well as the male
and female identity and phenotyping blocks to account for the variation from
genotypes of male and females in addition to experimental block effects. Model
fit and effects of factors was assessed with ANOVA tables produced by the
models. Because remating was scored as a categorical variable of whether or
not the female copulated on the two remating days, we used logistic regression
models to assess the effects of geographic location while controlling for male and
female genotypes and block effects on short- and long-term female receptivity to
22
remating. The significance of longitudinal and latitudinal coordinates and model
fits were assessed using analysis of deviance tables.
We performed a permutation test to investigate the significance of the lower
hatchability rates in the three central locations as revealed by linear models as
well as visual confirmation of plots. We calculated the difference in hatchability
between the five lines from our three central locations and the hatchability of all
other fly lines (18 lines). We then randomly assigned fly lines into groups of five
and 18 and calculated the difference in hatchability between these two groups.
These permutations were repeated 10,000 times. P-values were calculated by
the number of times the difference in hatchability between these two groups were
equal to or greater than our observed value divided by our 10,000 permutations.
The line with the lowest hatchability was removed for a follow up permutation test
to evaluate that the lower hatchability was only due to the effect of one line.
Similar permutation tests was conducted on total egg counts to determine that
lower hatchability was also not due to lower egg counts. Hatchability of eggs laid
by females mated to American and Caribbean males were performed separately.
P-values from these tests were corrected using the Bonferroni method.
All analysis was performed in R and code is available for the permutation test
upon request.
23
Longevity data analysis
Survival analysis is used for temporal data of waiting times to an event with
censored data. We employed methods from survival analysis to examine our
data. We analyzed the waiting times of female death after experiencing a
homotypic or heterotypic mating. Females that escaped or survived past our
observational periods were considered censored data points. The first step of
survival analysis is to estimate survival functions for each of our crosses, S(t),
which in our case is the probability of a female living longer than time, t. This can
be done non-parametrically using the Kaplan-Meier method (Kleinbaum and
Klein 2012). Parametric models were tried (i.e. exponential, log-normal, log-
logistic, generalized gamma), but none yielded a good fit (data not shown). After
survival curves were fitted, we then used the fitted survival curves to estimate the
cumulative hazard function, H(t), for each type of cross. The cumulative hazard
function shows the cumulative probability that a female has expired up to time, t.
The relationship between the survival function and the cumulative hazard
function is:
H(t) = −ln(S(t))
or
S(t) = e
!!(!)
24
The most common statistical test used for comparing survival distributions is the
log-rank test. However, this test has the proportional hazards assumption which
requires that the hazard functions of the two groups being compared are parallel.
Hazard functions for our comparisons of survival after heterotypic and homotypic
matings were plotted and visually checked for the crossing of hazard curves.
When hazard curves cross, the proportional hazards assumption is violated so
another test must be used because the standard log-rank test has little to no
power in this situation (Klein and Moeschberger, 2003). We chose to use a
combined weighted log-rank test (Zhou et al., 2006), which takes into account
crossing hazard curves. This improved log-rank test has more power than the
standard log-rank tests when the hazard functions cross and the hazard ratio is
not proportional.
All analysis was performed in R using the ‘survival’ package to estimate the
survival curves and hazard functions. The package emplik was used as part of
the improved log-rank test, which code in R can be found here:
http://www.ms.uky.edu/%7Emai/research/LogRank2006.pdf
25
2.3 Results
Egg counts
Egg counts for each line phenotyped are shown in a graph using side-by-side
boxplots with locations arranged from the northernmost to the southernmost
location, left to right (FIGURE 2.2). It does not appear that egg counts follow a
clinal pattern in either case of females mated to American or Caribbean males.
The graph reveals there is much variation within lines, but the median egg count
for each location is approximately the same except for in the case of females
from location 28 (Sebastian, FL) when mated to Caribbean males (FIGURE 2.2).
26
FIGURE 2.2: Egg counts of females mated with American males (top) and Caribbean males
(bottom). Each box plot is a isofemale line arranged from the northernmost location (left) to the
southernmost location (right)
27
The full regression model showed that longitude and latitude were not significant
influences (p = 0.3324) on egg laying and that most variance was accounted for
by male (p < 0.001) and female (p<0.001) genotypes as well as block effects (p =
0.0018). Comparing the full model with the reduced model where longitude and
latitude were omitted showed that the addition of longitude and latitude as
predictive variable did not help the predictive power of the full model (p =
0.4994). (TABLE A1, A2, A3)
Remating
Short- and long-term remating rates for each isofemale line were plotted against
latitude and longitude coordinates (FIGURE A1, A2). Short-term remating rates
were generally lower (range of rates : 0-30%) than long-term remating rates
(range of ranges: 0-60%). Remating rates do not appear to be influenced by
location, which investigated further with logistic regression.
The full logistic regression model evaluating effects of latitude and longitude
while controlling for male and female genotypes as well as block effects found
that latitude (p = 0.11) or longitude (p = 0.35) were not useful in predicting short-
term remating rates with similar results for long-term remating rates (lon p =
0.7616, lat p = 0.6361). Male genotype also was not a significant influence on
short-term or long-term remating rates (p = 0.4848 and p =
28
0.1240). The reduced models removing latitude and longitude as predictors also
showed that they were not significantly influencing remating rates. Female
identities in both logistic models for short- and long- term remating rates were
significant giving evidence that female genotypes could influence remating rates.
However, when we fitted a model for long-term remating rates with a male x
female interaction term, results showed that this interaction term was not
significant (p = 0.0959). (TABLE A4, A5, A6, A7, A8, A9, A10, A11)
Hatchability
Hatchability for the various locations in the southeast US and Caribbean Islands
were visualized in a graph using side-by-side boxplots with locations arranged
from the northernmost to the southernmost location, left to right (FIGURE 2.3).
Hatchability in the three middle locations (TABLE 2.1: locations 5,6,7) at the
border of the southeast US and Caribbean Islands appear lower than the
locations on the edges in both the graphs displaying hatchability of females
mated to American males and Caribbean males (FIGURE 2.3).
29
FIGURE 2.3: Hatchability of females mated with American males (top) and Caribbean males
(bottom). Each box plot is a isofemale line arranged from the northernmost location (left) to the
southernmost location (right)
30
Our full linear regression model took into account male and female identities on
hatchability as well as experimental block effects while assessing influences of
longitude (p = 0.048) and latitude. Longitude had a significant effect on
hatchability (p =0.0483) while latitude did not (p = 0.2396). However when we
compare the reduced model with latitude removed with the full model, we find
that latitude did help significantly in explaining hatchability (0.0302). (TABLE 2.2,
2.3, 2.4)
Df Sum Sq Mean Sq F value Pr(>F)
Block 14 2.9337 0.20955 4.6414 4.007e-08
Female 22 7.7220 0.35100 7.7745 <2.2e-16
Male 1 0.8869 0.88692 19.6448 1.084e-05
Latitude 1 0.6255 0.06255 1.3856 0.23956
Longitude 1 0.17674 0.17674 3.9147 0.04826
Residuals 694 31.3326 0.04515
TABLE 2.2: ANOVA Table for Full Model of Hatchability
31
Df Sum Sq Mean Sq F value Pr(>F)
Block 14 2.9337 0.20955 4.6168 4.557e-08
Female 22 7.7220 0.35100 7.7332 <2.2e-16
Male 1 0.8869 0.88692 19.5403 1.143e-05
Longitude 1 0.0264 0.02639 0.5813 0.446
Residuals 6945 31.5455 0.04539
TABLE 2.3: ANOVA Table for Reduced Model of Hatchability with no Latitude
Res Df RSS DF Sum of
Sq
F Pr(>F)
Full 694 31.333
Reduced 695 31.546 -1 -0.21291 4.47158 0.03022
TABLE 2.4: ANOVA table of model comparisons
To evaluate the significance of the dip in hatchability rates, we performed
permutation tests as described in our methods section. We found that the
hatchability in the middle three locations was significantly lower than the rates in
the surrounding locations regardless of being mated to an American male (p <
0.0001) or Caribbean male (p <0.0001). Results were similar when the location
with the lowest hatchability rate was removed (28: Sebastian, FL, USA) and the
permutation tests performed again (females mated to American male: p =
0.0056, females mated to Caribbean male: p = 0.0272). Similar tests were
conducted on egg counts to investigate whether the lower
32
hatchability was due to lower egg counts. No significant differences in egg counts
between females from the middle locations and the outer locations were found
regardless of whether they were mated to American males (p = 0.3192) or
Caribbean males (p = 0.7584). The same results were yielded when we removed
the influence of the middle location, 28: Sebastian, FL, USA, (mated to American
males: p=0.3016, mated to Caribbean males: p = 1.0). These results suggest a
generalizable central location effect on hatchability.
Longevity
Five female lines representing various locations in the Southeastern U.S. and
Caribbean Islands were measured for longevity after experiencing homotypic or
heterotypic matings. The homotypic cross survival curves for females from
isofemale lines 13,34, 20,17, and H,25 were consistently higher than the survival
curves of females in heterotypic crosses (FIGURE 2.4, 2.5, 2.6). There were no
apparent differences between homotypic and heterotypic survival curves of
females originating from 33,11 or 40,23 (FIGURE 2.7, 2.8).
33
FIGURE 2.4: Survival curves of females from line 20,17 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
FIGURE 2.5: Survival curves of females from line H,25 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
34
FIGURE 2.6: Survival curves of females from line 13,34 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
35
FIGURE 2.7: Survival curves of females from line 33,11 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
FIGURE 2.8: Survival curves of females from line 40,23 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
36
Hazard curves for all crosses and lines revealed non-proportional hazards in
almost all cases of homotypic and heterotypic matings. (Supplementary Figure
A3, A4, A5, A6, A7). Crossing points of all hazard functions were visually
estimated for use in the improved log-rank tests (TABLE 2.5).
The improved log-rank tests showed evidence that females after heterotypic
matings had shorter lifespans than females in homotypic matings for females
from 13,34, and H,25 (p = 0.0410 and p = 0.0271). Females of line 20,17 showed
a reduced lifespan when involved in heterotypic matings (FIGURE 2.4), but these
results were not significant (p = 0.3130).
Female line
T~approx time of crossing
hazards
pvalue from improved log
rank
13,34 37 0.04096407
40,23 42 0.4246727
33,11 40 0.6260448
H,25 23 0.02706502
20,17 61 0.3129819
TABLE 2.5: Improved Log-rank Test Results
37
2.4 Discussion
How species form is a complex problem dependent on a plethora of factors from
interactions with the environment or other individuals to the selectional force
driving the process. We examined several potential postmating reproductive
barriers such as remating rates, egg laying rates, hatchability, and female
longevity that may influence a system experiencing incipient speciation.
We did not find any evidence that egg laying rates or remating rates influenced
the reproductive success in a systematic way with regards to these isofemale
lines from the southeast United States and Caribbean Islands. However, the lack
of evidence does not imply that behaviors are not influential postmating
reproductive barriers. Current views of speciation view the process as a sliding
continuum where speciation can move forward or step back and may even be
arrested at intermediate stages (Seehausen et al., 2014). Depending on the
driving force of speciation, different types of reproductive barriers form at
different stages (Seehausen et al., 2014) thus it may be that postmating
behaviors could be of importance at other stages in the speciation continuum.
We also examined female longevity postmating with males that were more or
less related to them as defined by physical distance. There was no apparent
geographical pattern in our female longevity findings, but we
38
did find a mosaic pattern of heterotypic matings reducing female lifespan in a
couple isofemale lines. Previous laboratory evolution studies indicate that
females develop ‘resistance’ against males they coevolve with in the same
environment (Arbuthnott et al., 2014). With our findings, this may be the case in
natural populations outside of the laboratory, however, due to the low number of
lines we tested, our study may lack power to detect larger effects of this extrinsic
postmating barrier.
We did observed an interesting hatchability rate valley produced by the isofemale
lines originating from our three central locations spanning the border of the
United States and the Caribbean Island. This result may be evidence that there
are essential genetic differences between American and Caribbean fly
populations, which could have manifested as a intrinsic postzygotic barrier
between American and Caribbean populations. This may be indicative of the
presence of Bateson-Dobzhansky-Muller incompatibilities (DMI) which are
negative epistatic interactions and the most common form of intrinsic postzygotic
isolation (Presgraves, 2010). A more thorough investigation of these lines and
genome sequences that are beyond the scope of this study are required to
confirm the presence of DMI’s in this incipient speciation system.
39
2.5 Chapter References
Arbuthnott D, Dutton EM, Agrawal AF, Rundle HD. The ecology of sexual
conflict: ecologically dependent parallel evolution of male harm and female
resistance in Drosophila melanogaster. Ecol. Lett. 2014 Feb 1;17(2):221–8.
Arnqvist G, Rowe L. Sexual conflict: monographs in behavior and ecology.
Princeton, NJ: Princeton University Press; 2005.
Chapman T, Bangham J, Vinti G, Seifried B, Lung O, Wolfner MF, et al. The sex
peptide of Drosophila melanogaster: Female post-mating responses analyzed
by using RNA interference. Proceedings of the National Academy of Sciences.
National Acad Sciences; 2003 Aug 19;100(17):9923–8.
Civetta A, Clark AG. Correlated effects of sperm competition and postmating
female mortality. Proceedings of the National Academy of Sciences. National
Acad Sciences; 2000;97(24):13162–5.
Coyne JA, Orr HA. Speciation. Sinauer Associates Incorporated; 2004.
Cozzolino S, Scopece G. Specificity in pollination and consequences for
postmating reproductive isolation in deceptive Mediterranean orchids. … of
the Royal …. 2008.
Fowler K, Partridge L. A cost of mating in female fruitflies. , Published online: 27
April 1989; | doi:10.1038/338760a0. Nature Publishing Group; 1989 Apr
27;338(6218):760–1.
Holland B, Rice WR. Perspective: Chase-away sexual selection: Antagonistic
seduction versus resistance. Evolution. 1998 Feb;52(1):1–7.
Hollocher H, Ting CT, Wu ML, Wu CI. Incipient speciation by sexual isolation in
Drosophila melanogaster: Extensive genetic divergence without reinforcement.
Genetics. 1997 Nov;147(3):1191–201.
Johnstone RA, Keller L. How males can gain by harming their mates: sexual
conflict, seminal toxins, and the cost of mating. The American Naturalist. 2000.
Klein JP, Moeschberger ML. Survival Analysis. Springer; 2003.
Kleinbaum DG, Klein M. Kaplan-Meier survival curves and the log-rank test.
Survival analysis; 2012.
40
Knowles LL, Markow TA. Sexually antagonistic coevolution of a postmating-
prezygotic reproductive character in desert Drosophila. Proceedings of the
National Academy of Sciences. National Acad Sciences; 2001 Jul
17;98(15):8692–6.
Martin OY, HOSKEN DJ. The evolution of reproductive isolation through sexual
conflict. Nature. 2003;423(6943):979–82.
Presgraves DC. Darwin and the Origin of Interspecific Genetic Incompatibilities.
Am Nat. The University of Chicago Press; 2010 Dec 18;176(S1):S45–S60.
Rice WR. Sexually antagonistic male adaptation triggered by experimental arrest
of female evolution. Nature. 1996.
Rowe L, Arnqvist G. Sexually antagonistic coevolution in a mating system:
Combining experimental and comparative approaches to address evolutionary
processes. Evolution. 2002 Apr;56(4):754–67.
Seehausen O, Butlin RK, Keller I, Wagner CE. Genomics and the origin of
species. Nature Reviews …. 2014.
Wolfner MF. Tokens of love: Functions and regulation of drosophila male
accessory gland products. Insect Biochemistry and Molecular Biology. 1997
Mar;27(3):179–92.
Yukilevich R, True JR. African morphology, behavior and phermones underlie
incipient sexual isolation between US and Caribbean Drosophila melanogaster.
Evolution. Blackwell Publishing Inc; 2008a Nov 1;62(11):2807–28.
Yukilevich R, True JR. Incipient sexual isolation among cosmopolitan Drosophila
melanogaster populations. Evolution. Blackwell Publishing Inc; 2008b Aug
1;62(8):2112–21.
Zhou M, Bathke A, Kim M. Combined Multiple Testing by Censored Empirical
Likelihood. Univ. Kentucky, Dept. of Statistics Tech Report. 2006
41
Chapter 3
Circadian Rhythms of American and Caribbean Drosophila
melanogaster
3.1 Introduction
The timing at which events occur is crucial for many biological processes. At the
molecular level, gene expression is coordinated in a complex symphony where if
even one gene is expressed too early, too late, too much, or not at all, it could
spell disaster for the entire system (Yasuda et al., 1991). At the organismal scale,
correct timing could mean a chance to mate and pass on genes and incorrect
timing could result in a loss of fitness (Harrison, 1984). The timing of molecular
processes, which lead to events occurring at the organismal level can lead to
consequences in biological systems at the population scale as well. Temporal
isolation or also known as allochronic isolation can affect gene flow between or
within populations especially if these populations breed at different times and, in
theory, can lead to speciation (Coyne and Orr, 2004).
The basis of the study of temporal isolation lies with circadian rhythms. Circadian
rhythms are innate systems of complicated gene and
42
environmental interactions common to all living organisms from simple single-
celled life forms to complex multi-cellular animals (Young and Kay, 2001;
Clodung et al., 2007; Dibner et al., 2010). These cycles are highly connected with
the timing when organisms eat, sleep, mate, etc. and are roughly around 24
hours for most animals (Young and Kay 2001; Panda et al., 2002). Perhaps one
of the most studied model circadian rhythms are those of Drosophila
melanogaster which is entrain on dawn and dusk cues. They are most active at
dawn and dusk in anticipation of lights-on and lights-off and activity is lulled
midday in a siesta (Greenspan 2007).
Since the discovery of the first clock mutants (Konopka and Benzer, 1971), the
breadth of knowledge about the neurobiology of D. melanogaster has
exponentially increased. Much of the complex interactions between essential
genes and their products as well has how light cues play a role has been
described in addition to the critical neurons in the fly brain where all these
interactions take place (Helfrich-Förster, 2004). Moreover, the natural distribution
of different variants of certain clock genes have been investigated (Kyriacou et
al., 2008). For example, the gene, period, was shown to be in a latitudinal cline in
Europe (Costa et al., 1991), which is related to how those D. melanogaster
populations confer different thermal stability on their circadian rhythms (Ewers et
al., 1990; Sawyer et al., 1997).
43
In our study, we are interested whether aspects of circadian behavior in fly lines
from the southeast United States and Caribbean islands could be putative
reproductive barriers. This study was motivated by results from Yukilevich and
True (2008b) where they demonstrated a cline where American male flies were
more vigorous in courtship when compared to Caribbean male courtship. The
varying intensity in courtship was also observed when we conducted our post-
mating trials in the previous chapter. It is known that circadian rhythms greatly
influence courtship activity and mating (Hardeland, 1972; Kyriacou and Hall,
1980; Sakai and Ishida, 2001) so naturally we investigated circadian cycles of
flies in our study system.
3.2 Materials and Methods
Fly Lines and Rearing Conditions
We used 21 isofemale lines of Drosophila melanogaster collected in the summer
of 2004 and 2005 (Yukilevich and True 2008). Origins are as following : Selba,
AL (ID#: 20, 28 and 20, 17); Thomasville, GA (ID#: 13, 34 and 13, 29); Tampa
Bay, FL (ID#: 4, 12); Birmingham, AL (ID#: 21, 39 and 21, 36); Meridian, MS
(ID#: 24, 2 and 24, 9); Freeport, Grand Bahamas-west (ID#: 33, 16 and 33, 11);
George Town, Exumas (ID#: 36, 9 and 36, 12); Bullock’s
44
Harbor, Berry Islands (ID#: 40, 23 and 40, 10); Cockburn Town, San Salvador
(ID#: 42, 23 and 42, 20); Mayaguana, Mayaguana (ID#: 43, 19 and 43, 18); Port
Au Prince, Haiti (ID#: H, 29 and H, 25). All flies were maintained at 25 °C in vials
on a standard cornmeal diet and entrained under a 12hr light:12hr dark regime.
Fly collection and experimental set up
Sixteen males within 24 of eclosion were collected from each line using light CO2
anesthesia. Each male was placed singly in an activity chamber of a Drosophila
Activity Monitor (TriKinetics). Each monitor held up to 32 individual flies so we
randomly designated two fly lines per monitor and randomized their position in
the monitors. Monitors were placed in a LD incubator kept at 25 °C on a 12 hour
light: 12 hour dark schedule. We allowed a 12 hour adjustment time before
recording locomotor activities. Recordings were taken every five minutes.
Monitors were in the LD incubator for four days before being moved to another
25 °C incubator which was in constant darkness (DD). Locomotor activity of the
flies were recorded in DD for 10 days.
Circadian data analysis
Locomotor data from DD conditions and LD conditions were analyzed separately.
For each fly we used the data collected in DD conditions to
45
estimate the period length (i.e. number of hours in a circadian cycle) as well as
the periodicity (i.e. strength of the period over 10 days DD). Periodicity was
represented using the metric, power, which assesses how periodic a cycle is
using a Chi-squared periodogram (Sokolove and Bushell 1978). A power value
below 100 was an indication that the specimen was arrhythmic. Flies that expired
before the 10 day DD observational period were excluded from the analysis.
Activity plots of all flies over the 10 day DD period were visually inspected before
analysis. Period length and power values were obtained using the Actimetric’s
Clocklab Analysis Software (Coulbourn Instruments).
Locomotor data from LD conditions were used to investigate sleep and activity.
Each 24 hour period of LD observation were averaged together to obtain a
summary of a ‘typical’ day. Sleep was defined as five minutes of inactivity. We
initially investigated sleep and activity on a per line basis, but due to results from
Nicolas Svetec at UC Davis (unpublished), we analyzed the data with all
American lines in one group and all Caribbean lines in another group. We used a
Wilcoxon rank sum test to evaluate differences in total amount of sleep during
lights on and lights off between these two groups. All LD analysis were done in R
using custom scripts.
46
3.3 Results
Period length and Periodicity
FIGURE 3.1: Box plots of period of all isofemale lines arranged from northernmost location on the
left to the southernmost location on the right. Numbers on the y-axis represent period length in
hours.
Period length was highly variable between lines and do not appear to be in a clinal
pattern in the southeast United States and Caribbean Islands (FIGURE 3.1).
47
FIGURE 3.2: Box plots of power of all isofemale lines arranged from northernmost location on the
left to the southernmost location on the right. Numbers on the y-axis represent period length in
hours.
Some lines had high within group variation in power values. Overall, there did not
seem to be a clinal pattern in how periodic lines were with decreasing latitude
(FIGURE 3.2).
Activity and Sleep
We examined the activity of American-grouped flies and compared them with
Caribbean grouped-flies and found no difference in the average activity level
during lights off and lights off (FIGURE 3.3).
48
FIGURE 3.3: Activity of American (red line) and Caribbean (black line) grouped over a 24 hour
period broken up by 5 minute bins (x-axis).
When we examined sleep in a box plot we saw that Caribbean flies appeared to
sleep slightly more than American flies and investigated this further by creating a
graph depicting amount of sleep over a 24 hour period (FIGURE 3.4).
49
FIGURE 3.4: Box plots of total sleep of American (SE US) and Caribbean (Carib) during lights on
(L) and lights off (D)
FIGURE 3.5: Average sleep of American flies (red line) and Caribbean flies (black line) over a 24
hour period by 5 minute increments.
The plot of the average amount of sleep of American flies and Caribbean flies over a 24
hour period revealed that Caribbean flies do generally sleep more than American flies
(FIGURE 3.5). This difference in overall sleep was found to be
50
statistically significant with a Wilcoxon rank sum test (W= 9719, p = 9.814e-07).
However, when sleep during lights on was examined separately from sleep during lights
off, it was found that Caribbean flies sleep significantly more than American flies during
lights off (W = 11429, p = 0.00296), but not for lights on sleep (W = 12711, p-value =
0.1261).
3.4 Discussion
In summary, period length and periodicity do not appear to be clinally distributed
among various lines from the southeast United States and the Caribbean islands.
We then examined these circadian patterns in a course view by grouping all
American isofemale line observations together and all Caribbean isofemale lines
together. There were no differences in activity between American and Caribbean
male Drosophila melanogaster. However, upon investigating sleep in these two
broad populations, Caribbean fly lines slept significantly more than American fly
lines. These findings are consistent with another independent study (Svetec and
Begun, unpublished) where a sleep cline has been described on the east coast
of the United States with northern flies sleeping less than southern flies.
The relative importance of temporal isolation and its role in speciation by
reducing gene flow has had differing views. Reviews on this type of reproductive
barrier generally surmise that it is unimportant in relation to
51
other types of isolating barriers and proper studies on whether or not temporal
isolation has a pivotal role in speciation are nonexistent (Coyne and Orr, 2004).
Most likely, temporal isolation is not greatly influential by itself, but may be acting
with other types of reproductive barriers in the speciation process (Coyne and
Orr, 2004).
Certainly, the results in this study do not indicate that temporal isolation is a
major contributing factor in the differentiation of these American and Caribbean
population of Drosophila melanogaster. The slight, but significant, differences in
sleep behavior is most likely not large enough to greatly impact gene flow on
such a small geographical scale. If it does influence gene flow, it is possible that
gene flow from the Caribbean Islands into the United States is slightly impeded
by missed matings as a result of increased sleep, but in the end, this is only
speculative because there is a lack of correlation between phenotypes and the
effect size on reproductive isolation (Seehausen et al., 2014). To test the effects
of increased sleep on gene flow is beyond the scope of this study and would
require a different set of experiments.
52
3.5 Chapter References
Clodong S, Dühring U, Kronk L, Wilde A, Axmann I, Herzel H, et al. Functioning
and robustness of a bacterial circadian clock. Molecular Systems Biology.
John Wiley & Sons, Ltd; 2007 Jan 1;3(1).
Costa R, Peixoto AA, Barbujani G, Kyriacou CP. A Latitudinal Cline in a
Drosophila Clock Gene. Proc Biol Sci. The Royal Society; 1992 Oct
22;250(1327):43–9.
Coyne JA, Orr HA. Speciation. Sinauer Associates Incorporated; 2004.
Dibner C, Schibler U, Albrecht U. The mammalian circadian timing system:
organization and coordination of central and peripheral clocks. Annual review
of physiology. 2010.
EWER J, HAMBLENCOYLE M, ROSBASH M, HALL JC. Requirement for Period
Gene-Expression in the Adult and Not During Development for Locomotor-
Activity Rhythms of Imaginal Drosophila-Melanogaster. Journal of
Neurogenetics. 1990;7(1):31–&.
Greenspan RJ. An Introduction to Nervous Systems. CSHL Press; 2007.
Hardeland R. Species differences in the diurnal rhythmicity of courtship
behaviour within the Melanogaster group of the genus Drosophila. Animal
Behaviour. 1972 Feb;20(1):170–4.
Harrison RG. Barriers to Gene Exchange Between Closely Related Cricket
Species. II. Life Cycle Variation and Temporal Isolation. Evolution. 1985
Mar;39(2):244.
Helfrich Förster C. Neurobiology of the fruit fly's circadian clock. Genes, Brain
and Behavior. Munksgaard International Publishers; 2005 Mar 1;4(2):65–76.
Konopka RJ, Benzer S. Clock Mutants of Drosophila melanogaster. Proceedings
of the National Academy of Sciences. National Acad Sciences; 1971 Sep
1;68(9):2112–6.
Kyriacou CP, HALL JC. Circadian-Rhythm Mutations in Drosophila-Melanogaster
Affect Short-Term Fluctuations in the Males Courtship Song. Proceedings of
the National Academy of Sciences. 1980;77(11):6729–33.
53
Kyriacou CP, Peixoto AA, Sandrelli F, Costa R, Tauber E. Clines in clock genes:
fine-tuning circadian rhythms to the environment. Trends in Genetics. 2008
Mar;24(3):124–32.
Panda S, Hogenesch JB, Kay SA. Circadian rhythms from flies to human. Nature.
Nature Publishing Group; 2002 May 16;417(6886):329–35.
Sakai T, Ishida N. Circadian rhythms of female mating activity governed by clock
genes in Drosophila. Proceedings of the National Academy of Sciences.
National Acad Sciences; 2001 Jul 31;98(16):9221–5.
Sawyer LA, Hennessy JM, Peixoto AA, Rosato E, Parkinson H, Costa R, et al.
Natural variation in a Drosophila clock gene and temperature compensation.
Science. 1997;278(5346):2117–20.
Seehausen O, Butlin RK, Keller I, Wagner CE. Genomics and the origin of
species. Nature Reviews …. 2014.
Sokolove PG, Bushell WN. The chi square periodogram: Its utility for analysis of
circadian rhythms. Journal of Theoretical Biology. 1978 May;72(1):131–60.
Svetec N, and Begun D. unpublished results
Yasuda GK, Baker J, Schubiger G. Temporal regulation of gene expression in
the blastoderm Drosophila embryo. Genes Dev. Cold Spring Harbor Lab; 1991
Oct 1;5(10):1800–12.
Young MW, Kay SA. Time zones: a comparative genetics of circadian clocks. Nat
Rev Genet. Nature Publishing Group; 2001 Sep 1;2(9):702–15.
Yukilevich R, True JR. African morphology, behavior and phermones underlie
incipient sexual isolation between US and Caribbean Drosophila melanogaster.
Evolution. Blackwell Publishing Inc; 2008 Nov 1;62(11):2807–28.
54
Chapter 4
Investigating the Sources and Extent of European and
African Admixture in North American Populations of
Drosophila melanogaster
4.1 Introduction
Out of the thousands of species in the genus, Drosophila, the single most
extensively studied species is Drosophila melanogaster (Powell, 1997). The
utility of D. melanogaster as a model organism can be seen in many fields of
research from medicine to evolutionary biology. The advent of next-generation
sequencing (NGS) enabling the high-throughput sequencing of genomes has
generated much interest in the population genomics of D. melanogaster species
(Mackay et al., 2012; Pool et al., 2012; Campo et al., 2013). Understanding
demographic models of D. melanogaster can be approached with whole genome
data.
According to the current demographic model, D. melanogaster originated in sub-
Saharan Africa with a migration event into the European
55
continent 10,000 years ago (David and Capy, 1988). Colonization of the
Americas is hypothesized to have happened in two waves. The first wave
occurred ~400-500 year ago with African flies being transported into the
Caribbean Islands along with the transatlantic slave trade. The second wave,
which happened in the mid-19th century, was the cosmopolitan flies arriving with
the first European settlers into North America (David and Capy, 1988). These
two waves created a hybrid zone in the southeast United States and Caribbean
Islands of cosmopolitan-adapted flies from Europe and African-like flies from
West Africa. The flies originating from the Caribbean islands have retained
African-like behavior and physical phenotypes despite its close proximity to the
US cosmopolitan populations (Yukilevich and True, 2008a; Yukilevich and True,
2008b; Yukilevich et al., 2010).
Previous studies looking at genome-wide effects of divergence in these
populations used tiling microarrays to detect highly differentiated regions
between the pooled genomes of cosmopolitan populations (including Caribbean
fly lines) and Zimbabwean populations and then sequenced a subset of
fragments to look at genetic divergence (Yukilevich et al., 2010). Most
differentiation was found between populations living in African versus out of
Africa and evidence supporting that most of the variation in North America and
African populations originated from the sorting of African standing genetic
variation into the New World through Europe (Yukilevich et al.,
56
2010). However, Caracristi and Schlötterer (2003) found high levels of
polymorphisms in North American populations where the proportion of shared
alleles between African and American populations were greater than the
proportion of shared alleles between African and European populations. This
evidence supports the hypothesis that there was a separate migration event to
the Caribbean and thus are the source of these putative African alleles in North
America (Li and Stephan 2006). More recently, Duchen et al. (2013) showed that
North American populations of D. melanogaster are most likely an admixture
between European and African ancestry with the African ancestry accounting for
15% of the mixture. It is not clear whether there was a second migration event to
the Caribbean from Africa. The Caribbean islands has been purported to be the
source of additional African alleles in the North American populations although it
has not been confirmed.
FIGURE 4.1: Map of populations with number of whole genome sequences. Grey arrows indicate
currently accepted migration history of D. melanogaster.
57
We have sequenced 23 Drosophila melanogaster genomes from various
locations in the southeast United States and the Caribbean Islands. Combined
with the current sequencing efforts of other fly populations from Raleigh, NC,
USA, Winters, CA, USA, Montpellier, France, Europe, and Oku, Cameroon,
Africa, we can explore African and European admixture of North American
populations and elucidate the history of D. melanogaster’s migration to the
Americas and to understand how Caribbean D. melanogaster populations can
retain African phenotypes while being influenced by genetic material from the
United States.
4.2 Materials and Methods
Fly Lines for Sequencing
The same 23 isofemale lines of Drosophila melanogaster that we used in
Chapter 2 and 3 were selected for sequencing. Origins are as following: Selba,
AL (ID#: 20, 28 and 20, 17); Thomasville, GA (ID#: 13, 34 and 13, 29); Tampa
Bay, FL (ID#: 4, 12 and 4, 27); Birmingham, AL (ID#: 21, 39 and 21, 36);
Meridian, MS (ID#: 24, 2 and 24, 9); Sebastian, FL (ID#: 28, 8); Freeport, Grand
Bahamas-west (ID#: 33, 16 and 33, 11); George Town, Exumas (ID#: 36, 9 and
36, 12); Bullock’s Harbor, Berry Islands (ID#: 40, 23 and 40,
58
10); Cockburn Town, San Salvador (ID#: 42, 23 and 42, 20); Mayaguana,
Mayaguana (ID#: 43, 19 and 43, 18); Port Au Prince, Haiti (ID#: H, 29 and H, 25).
All flies were maintained at 25 °C in vials on a standard cornmeal diet.
Libraries and sequencing of southeast US and Caribbean lines
All lines were subjected to full-sibling inbreeding for at least five generations
before we collected 15 - 20 females from each line for library preparation. DNA
was extracted using an Epicentre MasterPure kit (Madison, WI, USA) and
cleaned with the Zymo Quick-gDNA Miniprep kit (Irvine, CA, USA). Solexa
sequencing libraries were prepared according to Dunham and Friesen (2013)
with the exception that DNA was sheared with dsDNA Shearase Plus (Zymo:
Irving, CA, USA) and cleaned using Agencourt AMPure XP beads (Beckman-
Coulter: Indianapolis, IN, USA). Fragment size selection was also done using
beads instead of gel electrophoresis. Libraries were run on an Illumina HiSeq
2500 resulting in 100 basepair single end reads.
Sources of other sequenced populations
We used the 35 isogenic lines from Winters, CA, USA and 33 isogenic lines from
Raleigh, NC, USA described in Campo et al. (2013). Raleigh, NC, USA lines
were a subset of the Drosophila Genetic Reference Panel
59
(DGRP) (Mackay et al, 2012). The 10 isofemale lines from Oku, Cameroon,
Africa were sequenced as a part of the Drosophila Population Genetic Panel
(DPGP-2 African Survey) (Pool et al., 2012). Sequencing reads for 20 isofemale
lines from Montpellier, France, Europe were downloaded via the Bergman lab
webpage (Haddrill and Bergman, 2012).
Mapping
For each fly line, the raw sequencing reads were trimmed by quality using the
SolexaQA package (ver. 1.12) with default parameters and all trimmed reads
less than 25 bp were discarded (Cox et al. 2010). The quality trimmed reads
were then mapped to the D. melanogaster reference genome (FlyBase version
5.41) using Bowtie 2 (ver. beta 4) with the “very sensitive” and “-N=1” parameters
(Salzberg and Langmead, 2012). Following mapping, the GATK (ver. 1.1-23,
dePristo et al., 2011) IndelRealigner tool was used to perform local realignments
around indels and PCR and optical duplicates were identified with the
MarkDuplicates tool in the Picard package (http://picard.sourceforge.net).
SNP calling, phasing, and filtering
SNP variants were identified in all lines simultaneously using the GATK
UnifiedGenotyper (ver. 2.1-8) tool with all parameters set to
60
recommended default values. The raw SNP calls were further filtered following
the GATK best practices recommendations (Auwera et al., 2013) resulting in
4,021,717 SNPs. We then used BEAGLE to perform haplotype phasing as well
as impute missing data (Browning and Browning, 2007; Browning and Browning,
2009). SNPs were further filtered using VCFtools (http://vcftools.sourceforge.net/)
for 5% minor allele frequency and biallelic sites resulting in 1,047,913 SNPs
across the major chromosomal regions: 2L (222,464 SNPs), 2R (192,120 SNPs)
, 3L (212,601 SNPs), 3R (268,701 SNPs), and X (152,027 SNPs) to be
considered for further analysis.
Population statistics
We used VCFtools (Danecek et al., 2011) to calculate F
ST
via the Weir and
Cockerham estimates (1984). We also calculated nucleotide diversity, π, in 5000
bp windows across the genome. Tajima’s D in 10,000 bp windows was also
computed using VCFtools (TABLE A1).
Population structure analysis
We used the R package SNPRelate (Zheng et al., 2012) to perform principal
component analysis. We did PCA with all populations and then removed the
61
Cameroon population for another PCA to investigate North American patterns
further without the influence of the African population.
Overall admixture proportions were estimated using a subset of SNPs in the
software ADMIXTURE (Alexander et al., 2009), which performs maximum
likelihood estimation of individual ancestries from multilocus SNP genotype
datasets. The underlying theory is the same as STRUCTURE (Pritchard et al.,
2000), but has a faster run time to compute estimates due to algorithmic
optimizations.
Chromosome painting
We utilized the software Chromopainter (Lawson et al., 2012) to estimate which
parts of the genome each North American individual were contributed by
European or African ancestors. We ran Chromopainter for 60 iterations to
estimate parameters of the algorithm and then ran Chromopainter with the
estimated parameters to obtain the final results. We used a hierarchical
clustering algorithm to examine the similarity of Chromopainter results across
each chromosomal region between all the North American individuals.
62
Linkage Disequilibrium Analysis
To look at linkage disequilibrium decay over genomic distance, measures of D’
were estimated using VCFtools (Danecek et al., 2011) in 10,000 bp windows
across the genome.
4.3 Results
Investigating Population Structure by Principal Component Analysis
To explore initial relationships between populations, we performed PCA on the
4,021,717 quality-filtered SNPs using the R package SNPRelate. The first
principal component represented the separation between African and non-African
populations and the second principal component was the variation within the
Cameroon population (FIGURE 4.2). Upon closer inspection of the non-African
cluster (FIGURE 4.3), the first principal component could have also been a proxy
to how genetically close each non-African population is to the Cameroon
population with the Caribbean population located the closest. The non-African
populations were roughly grouped into two sub-clusters of Caribbean and non-
Caribbean. There were a few Caribbean fly lines that clustered close to and
within the non-Caribbean group. The four Caribbean lines that clustered with the
US populations were collected from locations on islands
63
closest to the US and Caribbean border (i.e. Freeport, Grand Bahamas-west and
Bullock’s Harbor, Berry Islands). Along with the four Caribbean lines, the
sequenced fly lines from locations in the southeast United States were
interspersed with fly lines from Raleigh, NC, USA indicating a potential east coast
US admixture zone. The Raleigh population clustered very closely with the
Winters, CA population, but both Raleigh and Winters appeared to still have
distinct populations. The 20 French lines appeared dispersed in the whole non-
Caribbean cluster, which supports the notion that there is much European
influence in North American populations.
64
FIGURE 4.2: First and second principal components (eigenvectors) of PCA with populations from
Cameroon (CAM), Caribbean Islands (CAR), France (FRA), Raleigh (RAL), southeast US (SEU),
and Winters (WIN)
65
FIGURE 4.3: Zoomed in view of North American plus French populations, first and second
principal components of PCA with populations from Cameroon (CAM-not pictured), Caribbean
Islands (CAR), France (FRA), Raleigh (RAL), southeast US (SEU), and Winters (WIN)
66
FIGURE 4.4: First four principal components of PCA including populations from Cameroon,
Caribbean Islands, France, Raleigh, southeast US, and Winters reveal that most variation
explained is within the Cameroon population
Upon inspection of additional principal components (FIGURE 4.4), principal
components 3 and 4 explained variation within the Cameroon population
indicating there was much diversity in the African population, which may have
been masking patterns in the non-African populations. We removed the
Cameroon population and performed a second PCA using non-African population
(FIGURE 4.5). The first principal component after removing the Cameroon
population explained the variation within the North American populations, while
the second principal component separated the French population from the North
American populations. Clustering patterns of the second PC analysis with
Cameroon removed were similar in the first PC analysis, but
67
we did that the French population formed a distinct cluster and was located
closest to the cluster containing the Winters, Raleigh, and southeast US
populations. The third and fourth principal components accounted for more
variation within the North American populations (FIGURE 4.6).
FIGURE 4.5: First and second principal components (eigenvectors) of PCA with the Cameroon
population removed, but including populations from Caribbean Islands (CAR), France (FRA),
Raleigh (RAL), southeast US (SEU), and Winters (WIN)
68
FIGURE 4.6: First four principal components of PCA with the Cameroon population removed, but
including populations from Caribbean Islands, France, Raleigh, southeast US, and Winters
Genetic differentiation between populations
To quantify general genetic relationships, we calculated Weir and Cockerham
estimates (1984) for F
ST
between all pairs of populations per SNP and averaged
the F
ST
estimates per chromosomal region. We found a consistent pattern that
Cameroon highly differentiated from all cosmopolitan populations, but was
closest to the Caribbean population (TABLE 4.1, 4.2, 4.3). The French and
Winters populations were the most differentiated from the Cameroon lines. The
greatest differentiation between the Cameroon population and the non-African
populations was on the X chromosome (TABLE 4.3) as
69
expected since the X chromosome evolves faster than the autosomes
(Presgraves, 2008).
The French population was the least genetically differentiated from the Winters
and Raleigh populations (TABLE 4.1, 4.2, 4.3). Interestingly enough, the
Caribbean population was slightly more differentiated from the Winters
population than from the French population in the 2L and 3R chromosomal
regions (TABLE 4.1, 4.2), perhaps indicating a slightly larger European influence
in the Caribbean than the west coast US.
CHR 2 FRA WIN RAL SEUS CAR CAM
FRA 0.03479 0.04292 0.06273 0.07378 0.17891
WIN 0.03903 0.02220 0.03786 0.04657 0.14145
RAL 0.03795 0.02431 0.02602 0.02985 0.11839
SEUS 0.05114 0.05487 0.03238 0.03815 0.13094
CAR 0.07435 0.09362 0.05798 0.04345 0.11311
CAM 0.14316 0.15647 0.11718 0.11065 0.09201
TABLE 4.1: Average F
ST
values between populations for chromosome 2 divided by regions 2L
(below diagonal) and 2R (above diagonal)
70
CHR 3 FRA WIN RAL SEUS CAR CAM
FRA 0.06091 0.04787 0.05638 0.08436 0.15828
WIN 0.03119 0.02519 0.05181 0.10992 0.19523
RAL 0.03769 0.02126 0.04005 0.08867 0.17283
SEUS 0.05548 0.03855 0.02791 0.05206 0.15105
CAR 0.07135 0.05789 0.03810 0.04468 0.13220
CAM 0.17481 0.16001 0.13335 0.14126 0.11451
TABLE 4.2: Average F
ST
values between populations for chromosome 3 divided by regions 3L
(below diagonal) and 3R (above diagonal)
X vs A FRA WIN RAL SEUS CAR CAM
FRA 0.04148 0.04374 0.05430 0.07596 0.16379
WIN 0.03924 0.02324 0.04577 0.07700 0.16329
RAL 0.04972 0.02218 0.03159 0.05365 0.13544
SEUS 0.07561 0.03851 0.03116 0.04459 0.13348
CAR 0.09046 0.05042 0.03761 0.04903 0.11295
CAM 0.28974 0.25222 0.23638 0.24526 0.23037
TABLE 4.3: Average F
ST
values between populations for chromosome X (lower diagonal) and all
autosomes (upper diagonal)
Structure of populations to investigate possible admixture patterns
Given a predetermined number of populations, we investigated the underlying
population structure in all our populations to understand the
71
extent of possible admixture present in North American populations. Despite the
number of populations, K, being two or three, the Cameroon population was a its
own distinct population and did not appear to be influenced by any other
population in all chromosomal regions (FIGURE 4.7, 4.8, 4.9, 4.10, 4.11). The
French population also seemed to be a distinct population for K = 2 and K =3
with influences from the Cameroon population, which is probable with the French
population’s proximity to Africa. When K = 3, the third color that emerges
seemed to associate with just North American populations, which may indicate
recent adaptations post-migration from Europe and Africa.
The southeast US and Caribbean population harbor the most influence from the
Cameroon population with some sequenced lines completely identifying with the
Cameroon population while others containing a large proportion of Cameroon-
classified SNPs. In particular, the 3R chromosomal region revealed a clinal
pattern of decreasing African influence with increasing latitude in the southeast
US and Caribbean populations (FIGURE 4.10). The Winters population seemed
to be heavily influenced by European genetic influences.
72
FIGURE 4.7: ADMIXTURE results of chromosomal region 2L for K = 2 and K = 3 number of
distinct populations represented by different colors. Blocks represent from left to right French,
Winters, Raleigh, southeast US and Caribbean, and Cameroon populations.
FIGURE 4.8: ADMIXTURE results of chromosomal region 2R for K = 2 and K = 3 number of
distinct populations represented by different colors. Blocks represent from left to right French,
Winters, Raleigh, southeast US and Caribbean, and Cameroon populations.
73
FIGURE 4.9: ADMIXTURE results of chromosomal region 3L for K = 2 and K = 3 number of
distinct populations represented by different colors. Blocks represent from left to right French,
Winters, Raleigh, southeast US and Caribbean, and Cameroon populations.
FIGURE 4.10: ADMIXTURE results of chromosomal region 3R for K = 2 and K = 3 number of
distinct populations represented by different colors. Blocks represent from left to right French,
Winters, Raleigh, southeast US and Caribbean, and Cameroon populations.
74
FIGURE 4.11: ADMIXTURE results of chromosome X for K = 2 and K = 3 number of distinct
populations represented by different colors. Blocks represent from left to right French, Winters,
Raleigh, southeast US and Caribbean, and Cameroon populations.
Chromosome painting reveals genome-wide African and European influences
While results from ADMIXTURE are useful in understanding how populations are
structured and points towards approximate the influences of African and
European ancestors, we cannot determine the pattern of influence across a
genome with those results. We used Chromopainter to predict the ancestry of all
the North American sequenced fly lines across the genome. The most striking
result from visualizing the local ancestry of all genomes (FIGURE 4.12, 4.13,
4.14, 4.15, 4.16) was that larger chunks of African or European ancestry seemed
to be retained in telomeric and centromeric regions known to have low
recombination (Comeron et al., 2012).
75
When we clustered individual genomes by heat signature, individuals within one
population clustered more with each other than with other populations except for
chromosomal region 2R where Caribbean and southeast US individuals seem to
be evenly dispersed between Winters and Raleigh populations. Chromosome X
appeared to be the least influenced by African ancestry (FIGURE 4.16), which is
in agreement with the large X effect (Presgraves, 2008).
Individuals from the Caribbean populations and some from the southeast US
seemed to have a larger percentage of African painted alleles, which was
especially apparent in the chromosomal regions of 2L and 3R (FIGURE 4.12,
4.15). The long stretches of the African-painted SNPs in these chromosomal
regions coincided with the locations of common cosmopolitan inversions, In(2L)t
and In(3R)P (Corbett-Detig and Hartl, 2012).
FIGURE 4.12: Painted 2L chromosomal region heatmap with hierarchical clustering of individuals
with telomeric region on the left and centromeric region on right. (Green: Winters, CA, Blue:
Raleigh, NC, Pink: Southeast US, Purple: Caribbean). Red represents SNPs that are most similar
to the Cameroon donor population; Yellow represents SNPs that are most similar to the French
donor population
76
FIGURE 4.13: Painted 2R chromosomal region heatmap with hierarchical clustering of individuals
with telomeric region on the right and centromeric region on the left. (Green: Winters, CA, Blue:
Raleigh, NC, Pink: Southeast US, Purple: Caribbean). Red represents SNPs that are most similar
to the Cameroon donor population; Yellow represents SNPs that are most similar to the French
donor population
FIGURE 4.14: Painted 3L chromosomal region heatmap with hierarchical clustering of individuals
with telomeric region on the left and centromeric region on the right. (Green: Winters, CA, Blue:
Raleigh, NC, Pink: Southeast US, Purple: Caribbean). Red represents SNPs that are most similar
to the Cameroon donor population; Yellow represents SNPs that are most similar to the French
donor population
FIGURE 4.15: Painted 3R chromosomal region heatmap with hierarchical clustering of individuals
with telomeric region on the right and centromeric region on the right (Green: Winters, CA, Blue:
Raleigh, NC, Pink: Southeast US, Purple: Caribbean). Red represents SNPs that are most similar
to the Cameroon donor population; Yellow represents SNPs that are most similar to the French
donor population
77
FIGURE 4.16: Painted X chromosomes heatmap with hierarchical clustering of individuals with
telomeric region on the right and centromeric region on the left. (Green: Winters, CA, Blue:
Raleigh, NC, Pink: Southeast US, Purple: Caribbean). Red represents SNPs that are most similar
to the Cameroon donor population; Yellow represents SNPs that are most similar to the French
donor population
Overall the expected proportion of probable African ancestry ranged between
3.6% (Winters, CA) to 47% (Caribbean island) for the painted genomes. On
average over the whole genome, the expected percentage of African ancestry
was highest in the Caribbean population at 24.75% and the lowest in the Winters
population at 8.68%. Raleigh and southeast US populations had 14% and 15.6%
of predicted African ancestry, which is consistent with previous findings of 15%
(Duchen et al., 2013). In summary, populations had decreasing African ancestry
with respect to distance from the Caribbean Islands in all genomic areas
(FIGURE 4.17). Out of all the chromosomes, the X chromosome had the lowest
expected percentage of African-inherited alleles for all populations (FIGURE
4.17).
78
FIGURE 4.17: Expected proportion of African ancestry for each population by chromosomal
region
Linkage disequilibrium patterns
Elevated levels of linkage disequilibrium can be an indicator of admixture in
populations because inherited ancestral tracts have not had sufficient time to be
broken down by recombination (Loh et al., 2013). We calculated D’ as a measure
of LD and averaged the absolute value of D’ to get approximate LD levels in our
populations across different genomic regions. We found that on average
Cameroon and France populations have lower LD values than North American
populations (FIGURE 4.18). Out of all the North American populations, the
Caribbean population had one of the lowest LD values on most chromosomal
regions except on the X chromosome. This is consistent with the notion that
African flies colonized the Caribbean Islands a good 200 years before European
79
flies arrived on the east coast of the US making the Caribbean population older
than the US populations (David and Capy, 1988).
FIGURE 4.18: Average D′ as a measure of linkage disequilibrium by population and
chromosome
4.4 Discussion
Understanding the origins of North American Drosophila melanogaster
populations is useful for researchers working with populations from this area of
the world. We have presented here the the genome analyses of southeast US
and Caribbean fly populations in relation to other North American populations
and to their African and European ancestral populations.
80
Caribbean flies established by African ancestors
Although all non-African population F
ST
values were high throughout the
genome, the Caribbean population had on average the lowest values compared
to the other non-African populations. With the Caribbean population located
closest in the first PC analysis to the Cameroon population and the highest
percentage of predicted African ancestry out of all the North American
populations, these pieces of evidence do seem to further support the migration
event of west African flies to the Caribbean islands via the transatlantic slave
trade (David and Capy, 1988).
African and European admixture in North America
Admixed populations exhibit more linkage disequilibrium than in older long-
established populations (Loh et al., 2013). This is because newer populations,
which are a combination of genetic material from older base populations have not
gone through enough generations for recombination to break down LD blocks.
We do detect higher LD in the North American populations than in our African
and European populations. Although, this is a common signature of admixture in
populations, higher LD values can also result from other events such as a
population bottleneck. To uncover the sources of elevated LD, we would need to
run different demographic models on our data to rule out LD by
81
population bottleneck or selection. However, previous studies have already
established the existence of admixture in North American populations,
particularly Raleigh, (Duchen et al, 2013) which would support that elevated LD
in our case is most likely due to admixture.
We are able to extend the admixture scenario in North America with our 23
sequenced genomes from the southeast US and Caribbean islands. It has been
postulated that American D. melanogaster are more genetically variable than
European D. melanogaster due to admixture from the Caribbean islands
(Caracristi and Schlötterer, 2003). Our results from ADMIXTURE (FIGURE 4.8,
4.9, 4.10, 4.11) and chromosome painting (FIGURE 4.12, 4.13, 4.14, 4.15, 4.15)
clearly show a clinal pattern of African introgression into the United States, which
supports the notion that these non-European African alleles in the US are
originating from the Caribbean Islands. Furthermore, the PCA groupings also
illustrates that the southeast US is where Caribbean and east coast US fly
populations are experiencing the most admixture.
Westward expansion of Drosophila melanogaster
Our analysis of the Winters, CA genomes revealed that the Winters population is
more related to our European population than the other US population. There
appears to be very little to no African ancestry in the genomes
82
from Winters, CA. Either there was a separate colonization event in the west or
when D. melanogaster arrived in North America with European settlers, it quickly
expanded west shortly after arriving. The latter explanation may be more
plausible given that the first sighting of D. melanogaster was in the mid-19th
century (David and Capy, 1988) and European settlers at that point in US history
were actively building railroads and expanding westward with a heavy agricultural
agenda.
4.5 Chapter References
Alexander DH, Novembre J, Lange K. Fast model-based estimation of ancestry
in unrelated individuals. Genome Research. Cold Spring Harbor Lab; 2009
Sep 1;19(9):1655–64.
Auwera GA, Carneiro MO, Hartl C, Poplin R, del Angel G, Levy Moonshine A, et
al. From FastQ Data to High‐Confidence Variant Calls: The Genome Analysis
Toolkit Best Practices Pipeline. John Wiley & Sons, Inc; 2013. p. 11.10.1–
11.10.33.
Ben Langmead, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nature
methods. Nature Publishing Group; 2012 Apr 1;9(4):357–9.
Browning BL, Browning SR. A Unified Approach to Genotype Imputation and
Haplotype-Phase Inference for Large Data Sets of Trios and Unrelated
Individuals. The American Journal of Human Genetics. 2009 Feb;84(2):210–
23.
Browning SR, Browning BL. Rapid and Accurate Haplotype Phasing and
Missing-Data Inference for Whole-Genome Association Studies By Use of
Localized Haplotype Clustering. The American Journal of Human Genetics.
2007 Nov;81(5):1084–97.
83
Campo D, Lehmann K, Fjeldsted C, Souaiaia T, Kao J, Nuzhdin SV. Whole-
genome sequencing of two North American Drosophila
melanogasterpopulations reveals genetic differentiation and positive selection.
Molecular Ecology. 2013 Sep 19;22(20):n/a–n/a.
Caracristi G, Schlötterer C. Genetic differentiation between American and
European Drosophila melanogaster populations could be attributed to
admixture of African alleles. Molecular Biology and Evolution. 2003;20(5):792.
Cockerham C. Estimating F-statistics for the analysis of population structure.
Evolution. 1984.
Comeron JM, Ratnappan R, Bailin S. The Many Landscapes of Recombination in
Drosophila melanogaster. PLoS Genet. Public Library of Science; 2012 Oct
11;8(10):e1002905.
Corbett-Detig RB, HARTL DL. Population Genomics of Inversion Polymorphisms
in Drosophila melanogaster. PLoS Genet. Public Library of Science; 2012 Dec
20;8(12):e1003056.
Cox MP, Peterson DA, Biggs PJ. SolexaQA: At-a-glance quality assessment of
Illumina second-generation sequencing data. BMC bioinformatics. BioMed
Central Ltd; 2010 Sep 27;11(1):485.
Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, et al. The
variant call format and VCFtools. Bioinformatics. Oxford University Press;
2011 Aug 1;27(15):2156–8.
David J, Capy P. Genetic variation of Drosophila melanogaster natural
populations. Trends Genet. 1988;4(4):106–11.
DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A
framework for variation discovery and genotyping using next-generation DNA
sequencing data. Nature Genetics. Nature Publishing Group; 2011 May
1;43(5):491–8.
Duchen P, Živković D, Hutter S, Stephan W, Laurent S. Demographic inference
reveals African and European admixture in the North American Drosophila
melanogaster population. Genetics. 2013.
Dunham JP, Friesen ML. A Cost-Effective Method for High-Throughput
Construction of Illumina Sequencing Libraries. Cold Spring Harb Protoc. Cold
Spring Harbor Laboratory Press; 2013 Sep
1;2013(9):pdb.prot074187–7.
84
Haddrill P and Bergman CM. 20 Drosophila melanogaster genomes from
Montpellier, France. 2012. http://bergmanlab.smith.man.ac.uk/?p=1685
Jeffrey R. Powell Department of Biological Sciences Yale University. Progress
and Prospects in Evolutionary Biology : The Drosophila Model. Oxford
University Press; 1997.
Lawson DJ, Hellenthal G, Myers S, Falush D. PLOS Genetics: Inference of
Population Structure using Dense Haplotype Data. PLoS Genet. 2012.
Li H, Stephan W. Inferring the demographic history and rate of adaptive
substitution in Drosophila. PLoS Genet. 2006 Oct 13;2(10):e166. PMCID:
PMC1599771
Loh P-R, Lipson M, Patterson N, Moorjani P, Pickrell JK, Reich D, et al. Inferring
Admixture Histories of Human Populations Using Linkage Disequilibrium.
Genetics. Genetics Society of America; 2013 Apr 1;193(4):1233–54.
Mackay TFC, Richards S, Stone EA, Barbadilla A, Ayroles JF, Zhu D, et al. The
Drosophila melanogaster Genetic Reference Panel. Nature. Nature Publishing
Group; 2012 Feb 8;482(7384):173–8.
Pool JE, Corbett-Detig RB, Sugino RP, Stevens KA, Cardeno CM, Crepeau MW,
et al. Population Genomics of Sub-Saharan Drosophila melanogaster: African
Diversity and Non-African Admixture. PLoS Genet. Public Library of Science;
2012 Dec 20;8(12):e1003080.
Presgraves DC. Sex chromosomes and speciation in Drosophila. Trends in
Genetics. 2008.
Pritchard J, Stephens M, Donnelly P. Inference of population structure using
multilocus genotype data. Genetics. 2000;155(2):945.
Yukilevich R, True JR. African morphology, behavior and phermones underlie
incipient sexual isolation between US and Caribbean Drosophila melanogaster.
Evolution. 2008a Nov 1;62(11):2807–28.
Yukilevich R, True JR. Incipient sexual isolation among cosmopolitan Drosophila
melanogaster populations. Evolution. Blackwell Publishing Inc; 2008b Aug
1;62(8):2112–21.
85
Yukilevich R, Turner TL, Aoki F, Nuzhdin SV, True JR. Patterns and Processes
of Genome-Wide Divergence Between North American and African Drosophila
melanogaster. Genetics. Genetics Society of America; 2010 Sep
1;186(1):219–39.
Zheng X, Levine D, Shen J, Gogarten SM, Laurie C, Weir BS. A high-
performance computing toolset for relatedness and principal component
analysis of SNP data. Bioinformatics. Oxford University Press; 2012 Dec
1;28(24):3326–8.
86
Chapter 5
Santa Catalina Island as a study system for interspecies
interactions: Island-wide survey of Drosophila species
5.1 Introduction
Santa Catalina Island is part of the Channel Islands located off of the southern
coast of the U.S. state of California. The island is approximately 35.4 kilometers
long and 12.9 kilometers across and has two towns on either end of the island:
Two Harbors, on the north side, and Avalon, on the south side. The proximity to
several research institutions and research facilities on the island provided by the
Wrigley Marine Science Center and the Catalina Island Conservancy adds to the
convenience of field research on the island. Both marine and terrestrial
researches have been conducted on the island with an emphasis on the marine
front. The vertebrate and plant species on the island are well documented, but
with regards to invertebrates, the documentation is sparse. Some studies have
focused on specific arthropods endemic to the island, but there have been no
island-wide surveys of arthropod species on the island. Out of the arthropods on
the island, one of the most studied are the Drosophila flies (Reed et al. 2008;
Hurtado et al. 2004) and in particular the repleta species group,
87
which utilize cacti as their plant host. Though having previously been studied, the
spatial and temporal aspects of this species group (as well as other Drosophila
species) has not been thoroughly investigated.
Previous collection data and studies suggest that at least eight different
Drosophila species live on the island (list from the Drosophila Species Stock
Center, San Diego, CA). These species include D. melanogaster, D. simulans,
D. hamatofila, D. mojanvensis, D. mettleri, D. mainland, D. pseudoobscura, and
D. wheeleri. However, the spatial distribution of these species on the island is not
well known. From an informal island-wide insect survey conducted April 2011
(unpublished data), we have data suggesting that species distribution on the
island is not ubiquitous despite the prevalence of host cacti on the island
meaning that certain parts of the island harbor primarily more of one species over
another species. We returned to the island for a formal in depth island-wide
survey of Drosophila species in July and August 2012. According to our latest
collections on the island at 24 sites, we have determined where the repleta
subspecies group of flies is most abundant on the island and where they are
least abundant. We noted not only the locations of collection sites, but also the
elevation because altitudinal factors also play a role in population density
(Guruprasad et al. 2010). We have also conducted seasonal collections at select
locations on the island to roughly determine species composition over a year
88
since seasonality also plays a role in density of populations (Guruprasad et al.
2010; Torres & Madi-Ravazzi 2006).
5.2 Materials and Methods
Island-wide collections
In the months of July 2012 and August 2012, we assembled fly traps made from
plastic water and soda bottles baited with a mix of rotten banana mash, yeast,
and Opuntia cactus powder. [Markow & O’Grady 2006] This particular bait was
found in preliminary trapping to be the most efficient in attracting the most
specimens and species compared to other baits that were tried. Two traps were
distributed per site at 24 sites across Catalina Island (FIGURE 5.1). After 24
hours of trap deployment, the traps were subsequently collected. Flies were
removed from the traps and sorted into repleta species subgroup and non-repleta
species subgroup.
89
FIGURE 5.1: Map of collection sites on Santa Catalina Island.
Seasonal collections
In addition to summer collections, we returned to our sites at WMSC and Little
Harbor campgrounds in November 2012, January 2013, and April 2013 for fall,
winter, and spring collections. Flies were also collected for an
90
additional summer season in July 2013 at WMSC. Traps were assembled,
deployed, and retrieved as previously described.
Species identification
The species in the repleta subgroup are nontrivial to distinguish because
pigmentation patterns are very similar between these species. Only males can be
visually identified by looking at the morphology of the aedeagus via genitalia
dissections. Females of this subgroup are not distinguishable from each other
except by establishing isofemale lines and examining male progeny. Collected
males were dissected for species identification and females were preserved in
ethanol and are available for sequencing. Some of the females caught were
mailed to collaborators at Texas A&M University (TAMU) for identification by
establishing isofemale lines and examining male progeny.
Non-repleta species are fairly distinct in pigmentation and morphology and were
examined under a dissecting scope for identification.
91
5.3 Results
Density and distribution of repleta species
From the initial summer season (2012), we dissected 279 males and preserved
177 females in ethanol. Another 120 females were mailed to collaborators at
Texas A&M University (TAMU) for identification after the establishment of
isofemale lines. Species compositions at each of the sites were found to be
statistically different from each other (TABLE 5.1, Chi-squared, p <0.0001). We
classified collection sites as “hot”, “warm”, or “cold” according to the number of
flies present in traps. Hot spots are defined as locations where a trap collected
more than 20 male flies. Warm spots are defined as locations where a trap
contained 10 to 20 male flies. Cold spots are sites where there were less than
five flies were present in the traps set. We identified three collection hot spots on
the island and two warm spots. The three collection hot spots on the island
include Starlight Beach, Wrigley Marine Science Center (WMSC), and Little
Harbor Campgrounds. The two warm spots were at Eagle’s Nest Lodge and
Middle Ranch Junction. Out of the four repleta species identified, the most
abundant were D. mainlandi, D. mettleri, and D. hamatofila. Only four D.
mojavensis males were collected.
92
TABLE 5.1: Species compositions at each sampling site from summer 2012 collections.
Site
Name
Site
#
Lat. Lon. Elev. mainlandi
hamatofila
mettleri
mojavensis
Site
total:
Two
Harbors
1 33.44 -118.49 61' 3 3 2 0 8
Cherry
Cove
2 33.45 -118.51 201' 1 0 0 0 1
Emerald
Bay
3 33.46 -118.53 106' 2 0 0 0 2
Parson’s
Landing
4 33.47 -118.54 86' 1 0 0 0 1
WMSC 5 33.44 -118.48 66' 11 28 28 0 67
El
Rancho
Escondid
o
6 33.38 -118.45 538' 2 4 0 0 6
Airport in
the Sky
7 33.40 -118.41 1589' 2 2 0 2 6
Soapston
e
Trailhead
8 33.40 -118.41 1547' 2 2 0 0 4
Blackjack
Junction
9 33.39 -118.38 1391' 0 0 0 0 0
Middle
Ranch
Junction
10 33.36 -118.37 1405' 6 7 3 0 16
Avalon
Gate
11 33.34 -118.33 551' 6 1 0 0 7
Avalon 12 33.32 -118.34 391' 1 0 0 0 1
Fruit
trees
13 33.34 -118.40 833' 5 2 2 0 9
Little
Harbor
Campgro
und
14 33.38 -118.47 60' 25 28 30 0 83
Howland'
s
Landing
15 33.45 -118.52 74' 0 0 0 0 0
Road to
Starlight
Beach
16 33.47 -118.57 543' 0 0 0 0 0
Starlight
Beach
17 33.47 -118.58 117' 12 11 9 1 33
Junction
to Silver
Peak
18 33.46 -118.58 670' 1 3 1 0 5
Silver
Peak
19 33.46 -118.56 1791' 0 0 0 0 0
Junction
at
Fencelin
e Road
20 33.45 -118.55 1570' 1 1 0 0 2
"Machete
path"
21 33.44 -118.52 1449' 0 2 1 0 3
93
Gate at
Silver
Peak
Trail
22 33.43 -118.51 454' 0 2 0 0 2
Middle
Ranch
23 33.35 -118.44 625' 2 1 0 0 3
Eagle’s
Nest
Lodge
24 33.35 -118.45 498' 3 14 1 0 18
Speci
es
totals:
86 111 77 3
TABLE 5.1 (continued): Species compositions at each sampling site from summer 2012
collections
The species compositions at the collection hot and warm spots were compared.
Sites with less than 10 flies were excluded from the comparison. Species
compositions at different sites across the island were varied as shown in TABLE
5.1. Little Harbor campgrounds and Starlight beach were very similar in species
composition with D. mainlandi, D. hamatofila, and D. mettleri in approximately
equal proportions. At Eagle’s Nest Lodge, D. hamatofila was the main species
collected and at Middle Ranch Junction D. mainlandi was the most abundant
species. D. hamatofila and D. mettleri were dominant at the WMSC. It appears
that there is a significant correlation of D. hamatofila, D. mettleri, and D.
mainlandi occurring together at collection sites. There also appears to be a
marginally insignificant negative trend of number of flies collected and elevation
(TABLE 5.2).
94
Elevation D.
mainlandi
D.
hamatofila
D. mettleri D.
mojavensis
Site totals
Elevation
-0.345519 -0.3303370 -0.3557935 0.19761726 -
0.3517674
D. mainlandi 0.09819 0.8455475 0.88083836 0.09086256 0.9322748
D.
hamatofila
0.1149 <0.001 0.93989023 0.01360863 0.9712776
D. mettleri 0.08795 <0.001 <0.001 -0.00741057 0.9806853
D.
mojavensis
0.3546 0.6728 0.9497 0.9726 0.0477909
Site totals 0.09186 <0.001 <0.001 <0.001 0.8245
TABLE 5.2: Correlations and associated p-values of Drosophila species on Catalina Island.
Pearson’s r is in the top half of the table and associated p-values are in the lower half of the table.
Significant p-values and associated correlations are highlighted in light grey.
For the fall collection time point, we found only one D. mainlandi male at WMSC
and 12 D. mainlandi males at Little Harbor Campgrounds. No flies were caught in
the winter and very few D. mainlandi males were collected in the spring with four
males at Little Harbor Campgrounds and one male at WMSC. Our spring
collections were far lower than what we collected in an island-wide preliminary
collection in spring of 2011 (unpublished data). In the following summer season
(2013), we collected a total of 168 male repleta species specimens from the
WMSC sites. There were 81 D. mainland, 55 D. mettleri, 25 D. hamatofila, 4 D.
95
mojavensis, and 3 D. wheeleri. These species compositions were vastly different
from the composition in the previous summer at WMSC (FIGURE 5.2).
FIGURE 5.2: Species composition at the Wrigley Marine Science Center between summers.
Non-repleta species
Drosophila melanogaster and Drosophila simulans were found on the island at
many of the sites and males were identified based on genitalia morphology under
a scope.
One unidentified male was collected from one of our sites at Emerald Bay during
the summer collections (FIGURE 5.3). After consulting with the
96
Drosophila Species Stock Center in San Diego, CA, the specimen had tergite
pigmentation most similar to Drosophila busckii. However, thorax pigmentation
was darker than the species standard. Efforts of collecting more flies like the
unidentified specimen in the summer were unsuccessful, but a second male
specimen was collected when we returned for collections in the fall at WMSC.
97
FIGURE 5.3: Photos of unknown specimen
98
5.4 Discussion
We have assembled information on the distribution of fly species on Santa
Catalina Island for future field researchers interested in collecting specimens
from the Drosophila repleta species group. We have also sighted D.
melanogaster and D. simulans on the island as well as possibly identified a new
species not previously seen before on the island.
We have shown that some species are more prevalent on some parts of the
island than others. Furthermore, these results may point towards a seasonality of
overall fly population with the summer season being the best for specimen
collection in terms of numbers and diversity. Comparing species composition
within the same location and season between years may be quite volatile as well
according to our spring and summer sampling data. Reasons for why our spring
collections were so low this season could be due to the general lower rainfall
compared to the previous couple of years according to the National Climatic Data
Center. The disparate values between years suggest that species composition
between seasons and years is highly sensitive to environmental factors. More
sampling at regular time points over a few years would be needed to determine if
this were the case.
99
D. melanogaster and D. simulans are known human commensals and are
abundant in areas with high human occupancy due to the production of food
waste [Powell 1997]. Therefore, it was surprising to find that the towns on
Catalina (i.e. Avalon and Two Harbors) had very low yields of fly collections in all
species including the human commensal species. We attempted trapping on
multiple occasions for varying lengths of time to confirm these results. One
hypothesis why Two Harbors is devoid of flies is due to the “isthmus fan”, which
is a strong wind that blows through the town every afternoon (personal
communication with island residents). Reasons why Avalon has a low population
of flies are unclear. Further investigation into weather and wind patterns and
possible pesticide use might give clues as to why these areas are not inhabited
by many flies.
5.5 Chapter References
Guruprasad BR, Hegde SN, Krishna MS. Seasonal and altitudinal changes in
population density of 20 species of Drosophila in Chamundi hill. Journal of
Insect Science. 2010.
Hurtado LA, Erez T, Castrezana S, Markow TA. Contrasting population genetic
patterns and evolutionary histories among sympatric Sonoran Desert
cactophilic Drosophila. Molecular Ecology. Blackwell Science Ltd; 2004 Jun
1;13(6):1365–75.
Markow TA, O'Grady P. Drosophila. Academic Press; 2005.
100
Powell JR. Progress and Prospects in Evolutionary Biology : The Drosophila
Model. Oxford University Press; 1997.
Reed LK, LaFlamme BA, Markow TA. Genetic Architecture of Hybrid Male
Sterility in Drosophila: Analysis of Intraspecies Variation for Interspecies
Isolation. PLoS ONE. Public Library of Science; 2008 Aug 27;3(8):e3076.
Torres FR, Madi-Ravazzi L. Seasonal variation in natural populations of
Drosophila spp. (Diptera) in two woodlands in the State of São Paulo, Brazil.
Iheringia. Série Zoologia. Fundação Zoobotânica do Rio Grande do Sul; 2006
Dec 1;96(4):437–44.
101
References
Arbuthnott D, Dutton EM, Agrawal AF, Rundle HD. The ecology of sexual
conflict: ecologically dependent parallel evolution of male harm and female
resistance in Drosophila melanogaster. Ecol. Lett. 2014 Feb 1;17(2):221–8.
Arnqvist G, Rowe L. Sexual conflict: monographs in behavior and ecology.
Princeton, NJ: Princeton University Press; 2005.
Auwera GA, Carneiro MO, Hartl C, Poplin R, del Angel G, Levy Moonshine A, et
al. From FastQ Data to High‐Confidence Variant Calls: The Genome Analysis
Toolkit Best Practices Pipeline. John Wiley & Sons, Inc; 2013. p. 11.10.1–
11.10.33.
Ben Langmead, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nature
methods. Nature Publishing Group; 2012 Apr 1;9(4):357–9.
Browning BL, Browning SR. A Unified Approach to Genotype Imputation and
Haplotype-Phase Inference for Large Data Sets of Trios and Unrelated
Individuals. The American Journal of Human Genetics. 2009 Feb;84(2):210–
23.
Browning SR, Browning BL. Rapid and Accurate Haplotype Phasing and
Missing-Data Inference for Whole-Genome Association Studies By Use of
Localized Haplotype Clustering. The American Journal of Human Genetics.
2007 Nov;81(5):1084–97.
Campo D, Lehmann K, Fjeldsted C, Souaiaia T, Kao J, Nuzhdin SV. Whole-
genome sequencing of two North American Drosophila
melanogasterpopulations reveals genetic differentiation and positive selection.
Molecular Ecology. 2013 Sep 19;22(20):n/a–n/a.
Caracristi G, Schlötterer C. Genetic differentiation between American and
European Drosophila melanogaster populations could be attributed to
admixture of African alleles. Molecular Biology and Evolution. 2003;20(5):792.
Chapman T, Bangham J, Vinti G, Seifried B, Lung O, Wolfner MF, et al. The sex
peptide of Drosophila melanogaster: Female post-mating responses analyzed
by using RNA interference. Proceedings of the National Academy of Sciences.
National Acad Sciences; 2003 Aug 19;100(17):9923–8.
102
Civetta A, Clark AG. Correlated effects of sperm competition and postmating
female mortality. Proceedings of the National Academy of Sciences. National
Acad Sciences; 2000;97(24):13162–5.
Clodong S, Dühring U, Kronk L, Wilde A, Axmann I, Herzel H, et al. Functioning
and robustness of a bacterial circadian clock. Molecular Systems Biology.
John Wiley & Sons, Ltd; 2007 Jan 1;3(1).
Cockerham C. Estimating F-statistics for the analysis of population structure.
Evolution. 1984.
Comeron JM, Ratnappan R, Bailin S. The Many Landscapes of Recombination in
Drosophila melanogaster. PLoS Genet. Public Library of Science; 2012 Oct
11;8(10):e1002905.
Corbett-Detig RB, HARTL DL. Population Genomics of Inversion Polymorphisms
in Drosophila melanogaster. PLoS Genet. Public Library of Science; 2012 Dec
20;8(12):e1003056.
Costa R, Peixoto AA, Barbujani G, Kyriacou CP. A Latitudinal Cline in a
Drosophila Clock Gene. Proc Biol Sci. The Royal Society; 1992 Oct
22;250(1327):43–9.
Cox MP, Peterson DA, Biggs PJ. SolexaQA: At-a-glance quality assessment of
Illumina second-generation sequencing data. BMC bioinformatics. BioMed
Central Ltd; 2010 Sep 27;11(1):485.
Coyne JA, Orr HA. Patterns of speciation in Drosophila. Evolution. 1989.
Coyne JA, Orr HA. Speciation. Sinauer Associates Incorporated; 2004.
Cozzolino S, Scopece G. Specificity in pollination and consequences for
postmating reproductive isolation in deceptive Mediterranean orchids. … of
the Royal …. 2008.
Dallerac R, Labeur C, Jallon J-M, Knipple DC, Roelofs WL, Wicker-Thomas C. A
Δ9 desaturase gene with a different substrate specificity is responsible for the
cuticular diene hydrocarbon polymorphism in Drosophila melanogaster.
Proceedings of the National Academy of Sciences. National Acad Sciences;
2000 Aug 15;97(17):9449–54.
Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, et al. The
variant call format and VCFtools. Bioinformatics. Oxford University Press;
2011 Aug 1;27(15):2156–8.
103
David J, Capy P. Genetic variation of Drosophila melanogaster natural
populations. Trends Genet. 1988;4(4):106–11.
DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A
framework for variation discovery and genotyping using next-generation DNA
sequencing data. Nature Genetics. Nature Publishing Group; 2011 May
1;43(5):491–8.
Dibner C, Schibler U, Albrecht U. The mammalian circadian timing system:
organization and coordination of central and peripheral clocks. Annual review
of physiology. 2010.
Duchen P, Živković D, Hutter S, Stephan W, Laurent S. Demographic inference
reveals African and European admixture in the North American Drosophila
melanogaster population. Genetics. 2013.
Dunham JP, Friesen ML. A Cost-Effective Method for High-Throughput
Construction of Illumina Sequencing Libraries. Cold Spring Harb Protoc. Cold
Spring Harbor Laboratory Press; 2013 Sep 1;2013(9):pdb.prot074187–7.
Endler JA. Geographic Variation, Speciation, and Clines. Princeton, NJ:
Princeton University Press; 1977.
EWERS J, HAMBLENCOYLE M, ROSBASH M, HALL JC. Requirement for
Period Gene-Expression in the Adult and Not During Development for
Locomotor-Activity Rhythms of Imaginal Drosophila-Melanogaster. Journal of
Neurogenetics. 1990;7(1):31–&.
Fowler K, Partridge L. A cost of mating in female fruitflies. , Published online: 27
April 1989; | doi:10.1038/338760a0. Nature Publishing Group; 1989 Apr
27;338(6218):760–1.
Greenspan RJ. An Introduction to Nervous Systems. CSHL Press; 2007.
Guruprasad BR, Hegde SN, Krishna MS. Seasonal and altitudinal changes in
population density of 20 species of Drosophila in Chamundi hill. Journal of
Insect Science. 2010.
Haddrill P and Bergman CM. 20 Drosophila melanogaster genomes from
Montpellier, France. 2012. http://bergmanlab.smith.man.ac.uk/?p=1685
Hardeland R. Species differences in the diurnal rhythmicity of courtship
behaviour within the Melanogaster group of the genus Drosophila. Animal
Behaviour. 1972 Feb;20(1):170–4.
104
Harrison RG. Barriers to Gene Exchange Between Closely Related Cricket
Species. II. Life Cycle Variation and Temporal Isolation. Evolution. 1985
Mar;39(2):244.
Harrison RG. Hybrid zones: windows on evolutionary process. Oxford Surveys in
Evolutionary Biology. 1990;7:69–128.
Harrison RG. Hybrid Zones and the Evolutionary Process. New York, NY: Oxford
University Press; 1993.
Helfrich Förster C. Neurobiology of the fruit fly's circadian clock. Genes, Brain
and Behavior. Munksgaard International Publishers; 2005 Mar 1;4(2):65–76.
Holland B, Rice WR. Perspective: Chase-away sexual selection: Antagonistic
seduction versus resistance. Evolution. 1998 Feb;52(1):1–7.
Hollocher H, Ting CT, Wu ML, Wu CI. Incipient speciation by sexual isolation in
Drosophila melanogaster: Extensive genetic divergence without reinforcement.
Genetics. 1997 Nov;147(3):1191–201.
Hurtado LA, Erez T, Castrezana S, Markow TA. Contrasting population genetic
patterns and evolutionary histories among sympatric Sonoran Desert
cactophilic Drosophila. Molecular Ecology. Blackwell Science Ltd; 2004 Jun
1;13(6):1365–75.
Huxley JS. Clines: an auxiliary taxonomic principle. Nature. 1938.
Johnstone RA, Keller L. How males can gain by harming their mates: sexual
conflict, seminal toxins, and the cost of mating. The American Naturalist. 2000.
Klein JP, Moeschberger ML. Survival Analysis. Springer; 2003.
Kleinbaum DG, Klein M. Kaplan-Meier survival curves and the log-rank test.
Survival analysis; 2012.
Knowles LL, Markow TA. Sexually antagonistic coevolution of a postmating-
prezygotic reproductive character in desert Drosophila. Proceedings of the
National Academy of Sciences. National Acad Sciences; 2001 Jul
17;98(15):8692–6.
Konopka RJ, Benzer S. Clock Mutants of Drosophila melanogaster. Proceedings
of the National Academy of Sciences. National Acad Sciences; 1971 Sep
1;68(9):2112–6.
105
Kyriacou CP, HALL JC. Circadian-Rhythm Mutations in Drosophila-Melanogaster
Affect Short-Term Fluctuations in the Males Courtship Song. Proceedings of
the National Academy of Sciences. 1980;77(11):6729–33.
Kyriacou CP, Peixoto AA, Sandrelli F, Costa R, Tauber E. Clines in clock genes:
fine-tuning circadian rhythms to the environment. Trends in Genetics. 2008
Mar;24(3):124–32.
Lawson DJ, Hellenthal G, Myers S, Falush D. PLOS Genetics: Inference of
Population Structure using Dense Haplotype Data. PLoS Genet. 2012.
Li H, Stephan W. Inferring the demographic history and rate of adaptive
substitution in Drosophila. PLoS Genet. 2006 Oct 13;2(10):e166. PMCID:
PMC1599771
Loh P-R, Lipson M, Patterson N, Moorjani P, Pickrell JK, Reich D, et al. Inferring
Admixture Histories of Human Populations Using Linkage Disequilibrium.
Genetics. Genetics Society of America; 2013 Apr 1;193(4):1233–54.
Mackay TFC, Richards S, Stone EA, Barbadilla A, Ayroles JF, Zhu D, et al. The
Drosophila melanogaster Genetic Reference Panel. Nature. Nature Publishing
Group; 2012 Feb 8;482(7384):173–8.
Markow TA, O'Grady P. Drosophila. Academic Press; 2005.
Martin OY, HOSKEN DJ. The evolution of reproductive isolation through sexual
conflict. Nature. 2003;423(6943):979–82.
Panda S, Hogenesch JB, Kay SA. Circadian rhythms from flies to human. Nature.
Nature Publishing Group; 2002 May 16;417(6886):329–35.
Pool J, Hellmann I, Jensen J, Nielsen R. Population genetic inference from
genomic sequence variation. Genome Research. 2010;20(3):291.
Powell JR. Progress and Prospects in Evolutionary Biology : The Drosophila
Model. Oxford University Press; 1997.
Presgraves DC. Sex chromosomes and speciation in Drosophila. Trends in
Genetics. 2008.
Presgraves DC. Darwin and the Origin of Interspecific Genetic Incompatibilities.
Am Nat. The University of Chicago Press; 2010 Dec 18;176(S1):S45–S60.
Pritchard J, Stephens M, Donnelly P. Inference of population structure using
multilocus genotype data. Genetics. 2000;155(2):945.
106
Reed LK, LaFlamme BA, Markow TA. Genetic Architecture of Hybrid Male
Sterility in Drosophila: Analysis of Intraspecies Variation for Interspecies
Isolation. PLoS ONE. Public Library of Science; 2008 Aug 27;3(8):e3076.
Rice WR. Sexually antagonistic male adaptation triggered by experimental arrest
of female evolution. Nature. 1996.
Rowe L, Arnqvist G. Sexually antagonistic coevolution in a mating system:
Combining experimental and comparative approaches to address evolutionary
processes. Evolution. 2002 Apr;56(4):754–67.
Sakai T, Ishida N. Circadian rhythms of female mating activity governed by clock
genes in Drosophila. Proceedings of the National Academy of Sciences.
National Acad Sciences; 2001 Jul 31;98(16):9221–5.
Sawyer LA, Hennessy JM, Peixoto AA, Rosato E, Parkinson H, Costa R, et al.
Natural variation in a Drosophila clock gene and temperature compensation.
Science. 1997;278(5346):2117–20.
Seehausen O, Butlin RK, Keller I, Wagner CE. Genomics and the origin of
species. Nature Reviews …. 2014.
Servedio M, Noor M. The role of reinforcement in speciation: theory and data.
Annual Review of Ecology, Evolution, and Systematics. 2003;:339–64.
Sokolove PG, Bushell WN. The chi square periodogram: Its utility for analysis of
circadian rhythms. Journal of Theoretical Biology. 1978 May;72(1):131–60.
Sota T, Kubota K. Genital Lock-and-Key as a Selective Agent against
Hybridization. Evolution. 1998 Oct;52(5):1507.
Svetec N, and Begun D. unpublished results
Takahashi A, Tsaur S-C, Coyne JA, Wu C-I. The nucleotide changes governing
cuticular hydrocarbon variation and their evolution in Drosophila melanogaster.
Proceedings of the National Academy of Sciences. National Acad Sciences;
2001 Mar 27;98(7):3920–5.
Torres FR, Madi-Ravazzi L. Seasonal variation in natural populations of
Drosophila spp. (Diptera) in two woodlands in the State of São Paulo, Brazil.
Iheringia. Série Zoologia. Fundação Zoobotânica do Rio Grande do Sul; 2006
Dec 1;96(4):437–44.
107
Wolfner MF. Tokens of love: Functions and regulation of drosophila male
accessory gland products. Insect Biochemistry and Molecular Biology. 1997
Mar;27(3):179–92.
Yasuda GK, Baker J, Schubiger G. Temporal regulation of gene expression in
the blastoderm Drosophila embryo. Genes Dev. Cold Spring Harbor Lab; 1991
Oct 1;5(10):1800–12.
Young MW, Kay SA. Time zones: a comparative genetics of circadian clocks. Nat
Rev Genet. Nature Publishing Group; 2001 Sep 1;2(9):702–15.
Yukilevich R, True JR. African morphology, behavior and phermones underlie
incipient sexual isolation between US and Caribbean Drosophila melanogaster.
Evolution. Blackwell Publishing Inc; 2008a Nov 1;62(11):2807–28.
Yukilevich R, True JR. Incipient sexual isolation among cosmopolitan Drosophila
melanogaster populations. Evolution. Blackwell Publishing Inc; 2008b Aug
1;62(8):2112–21.
Yukilevich R, Turner TL, Aoki F, Nuzhdin SV, True JR. Patterns and Processes
of Genome-Wide Divergence Between North American and African Drosophila
melanogaster. Genetics. Genetics Society of America; 2010 Sep
1;186(1):219–39.
Zheng X, Levine D, Shen J, Gogarten SM, Laurie C, Weir BS. A high-
performance computing toolset for relatedness and principal component
analysis of SNP data. Bioinformatics. Oxford University Press; 2012 Dec
1;28(24):3326–8.
Zhou M, Bathke A, Kim M. Combined Multiple Testing by Censored Empirical
Likelihood. Univ. Kentucky, Dept. of Statistics Tech Report. 2006
108
Appendices
Appendix A: Supplemental Tables and Figures for Postmating
Project
DF Sum Sq Mean Sq F value Pr(>F)
Block 14 1371907 97993 10.1258 < 2.2e-16
Female 22 1897556 86253 8.9126 < 2.2e-16
Male 1 94512 94512 9.7661 0.001852
Latitude 1 9105 9105 0.9408 0.332414
Longitude 1 4350 4350 0.4495 0.502803
Residuals 694 6716252 9678
TABLE A.1: Full model for egg counts with Latitude and Longitude
DF Sum Sq Mean Sq F value Pr(>F)
Block 14 1371907 97993 10.1347 < 2.2e-16
Female 22 1897556 86253 8.9204 < 2.2e-16
Male 1 94512 94512 9.7747 0.001843
Residuals 696 6729707 9669
TABLE A.2: Reduced model for egg counts without Latitude and Longitude
Res. Df. RSS DF Sum of Sq F Pr(>F)
Full 694 6716252
Reduced 696 6729707 -2 -13454 0.6951 0.4994
TABLE A.3: ANOVA table of model comparisons for egg counts
109
FIGURE A.1: Short-term (top) and long-term (bottom) remating rates of females plotted against
longitude
110
FIGURE A.0.2: Short-term (top) and long-term (bottom) remating rates of females plotted against
latitude
111
DF Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 733 522.17
Female 22 35.152 711 487.02 0.037352
Male 1 0.488 710 486.53 0.484801
Block 14 30.945 696 455.59 0.005643
Latitude 1 2.602 695 452.98 0.106761
Longitude 1 0.865 694 452.12 0.352288
TABLE A.4: Analysis of Deviance Table for Full Model of Short-term Remating Rates.
DF Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 733 522.17
Female 22 35.152 711 487.02 0.037352
Male 1 0.488 710 486.53 0.484801
Block 14 30.945 696 455.59 0.005643
TABLE A.5: Analysis of Deviance Table for Reduced Model of Short-term Remating without
Longitude or Latitude
Res. Df. Resid Df. Df. Dev. Pr(>Chi)
Full 694 452.12
Reduced 696 455.59 -2 -3.4667 0.1767
TABLE A.6: Analysis of Deviance Table for Short-term Remating Model Comparison
DF Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 733 841.58
Female 22 76.961 711 764.62 5.088e-08
Male 1 2.366 710 762.26 0.124014
Block 14 35.137 696 727.12 0.001403
Latitude 1 0.092 695 727.03 0.761588
Longitude 1 0.224 694 726.80 0.636064
TABLE A.7: Analysis of Deviance Table for Full Model of Long-term
Remating Rates.
112
DF Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 733 841.58
Female 22 76.961 711 764.62 5.088e-08
Male 1 2.366 710 762.26 0.124014
Block 14 35.137 696 727.12 0.001403
TABLE A.8: Analysis of Deviance Table for Reduced Model of Long-term Remating without
Longitude or Latitude
Res. Df. Resid Df. Df. Dev. Pr(>Chi)
Full 694 726.80
Reduced 696 727.12 -2 -0.31598 0.8539
TABLE A.9: Analysis of Deviance Table for Long-term Remating Model Comparison
DF Dev Resid Df Resid Dev Pr(>Chi)
NULL 733 841.58
Female 22 76.961 711 764.62 2.088e-08
Male 1 2.366 710 762.26 0.124014
Block 14 35.137 696 727.12 0.001403
MxF 22 31.012 674 696.11 0.095870
TABLE A.10: Analysis of Deviance Table for Reduced Model with Female x Male Interaction
Term
Res. Df. Resid Df. Df. Dev. Pr(>Chi)
with FxM 674 696.11
Reduced 696 727.12 -22 -31.012 0.09587
TABLE A.11: Analysis of Deviance Table for Reduced Long-term Remating Model Comparison
with and without Female x Male Interaction Term
113
FIGURE A.3: Hazard curves of females from line 13,34 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
114
FIGURE A.4: Hazard curves of females from line 20,17 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
115
FIGURE A.5: Hazard curves of females from line 33,11 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
116
FIGURE A.6: Hazard curves of females from line 40,23 after experiencing homotypic (solid line)
or heterotypic (dashed line) matings
117
FIGURE A.7: Hazard curves of females from line H,25 after experiencing homotypic (solid line) or
heterotypic (dashed line) matings
118
Appendix B: Supplemental Tables and Figures for Admixture
Project
π (autosome
s)
π
(X chrom)
Tajima’s D
(autosomes)
Tajima’s D (X
chrom)
CAM 0.00220948 0.00175777 1.15908195 1.27258909
CAR 0.00270037 0.00187419 0.80182112 0.90729742
SEUS 0.00260921 0.00176941 0.64704300 0.77582442
RAL 0.00256202 0.00187254 1.08400666 1.09699319
WIN 0.00239636 0.00177271 0.92246565 0.92705809
FRA 0.00288295 0.00166469 1.39156225 0.86755548
TABLE B.1: Average nucleotide diversity (π) and Tajima’s D for autosomes and X chromosomes
by population
FIGURE B.1: F
ST
between Caribbean and Cameroon populations across each chromosomal
region
119
FIGURE B.2: F
ST
between Caribbean and Raleigh populations across each chromosomal region
FIGURE B.3: F
ST
between Caribbean and southeast United States populations across each
chromosomal region
FIGURE B.4: F
ST
between Caribbean and Winters populations across each chromosomal region
120
FIGURE B.5: F
ST
between France and Cameroon populations across each chromosomal region
FIGURE B.6: F
ST
between France and Caribbean populations across each chromosomal region
FIGURE B.7: F
ST
between France and Raleigh populations across each chromosomal region
121
FIGURE B.8: F
ST
between France and southeast United States populations across each
chromosomal region
FIGURE B.9: F
ST
between France and Winters populations across each chromosomal region
FIGURE B.10: F
ST
between Raleigh and Cameroon populations across each chromosomal region
122
FIGURE B.11: F
ST
between Raleigh and Winters populations across each chromosomal region
FIGURE B.12: F
ST
between Caribbean and Cameroon populations across each chromosomal
region
FIGURE B.13: F
ST
between southeast United States and Raleigh populations across each
chromosomal region
123
FIGURE B.14: F
ST
between southeast United States and Winters populations across each
chromosomal region
FIGURE B.15: F
ST
between Winters and Cameroon populations across each chromosomal region
Population Average D′
Cameroon 0.1912575
France 0.1838487
Caribbean 0.2507630
Southeast US 0.2716825
Raleigh 0.2839296
Winters 0.2756007
TABLE B.2: Average D′ of chromosomal region 2L for each population
124
Population Average D′
Cameroon 0.1930367
France 0.1885577
Caribbean 0.2584700
Southeast US 0.2675947
Raleigh 0.2730720
Winters 0.2609750
TABLE B.3: Average D′ of chromosomal region 3L for each population
Population Average D′
Cameroon 0.1874535
France 0.1841932
Caribbean 0.2610256
Southeast US 0.2690109
Raleigh 0.2740159
Winters 0.2648767
TABLE B.4: Average D′ of chromosomal region 3L for each population
Population Average D′
Cameroon 0.1874535
France 0.1841932
Caribbean 0.2610256
Southeast US 0.2690109
Raleigh 0.2740159
Winters 0.2648767
TABLE B.5: Average D′ of chromosomal region 3R for each population
125
Population Average D′
Cameroon 0.1744096
France 0.1663958
Caribbean 0.2486993
Southeast US 0.2741414
Raleigh 0.2932253
Winters 0.2706939
TABLE B.6: Average D′ of chromosomal region 3R for each population
Population Average D′
Cameroon 0.1551474
France 0.2314284
Caribbean 0.2928683
Southeast US 0.3011555
Raleigh 0.2784124
Winters 0.2767449
TABLE B.7: Average D′ of chromosome X for each population
Abstract (if available)
Abstract
Biological systems are complicated webs of interactions at many different levels. It is through the ebb and flow of these interactions do we get processes such as molecular pathways or social behavioral systems, which can rapidly change large populations of organisms over time. Speciation is the process by which new species are formed and underlying this process are these intricate interactions between cellular, behavioral, and ecological levels. To understand the fundamentals of speciation, we study these interactions on multiple levels and at different time points of the ‘speciation continuum’ where at one end we have freely mating organisms and the other end we have reproductively blocked separate species. ❧ We present here a study of Drosophila melanogaster populations from the southeast United States and Caribbean islands. These populations represent a hybrid zone of secondary contact between cosmopolitan flies from Europe and African‐like flies from West Africa, which diverged over 10,000 years ago. With the presence of previously established clines in premating reproductive barriers (i.e. male courtship behavior, cuticular hydrocarbons, etc.), it was proposed that these flies were undergoing incipient sexual isolation, which is the start of the speciation process. We investigated putative postmating reproductive barriers of remating rates, egg hatchability, and female longevity after mating to assess the extent and influence of these barriers. We found remating rates had no effect and female longevity after mating had varied and patchy effects on several different female lines from our study area. Interestingly enough, there existed a hatchability ‘dip’ where there was an area of lower hatchability around the border of the southern US and the Caribbean islands indicating possible presence of Dobzhansky‐Muller incompatibilities. In addition to these postmating barriers, we also investigated the circadian rhythms of the southeast US and Caribbean flies and found that these populations are consistent with an existing sleep cline on the east coast of the US with increasing sleep with decreasing latitude. The differences of sleep are small, but statistically significant and could potentially impact gene flow between the D. melanogaster in the US and Caribbean islands. ❧ To investigate African and European admixture, we sequenced the genomes of our 23 isofemale fly lines from 12 locations in the southeast US and Caribbean islands. We compared these genomes to previously sequenced genomes from Raleigh, NC and Winters, CA from North America as well as to sequenced genomes of D. melanogaster from Montpellier, France and Oku, Cameroon. The genome sequences revealed up to 25% African ancestry present in the Caribbean population and decreasing percentages of African ancestry with increasing latitudes into the southeast United States. Given our results, we also propose a westward expansion model of D. melanogaster in the United States. ❧ Finally, in the spirit of studying speciation on the other end of the continuum, we establish a new study system of many closely related interacting cactophilic Drosophila species on Santa Catalina Island off the coast of southern California. We report species found on the island as well as abundance and seasonality of species compositions.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Behavioral choice assays and alcohol preference in Drosophila melanogaster
PDF
Robustness and stochasticity in Drosophila development
PDF
Automatic tracking of flies and the analysis of fly behavior
PDF
Application of machine learning methods in genomic data analysis
PDF
Long term evolution of gene duplicates in arabidopsis polyploids
PDF
Molecular and genomic studies of sex determination in Drosophila
PDF
Probing the genetic basis of gene expression variation through Bayesian analysis of allelic imbalance and transcriptome studies of oil palm interspecies hybrids
PDF
The evolution of gene regulatory networks
PDF
Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the copepod Tigriopus californicus
Asset Metadata
Creator
Kao, Joyce Y.
(author)
Core Title
Biological interactions on the behavioral, genomic, and ecological scale: investigating patterns in Drosophila melanogaster of the southeast United States and Caribbean islands
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Computational Biology and Bioinformatics
Publication Date
07/01/2014
Defense Date
06/03/2014
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
admixture,computational biology,Drosophila melanogaster,Evolution,hybrid zone,OAI-PMH Harvest,population genomics,reproductive barriers,speciation
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Nuzhdin, Sergey V. (
committee chair
), Ralph, Peter L. (
committee member
), Siegmund, Kimberly D. (
committee member
)
Creator Email
joyce.y.kao@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-428903
Unique identifier
UC11286724
Identifier
etd-KaoJoyceY-2599.pdf (filename),usctheses-c3-428903 (legacy record id)
Legacy Identifier
etd-KaoJoyceY-2599.pdf
Dmrecord
428903
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Kao, Joyce Y.
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
Tags
admixture
computational biology
Drosophila melanogaster
hybrid zone
population genomics
reproductive barriers
speciation