University of Southern California Dissertations and Theses

Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the copepod Tigriopus californicus

(USC Thesis Other)

Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the copepod Tigriopus californicus

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the
copepodTigriopuscalifornicus
by
Barret C. Phillips
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In partial fulfillment of the
requirements for the degree
DOCTOR OF PHILOSOPHY
Biology (Marine Biology and Biological Oceanography)
December 2016
Dedication
This work is dedicated to my parents, Elaina McMillan and Philip Cohen, who are always looking out
for me.
1
Acknowledgements
This dissertation would not have been possible without the help and encouragement of many people.
Many, many thanks are due to my advisor, Suzanne Edmands, who has always encouraged me to follow
my interests and to apply for the fellowships and grants that have supported me and my research at USC.
Thanks are also due to my committee members—Dennis Hedgecock, Matt Dean, and Dave Conti—who
have always taken the time to meet with me. Regina Wetzer has not only been a mentor but also allowed me
access to resources, people, and instruments that made this work possible. And thanks also to Peter Ralph,
for his enthusiasm, advice, and general good cheer.
Brad Foley gave me my first lines of R code, a good story about a border crossing, and—when I needed
it most—a home. Wai Leong was a constant companion in the lab, in the field, and in friendship during our
shared time at USC. Eric Watson has provided advice, codes, enthusiasm, encouragement, and assistance in
the field. Nicole Adams has always been there when I needed a confidant or partner in adventure. Helen
Foley and Patrick Sun have shared laughter, occasional outrage, and ideas. Shoshanna Kahne was always
ready for a lively debate. Vicky Pritchard shared her data and cultures, which were integral to this research.
The work contained in the following pages could not have happened without the assistance of numerous
undergraduate assistants, especially Shiven Chaudry and Mindy Guo. Linda Bazilian, Don Bingham, Doug
Burleson, Dawn Burke, and Adolfo dela Rosa have all helped smooth the waters in my time at USC.
This work is deeply indebted to the members of the Tigriopus californicus genome project who are
outside USC: Felipe Barreto, Ron Burton, Chris Willett, and Thiago Lima. They have shared their data and
provided analysis suggestions and two of these chapters would not have been possible without this project.
My family has been a huge source of strength. My parents, Elaina McMillan and Phil Cohen, have
put up with me working during family vacations, missing holiday gatherings for labwork, and forgetting to
call home. And through all that, their love and support for me has never wavered, including flying in from
halfway across the world just to see me give a talk. I can’t thank them enough. Thanks also to Sam, for
always (eventually) calling me back; to Marissa always having a good story to take my mind off my work;
to Katie, for being my opera buddy; and to Ebe, for humoring my nerdiness about the world we live in.
Sean Price has been a cheerleader, a helpmeet, and a distraction all in one, and I’m so glad of it. Thank
you for being there, and thank you for letting me take over your office day after night, and thank you for
talking me across the finish line.
2
So many other friends have helped along the way. Allie, my wine bar buddy at the end of a long day.
Chase, Megan, and Bridgid have been my roommates, and my friends. Eliseo opened his home to me when
I needed it. I can’t ever thank him enough for that, but I hope the chiles are a start.
And, finally, a very special thanks to my very best cat, Kichi, who always knows that what I need most
is to pet him.
3
4
Table of Contents

Dedication 1
Acknowledgements 2
List of Tables 5
List of Figures 7
Abstract 13
Introduction 15
INTRODUCTION REFERENCES 20
Chapter 1: Does the speciation clock tick more slowly in the 25
absence of heteromorphic sex chromosomes?
CHAPTER 1 REFERENCES 32
Chapter 2: Phylogeography of Tigriopus californicus revisited: 34
Highly divergent populations in Baja California, Mexico
constitute a novel species
CHAPTER 2 REFERENCES 53
Chapter 3: Hybrid breakdown, transgressive segregation, and sex- 55
specific effects of hybridization in non-recombinant
backcrosses in the copepod Tigriopus californicus
CHAPTER 3 REFERENCES 98
Chapter 4: Emergence of a novel sex chromosome in an experimental 102
hybrid population of a species with polygenic sex
determination
CHAPTER 4 REFERENCES 134
Conclusions 138
CONCLUSIONS REFERENCES 143
Bibliography 146

5
List of Tables

Table 2-1 Sampling locations for Tigriopus populations 38
Table 2-2 Loci utilized 39
Table 2-3 mt12S mean pairwise genetic distance (± standard error) 41
Table 2-4 Population genetic summary statistics for mt12S and 18S
ribosomal DNA (± standard error) 42
Table 2-5 Mean genetic distance (± standard error) between
regional clades. Above diagonal: nuclear loci; below
diagonal: mitochondrial 12S 43
Table 2-6 Mean (± standard error) mt12S interpopulation
genetic distance and geographic distance between
sampled populations within regional clades 43
Table 2-7 Number (percentage) of total, variable, and parsimony–
informative sites per locus 47
Table 3-1 Populations, isofemale lines, and crosses 58
Table 3-2 Population-diagnostic SNPs genotyped 62
Table 3-3 Diagnostic loci per chromosome per cross 63
Table 3-4 Differences between parental lines for survivorship,
sex ratio, and morphometric measures. Deviations
from equal values were calculated using independent
two-sample t-tests (p-values are shown). Bolded values
remain significant after FDR correction for multiple
tests. For significant differences, the population with
larger hatch number, higher survivorship, or more male-
biased family sex ratio is in parentheses 66
Table 3-5 Hatch number, survivorship, and sex ratio measurements
for parental lines and non-recombinant backcrosses.
Deviations from 3/4-parent expectations were calculated
by ANOVA and planned contrast tests. Statistically
significant differences are bolded 68
Table 3-6 Differences between parental lines for morphometric
measures. Deviations from equal values were calculated
using independent two-sample t-tests (p-values are
shown). Bolded values remain significant after FDR
correction. For significant differences, the population
with longer/wider cephalothoraxes, urosomes, or claspers
is in parentheses 68
Table 3-7 Morphometric measurements for parental lines and
non-recombinant backcrosses. Morphometric measure-
ments are given in pixels. Deviations from 3/4-parent
expectations were calculated by ANOVA and planned
contrast tests. Statistically significant differences are bolded 71
Table 3-8 Average (± std. dev.) genotyping error rate for each cross 75

6
Table 3-9 Average heterozygosity by cross and sex. Deviations from
the expected proportion of 0.5 were tested for with one-
sample t-tests 76
Table 3-10 Sex-specific allele frequency differences for individual
loci. Differences were calculated using Fishers Exact Test.
Only statistically significant comparisons are shown 77
Table 3-11 Median, mean, and corrected mean number of distorted
chromosomes per cross when down-sampled to SD x FHL
sample sizes, from 10,000 permutations, and actual number
of distorted chromosomes observed in the full dataset. Not
all chromosomes were genotyped in all crosses, so means
are corrected based on the proportion of chromosomes that
were genotyped in both SD x FHL (all 12) and in each cross 82
Table 3-12 Pairwise epistatic loci 83
Table 4-1 Chromosome 10 indel (scaf-76) genotype, survivorship, and sex
ratios of families sequenced individually 110
Table 4-2 Number of loci with SD allele frequency changes (Fisher’s exact
test) between 3 and 21 months of admixture in the hybrid popula-
tion, for each sex and chromosome 113
Table 4-3 Mean SD allele frequency per chromosome (± standard deviation),
population–diagnostic SNPs 114
Table 4-4 Survivorship and family sex ratios of hybrid swarm and parental
populations 115
Table 4-5 Genotype frequencies of parents from hybrid swarm family sex
ratio assays at indel scaf-76 on chromosome 10 115
Table 4-6 Read number and mapping statistics, pools 119
Table 4-7 Read number and mapping statistics, individuals 123
Table 4-8 Linear regression results for the total number of SNPs per
chromosome versus chromosome length and number of SNPs
with male-female allele frequency differences per chromosome
versus chromosome length 127

7
List of Figures

Figure 1-1 Net diversification intervals (NDIs) in reptiles and birds. Mean
NDIs (± standard error) are mapped onto a phylogenetic tree.
Heteromorphic sex chromosomes are present in birds, common
in squamates (lizards and snakes), rare in turtles, and absent in
crocodilians. Sex determination data are from (Bull 1983, Eo
and DeWoody 2010, Pokorná and Kratochvíl 2009). Phylogeny
and NDI data are from (Eo and DeWoody 2010). 30
Figure 2-1 Mitochondrial 12S Bayesian consensus tree for 25 populations
of T. californicus and one population of T. japonicus. Branches
and individuals are colored by regional clades, which are
designated by the letters A–D (T. californicus), or by species (T.
japonicus). Branch numbers indicate posterior probabilities.
MCMC was run for 2,500,000 generations. 40
Figure 2-2 Pairwise genetic divergence increases with geographic distance
in both interspecific (left) and interspecific (right) comparisons
for the 12S mitochondrial locus (top) but not for the concatenated
set of nuclear loci (bottom). Line represents best-fit linear regression
for genetic and geographic distance for inter- and intraspecific
comparisons. Comparisons are colored by type—red: interspecific;
gray: intraspecifc. 44
Figure 2-3 Bayesian consensus phylogeny for nuclear 18S ribosomal RNA
gene. Individuals are colored by mitochondrial regional clade (T.
californicus), as identified in Figure 1 or species (T. japonicus and
T. brevicornis). Branch numbers indicate posterior probabilities.
MCMC was run for 2,500,000 generations. 45
Figure 2-4 Bayesian consensus tree for concatenated set of 7 nuclear loci.
Numbers on branches indicate posterior probabilities. MCMC
was run for 1,000,000 generations. Colors correspond to regional
clades identified in Figure 1. 46
Figure 2-5 Bayesian consensus phylogeny of 1 mitochondrial and 7 nuclear
loci and map of sampling locations. Numbers on branches indicate
posterior probabilities (only values < 1 are shown). MCMC was run
for 2,500,000 generations. Colors correspond to regional clades
identified in Figure 1. Populations on the map designated in paren-
theses were sequenced at the mt12S locus but not nuclear loci. 48
Figure 2-6 Consensus maximum parsimony tree for 1 mitochondrial and 7
nuclear loci (7 best trees were found). Numbers on branches indi-
cate support (only branches present in > 50% of trees are resolved).
Branches are colored by clade as in Figure 1. 49
Figure 3-1 Phylogeny and map of populations used in the crosses. (A) Popu-
lation-consensus phylogeny from 1 mitochondrial and 7 nuclear
loci. Data from Chapter 2. (B) Sampling sites are indicated by stars.
Colors as in A. (C) Schematic of cross design. F1 females were
crossed to generation-controlled males from the paternal population
to produce non-recombinant backcross (NRBC) offspring. 61

8
Figure 3-2 Comparison of backcross hatch number (A), survivorship (B) and
family sex ratio (C) to 3/4-parent expectations. Percent difference
between observed and additive expected values are shown. For hatch
number, positive values indicate more backcross offspring per clutch,
and negative values indicate fewer backcross offspring per clutch. For
survivorship, positive values indicate higher-than-expected survivorship,
and negative values indicate lower-than-expected survivorship. For fam-
ily sex ratio, positive values indicate more males than expected (suggest-
ive of female-biased mortality), and negative values indicate more fem-
ales than expected (suggestive of male-biased mortality). Crosses are
colored according to their expected genetic composition (red: AB; blue:
BR; violet: FHL; green: SD-A; yellow: SDB). Asterisks above/below
bars indicate statistically significant differences from the 3/4-parent
expectation. *p < 0.05; **p < 0.01; ***p < 0.001. 67
Figure 3-3 Comparison of morphometric measurements to 3/4-parent expecta-
tions. Percent difference between observed and additive expected
(3/4-parent) values is shown. For cephalothorax length (A), urosome
length (C), and left (E) and right (F) clasper lengths, positive values
indicate longer-than-expected body parts in hybrids, and negative
values indicate shorter-than-expected body parts. For cephalothorax
width (B) and urosome width (D), positive values indicate wider-than-
expected body parts in hybrids, and negative values indicate narrower-
than-expected body parts. Crosses are colored according to their expec-
ted genetic composition (red: AB; blue: BR; violet: FHL; green: SD-A;
yellow: SD-B). Asterisks above/below bars indicate statistically signif-
icant differences from the 3/4-parent expectation. *p < 0.05; **p < 0.01;
***p < 0.001 70
Figure 3-4 Linear regression of absolute value of percent deviation from the 3/4-
parent for morphometric traits versus pairwise genetic divergence (dS).
Points are colored by morphometric measure. 72
Figure 3-5 Frequency of potential recombination events versus genotyping error
frequency. Points represent chromosomes with at least two diagnostic
SNPs; points are colored by cross. Dashed line indicates 1:1 ratio of
potential recombination events to genotyping errors. Points on or
below the dashed line indicate that most “recombination events”
are likely attributable to genotyping errors for that chromosome.
Points above the dashed line may be indicative of the presence of
actual recombination events or may indicate genotyping bias for one
or more loci on that chromosomes. 74
Figure 3-6 Genotype frequencies for BR x SD crosses. (A) (BR x SD) x SD
males. (B) (SD x BR) x BR males. BR homozygotes are blue, SD
homozygotes are green, and heterozygotes are brown. Asterisks
indicate p-values for χ
2
tests for segregation distortion (deviation from
a 1:1 heterozygote: homozygote ratio): *p < 0.05, **p < 0.01, ***p <
0.001; black: significant after FDR correction; gray: not significant after
FDR correction. 78

9
Figure 3-7 Genotype frequencies for AB x BR crosses. (A) (AB x BR) x BR
males. (B) (BR x AB) x AB males. AB homozygotes are red, BR
homozygotes are blue, and heterozygotes are pink. Asterisks indicate
p-values for χ
2
tests for segregation distortion (deviation from a 1:1
heterozygote: homozygote ratio): *p < 0.05, **p < 0.01, ***p <
0.001; black: significant after FDR correction; gray: not significant
after FDR correction. 79
Figure 3-8 Genotype frequencies for AB x SD crosses. (A) (AB x SD) x SD
males. (B) (SD x AB) x AB males (C) (SD x AB) x AB females.
(D) (SD x AB) x AB males and females combined. SD homozygotes
are green, AB homozygotes are red, and heterozygotes are gold.
Asterisks indicate p-values for χ
2
tests for segregation distortion
(deviation from a 1:1 heterozygote: homozygote ratio): *p < 0.05,
**p < 0.01, ***p < 0.001; black: significant after FDR correction;
gray: not significant after FDR correction. 80
Figure 3-9 Genotype frequencies for (SD-B x FHL) x FHL cross. (A) Males.
(B) Females. (C) Males and females combined. SD homozygotes
are green, FHL homozygotes are violet, and heterozygotes are gray.
Asterisks indicate p-values for χ
2
tests for segregation distortion
(deviation from a 1:1 heterozygote: homozygote ratio): *p < 0.05,
**p < 0.01, ***p < 0.001; black: significant after FDR correction;
gray: not significant after FDR correction. 81
Figure 3-10 Number of chromosomes exhibiting segregation distortion per
cross in 10,000 replicate re-samplings of genotyped individuals at
sample sizes equal to those genotyped for (SD x FHL) x FHL after
FDR correction. Histograms are colored by cross. Dashed lines:
number of distortions detected in (SD x FHL) x FHL males (dark
purple), females (light blue), and males and females combined (light
blue) after FDR correction. Asterisks: the number of distorted chrom-
osomes detected for each cross with the full dataset. 82
Figure 3-11 Morphometric QTL for the BR x SD crosses under a single-locus
model. (A) Cephalothorax length; (B) cephalothorax width; (C)
urosome length; (D) urosome width; (E) left and (F) right clasper
length. Orange: (BR x SD) x SD; dark red: (SD x BR) x BR. Dashed
lines indicate p = 0.05 significance threshold as determined by
permutation. 85
Figure 3-12 Morphometric QTL for the AB x BR crosses under a single-locus
model. (A) Cephalothorax length; (B) cephalothorax width; (C)
urosome length; (D) urosome width; (E) left and (F) right clasper
length. Red: (AB x BR) x BR; green: (BR x AB) x AB. Dashed lines
indicate p = 0.05 significance threshold as determined by permutation. 86
Figure 3-13 Morphometric QTL for the AB x SD crosses under a single-locus
model. (A) Cephalothorax length; (B) cephalothorax width; (C)
urosome length; (D) urosome width; (E) left and (F) right clasper
length. Dark green: (AB x SD) x SD; light blue: (SD x AB) x AB.
Dashed lines indicate p = 0.05 significance threshold as determined
by permutation. 87

10
Figure 3-14 Morphometric QTL for the SD x FHL cross under a single-locus
model. (A) Cephalothorax length; (B) cephalothorax width; (C)
urosome length; (D) urosome width; (E) left and (F) right clasper
length. Dashed line indicates p = 0.05 significance threshold as
determined by permutation. 89
Figure 3-15 QTL mapping of sex determination in the (SD x AB) x AB cross.
(A) Single-locus scan for sex QTL. Dashed line indicates p = 0.05
significance threshold as determined by permutation. (B-D) Effect
plots of two-locus genotypes: (B) chromosomes 2 x 10; (C) chrom-
osomes 4 x 10; (D) chromosomes 4 x 2. 90
Figure 3-16 QTL mapping of sex determination in the (SD x FHL) x FHL cross.
(A) Single-locus scan for sex QTL. Dashed line indicates p = 0.05
significance threshold as determined by permutation. (B) Effect plot
of two-locus genotypes for chromosomes 6 and 10. 91
Figure 4-1 Mean SD allele frequency per chromosome per month in males
(orange triangles) and females (purple circles) in the hybrid pop-
ulation over its first 21 months of admixture. Error bars indicate
standard error. Asterisks identify chromosomes with >1 SNP
where allele frequencies differ between males and females (Fishers
exact test, 10% FDR correction). Data from (Pritchard and Edmands
2013). 112
Figure 4-2 Population–diagnostic genotyping of the experimental hybrid pop-
ulation after 4 years of admixture revealed sex-specific heterozy-
gosity across a single chromosome, which is the only region of the
genome that is strongly associated with sex. (A) SD allele frequency
at 187 parental-population–diagnostic SNPs (females: purple; males:
orange). Many hybrid population females have retained SD ancestry
across the entire mapped length of chromosome 10 despite genetic
swamping of SC alleles across the remainder of the genome. Dashed
lines indicate chromosome boundaries. (B). Association tests for sex
find a strong effect of chromosome 10 and only of chromosome 10
after 10,000 permutations. Each SNP is colored by chromosome.
Black line indicates p = 0.05. 116
Figure 4-3 Principal components and cluster analysis of population–diagnostic
SNPs. (A.) Principal components plot for the top two eigenvectors.
PC 2 separates males from some females. Black: females; red: males.
(B.) Correlation coefficient for each SNP for the top two eigenvectors.
Correlation coefficients for PC 2 are largest for chromosome 10. SNPs
are colored by chromosome. (C.) Identity-by-state cluster analysis
groups females with SD ancestry on chromosome 10 separately from
all but one male. Black: female; red: male. 117

11
Figure 4-4 Quantile-quantile family sex ratio plot. Family sex ratios do not fit
a binomial distribution—as would be expected under chromosomal
sex determination–in either the hybrid swarm or the parental popula-
tions. Plot depicts actual versus expected family sex ratios under a
binomial model with all theoretical family sizes equal to the average
clutch size for that population. Green: parental populations (SC and
SD combined); purple: hybrid population. Dashed gray line indicates
the expectation when family sex ratios are binomially distributed (as
with chromosomal sex determination). 118
Figure 4-5 Reads map disproportionately to chromosome 3 in hybrids and
SC and to chromosome 12 in SD, relative to assembled chromo-
some length. Each bar represents the percent difference between
the proportion of total reads unambiguously mapped to a chromo-
some per pool set (males and females summed across sequencing
lanes) and the proportion of the total assembled, mapped genome
assigned to each chromosome. Chi-square p < 0:001 for all pool
sets. A, B, C, E. Experimental hybrid population. D. Parental pop-
ulation SC. F. Parental population SD. A, C, E. Pools composed
of 50 siblings from families where mothers had with 0-1 (A), 1
(C), or 2 (E) SD alleles at the scaf-76 indel on chromosome 10.
B. Unrelated population sample of hybrid population. Bars are
colored by chromosome. 122
Figure 4-6 Degenerated sex chromosomes are absent in both the parental
populations and the hybrid population. Log2 normalized read
coverage per scaffold per individual is plotted for all scaffolds
mapped to chromosomes. Lines represent GAM-smoothed
averages across all samples. Purple: females; orange: males. 125
Figure 4-7 Male-female allele frequency differences for each sequenced
pool set. A, B. Pools composed of 50 siblings from families
from parental population A. SC and B. SD. C, D, E. Pools
composed of 50 siblings from families where mothers had 0-1
(A), 1 (C), or 1-2 (E) SD alleles at the scaf-76 indel on chromo-
some 10. F. Random sample of hybrid population. Black line
indicates FDR cutoff. 126
Figure 4-8 Boxplots of p-values (Cochran-Mantel-Haenszel test) for male-
female allele frequency differences on each mapped scaffold for
each sequenced pool set. A, B. Pools composed of 50 siblings
from families from parental population A. SC and B. SD. C, D, E.
Pools composed of 50 siblings from families where mothers had
0-1 (A), 1 (C), or 1-2 (E) SD alleles at the scaf-76 indel on
chromosome 10. F. Random sample of hybrid population. 128

12
Figure 4-9 More SNPs with statistically significant allele frequency differ-
ences between males and females are present on chromosome 10
in the experimental hybrid population but not in the parental popu-
lations, and chromosome length is correlated with both the number
of detected SNPs and the number of SNPs with different allele freq-
uencies in males and females in the hybrid population but not in the
parental populations. A. Proportion of all detected, mapped SNPs per
chromosome in each pool set. B. Proportion of all mapped SNPs with
different allele frequencies in males and females per chromosome per
pool set (CMH test, 10% FDR correction). Each color represents a
chromosome. C. Linear regressions of total number of detected SNPs
per chromosome versus chromosome length. D. Linear regressions of
the number of SNPs with allele frequencies that differ between males
and females versus length of assembled chromosomes. E., Linear
regressions of the number of male-female divergent SNPs versus the
total number of SNPs per chromosome. C, D, E. Points are colored by
chromosome as in (A, B). Purple line/text: hybrid population; green
line/text: parental populations. 129
Figure 4-10 Principal components analysis of whole–genome re-sequencing for
individual animals. A. PCA of individuals from all three sequenced
populations. Triangles: Parental population SC; Crosses: Parental
population SD; Circles: Hybrid population. Black: female; red: male.
B. PCA of variants for sequenced individuals for the hybrid popula-
tion. Squares: Individuals from Hybrid 0/1 families; circles: individ-
uals from families Hybrid 1 families; triangles: individuals from
Hybrid 1/2 families. C. PCA of variants for sequenced individuals
for the SC population. D. PCA of variants for sequenced individuals
for the SD population. Colors as in A. 130

Abstract
The processes and patterns underlying speciation are a major focus of evolutionary biology. These pro-
cesses often result in the evolution of reproductive isolation, and sex chromosomes often play an important
role in postzygotic reproductive isolation. However, sex chromosomes—and, more generally, sex determi-
nation mechanisms—are themselves highly evolutionarily labile, and many lineages lack sex chromosomes
but still manage to speciate. However, the vast majority of work on the evolution of postzygotic reproductive
isolation has been conducted in taxa that do possess sex chromosomes, leaving many open questions regard-
ing how postzygotic isolation may evolve in taxa without sex chromosomes. This dissertation examines how
the speciation process may differ in taxa that do not have sex chromosomes.
Chapter 1 asks whether the presence of sex chromosomes in a clade accelerates speciation. Sex chro-
mosomes disproportionately accumulate postzygotic incompatibilities (large X–effect) and contribute to
sex–specific hybrid inviability and infertility (Haldane’s rule). Therefore, in the absence of sex chromo-
somes, postzygotic reproductive isolation may take longer to accumulate to the same level, which could
ultimately result in slower speciation rates. We addressed this question in reptiles, where we found that
speciation intervals are shorter in snakes and lizards (where sex chromosomes are common) than in turtles
and crocodiles (where sex chromosomes are rare or absent).
Chapters 2–4 examine phylogeography, hybrid incompatibility, and sex determination in the intertidal
copepodTigriopuscalifornicus, a species with polygenic (genetic but non-chromosomal) sex determination.
Chapter 2 presents a phylogeographic study of T. californicus across the species range. T. californicus
populations have been documented from southern Alaska, USA to central Baja California, Mexico, and
previous phylogeographic studies have employed only a single mitochondrial locus (mtCOI). We used a
different mitochondrial locus (mt12S), as well as 7 nuclear loci (18S, ITS1, 5.8S, and 4 transcriptome–
based loci) to resolve relationships between major clades and to clarify the taxonomic status of populations
at the southern edge of the range, which are genetically and reproductively isolated from otherT.californicus
populations. We found definitive evidence that these southern populations constitute a novel species that is
geographically continuous withT.californicus, identify the presence of major regional clades, and document
a general isolation–by–distance pattern within regions.
In Chapter 3, non-recombinant backcrosses between multiple pairs of T. californicus populations were
used to map postzygotic reproductive isolation, morphometric features, and sex. We identified segregation
13
distortion in some, but not all crosses and detected few epistatic incompatibilities. We mapped associations
with sex to several chromosomes, consistent with polygenic sex determination, and we found evidence of
transgressive segregation for morphometric traits in some crosses. In these crosses, we found that males
consistently fare worse than females, despite not being heterogametic (the traditional explanations for sex-
specific deleterious effects invoke heterogamety). Thus, explanations for sex-specific deleterious effects of
hybridization may require the generation of alternative hypotheses in taxa without sex chromosomes.
Finally, Chapter 4 investigates polygenic sex determination and the emergence of a novel sex chromo-
some in a hybrid population of T. californicus. Polygenic sex determination, where sex determination is
genetic but non-chromosomal, is predicted to be unstable and prone to takeover by strong sex-determiners,
which can then become sex chromosomes. We found that alleles from one parental population (SD, from
San Diego, California, USA) had introgressed sex–specifically on one chromosome (chromosome 10), de-
spite swamping by the other population’s alleles (SC, from Santa Cruz, California, USA) across the rest of
the genome. SD alleles on chromosome 10 appear to have a strong feminizing effect, and—in the absence of
meiotic recombination in females—to have been transmitted almost exclusively to females. Whole-genome
re-sequencing of the parental progenitor populations did not reveal allele frequency differences between
males and females to be concentrated anywhere in the genome, whereas in the hybrid population these dif-
ferences were elevated on chromosome 10. The total evidence suggests that chromosome 10 is acting as
a novel sex chromosome within the hybrid population but not within the parental populations, making it
perhaps the first sex chromosome to emerge from a polygenic sex determination system in a laboratory.
The speciation process is currently poorly understood outside of a few model taxa, where sex chromo-
somes are involved in most aspects of postzygotic reproductive isolation. Increasing the taxonomic breadth
of this research is an imperative in the Anthropocene, where human activities are directly affecting the
speciation process through anthropogenically–mediated hybridization and extinction. Speciation in the ab-
sence of sex chromosomes may proceed fundamentally differently, as depicted in Chapters 1, 2, and 3.
Alternatively, as seen in Chapter 4, hybridization could in some cases lead to the emergence of a novel sex
chromosome, which could then modify the speciation process.
14
Introduction
Speciation is at the heart of biodiversity. Estimates of the number of species on the planet range from a
few million to one trillion, but all living organisms ultimately share a common evolutionary history (Tittensor
et al. 2010, Mora et al. 2011, Locey and Lennon 2016). How organisms have speciated and are speciating,
therefore, is a fundamental question of evolutionary biology that will require extensive studies in a variety
of taxa, including representatives of the numerous highly divergent lineages present on the planet today.
Much notorious debate in the biological community has revolved around the formal definition of the
seemingly simple term species, with at least 26 species concepts having been proposed, no single one of
which is expansive enough to encompass the full diversity of extant and extinct organisms (Coyne and
Orr 2004, Hausdorf 2011). However, the biological species concept proposed by Mayr (1942) provides a
convenient basis on which to base studies of speciation in sexually reproducing organisms. Thus, a species
is a group of organisms capable of producing viable, fertile offspring with each other but not with non–group
members (Mayr 1942).
Species barriers are, consequently, maintained by factors preventing gene flow and may be extrinsic (e.g.,
a geographic feature) or intrinsic (e.g., hybrids are sterile). Intrinsic factors that do not prevent mating or
fertilization but are manifested later as partial or complete sterility or inviability of hybrid offspring (whether
occurring in the first or some later generation) are collectively termed postzygotic isolation (Coyne and Orr
2004). The genetic mechanisms responsible for postzygotic isolation have been the subject of intense study
that is concentrated in a few taxa, where sex chromosomes appear to play a large role (Presgraves 2008,
2010).
Broadscale patterns relating to postzygotic isolation have been described as the three “rules” of speci-
ation (Coyne and Orr 1989b, Turelli and Moyle 2007). Of these rules, two are entirely attributable, and
the third is partially attributable, to the peculiar properties of sex chromosomes: Haldane’s rule is the ob-
servation that the heterogametic sex often fares more poorly during hybridization (e.g., XY males or ZW
females are more often sterile and/or inviable than XX females or ZZ males); the large-X effect notes that
genetic incompatibilities map disproportionately to the X (or Z) chromosome; and Darwin’s corollary is
the observation that fitness often varies between reciprocal crosses, which may be explained by divergences
on sex chromosomes or by cytonuclear incompatibilities (Haldane 1922, Coyne and Orr 1989b, Masly and
Presgraves 2007, Turelli and Moyle 2007, Ellegren 2009).
15
Additionally, sex chromosome turnovers may contribute directly to speciation (Kitano and Peichel
2012). The contribution of a neo-sex chromosome to speciation has been directly demonstrated in stickle-
backs, where loci controlling mating behavior reside on a neo-X chromosome, while hybrid sterility loci are
located on the ancestral X (Kitano et al. 2009), and the fixation of different inversions on the Y chromosome
may underlie the formation of species complexes in blackflies, where numerous paracentric inversions ap-
pear to create a gradation from polymorphic cytotypes to morphologically indistinguishable sibling species
separated by fixed inversions to morphologically distinguishable species (Rothfels 1989, Shields 2013). Fur-
thermore, transitions between sex chromosomes have recently been proposed as potential causative agents
for the splitting of the major mammal lineages (Graves in press).
However, though sex chromosomes have evolved in many lineages, they are far from a universal feature
in the genomes of gonochoristic organisms (Bull 1983, The Tree of Sex Consortium 2014). Alternatives
to sex chromosomes include environmental sex determination, haplodiploidy, and polygenic sex determina-
tion. Under polygenic sex determination, sex determination is genetic but non-chromosomal; rather, sex is
determined by multiple, often unlinked, loci. Polygenic sex determination was once thought to be so unsta-
ble that it might be nearly impossible to observe (Rice (1986) but see also van Doorn and Kirkpatrick (2007,
2010)) but has been documented in many organisms in recent years, including Lake Tanzania cichlids, lab
strains of zebrafish, and European frogs (Roberts et al. 2009, Ser et al. 2010, Liew et al. 2012, Rodrigues
et al. 2013, Anderson et al. 2012, Wilson et al. 2014, Rodrigues et al. 2015, 2016).
This dissertation examines how the speciation process may differ when sex chromosomes are absent.
The first chapter tackles a broad question about speciation rates in taxa with and without sex chromosomes,
while the other three address genetic differentiation, hybrid incompatibility, and the evolution of sex deter-
mination mechanisms in a marine copepod.
Chapter 1 asks the question, “Does the speciation clock tick more slowly in the absence of heteromorphic
sex chromosomes?”—that is, if sex chromosomes are primarily responsible for manifestations of postzy-
gotic reproductive isolation and/or reproductive isolation accumulates more quickly on sex chromosomes
than on autosomes, does it then follow that reproductive isolation accumulates more slowly in taxa that do
not have highly differentiated sex chromosomes? Addressing this question requires good estimates of speci-
ation rates in taxa with variable sex determination mechanisms. We therefore examined reptiles, where sex
determination mechanisms are highly variable and range from conserved heteromorphic sex chromosomes
in birds and snakes to conserved environmental sex determination in turtles and crocodilians to family- and
16
species-specific mechanisms in lizards.
The remaining chapters focus on the emerging marine model organismTigriopuscalifornicus, a harpacti-
coid copepod inhabiting splash pools above the mean high tide line from Baja California to Alaska (Dethier
1980, Ganz and Burton 1995, Powlik 1996, Edmands 2001). These pools are rarely tidally inundated, and,
as a result, T. californicus is exposed to extreme fluctuations in pool volume, temperature, and salinity on
a regular basis (Vittor 1971). Despite its broad range, T. californicus apparently experiences little gene
flow between rock outcrops, even on the range of a few kilometers, such that a single outcrop can thus be
considered to encompass an individual population (Burton and Feldman 1981, Burton and Swisher 1984).
However, though genetic divergence between populations has been extensively documented in T. cali-
fornicus, range–wide phylogeographic studies to date have used sequences only from the mitochondrial COI
gene (Edmands 2001, Peterson et al. 2013). These studies found low divergence between northern popu-
lations, possibly as a result of recent post-glacial range expansion, and could not resolve the relationships
between populations. Conversely, populations at the southern edge of the range had previously been pro-
posed to be a semi-species or sibling species due to their reproductive isolation and high genetic divergence
from otherT.califonrnicus populations (Ganz and Burton 1995, Peterson et al. 2013). Chapter 2 re-examines
the phylogenetic structure ofT.californicus populations from the San Juan Islands in Washington to central
Baja California using both nuclear and mitochondrial loci.
Sex determination in T. californicus is polygenic. Karyotypic studies have demonstrated that T. cali-
fornicus lacks heteromorphic sex chromosomes (Ar-Rushdi 1963, Lazzaretto and Libertini 1985), and all
of the predicted hallmarks of polygenic sex determination have been observed: family sex ratios exhibit
extrabinomial variation (V oordouw and Anholt 2002a,b, V oordouw et al. 2005b,a, 2008), both maternal and
paternal effects on family sex ratio are heritable (V oordouw and Anholt 2002b, V oordouw et al. 2005a,
2008), and family sex ratios respond to selection for sex bias (Ar-Rushdi 1958, Alexander et al. 2014).
Additionally, QTL mapping of sex has identified 6 loci on 5 of the 12 chromosomes in the T. californicus
genome (Alexander et al. 2015). Several environmental factors, including temperature (V oordouw and An-
holt 2002a), UV–B radiation (Chalker-Scott 1995), and hydrostatic pressure (Vacquier 1962, Vacquier and
Belser 1965), are known to exert a small but (at least for temperature) heritable influence on primary sex
ratios, but the magnitude of their effect is much smaller than the between–family variance observed and
cannot be considered the primary sex determination mechanism. Finally, attempted PCR amplification of
Wolbachia and microsporidian rDNA sequences, visual microscopy, and antibiotic treatment have all failed
17
to find evidence of cytoplasmic sex ratio distorters (V oordouw et al. 2008, S. Edmands et al. unpub. data).
However, meiosis in females is achiasmatic (no recombination), a feature frequently (but not always) associ-
ated with heterogamety (known as the Haldane–Huxley effect; Ar-Rushdi 1963, Burton et al. 1981, Harrison
and Burton 2006; but see Lenormand and Dutheil 2005), which could pre–disposeT.californicus to female
heterogamety if a strong sex–determiner emerged within a population (Rice 1986, Wright et al. 2016).
In taxa with sex chromosomes, hybrid sterility in the heterogametic sex often arises before hybrid steril-
ity in the homogametic sex and before hybrid inviability in either sex (Coyne and Orr 1989a, Wu 1992,
Coyne and Orr 1997, Presgraves and Orr 1998, Sasa et al. 1998, Naisbit et al. 2002, Presgraves 2002, Price
and Bouvier 2002, Tubaro and Lijtmaer 2002, Lijtmaer et al. 2003). However, in T. californicus, hybrid
sterility in either sex is rare and does not accumulate faster than hybrid inviability (Willett 2008), which
is often observed even in low-divergence crosses (Palmer and Edmands 2000), but postzygotic isolation
has been documented in many crosses (Burton 1990a,b, Burton et al. 1999, Edmands 1999, Willett and
Burton 2001, Ellison and Burton 2006, Harrison and Edmands 2006, Willett and Berkowitz 2007, Ellison
and Burton 2008, Edmands et al. 2009, Ellison and Burton 2010, Pritchard et al. 2011, Foley et al. 2013,
Willett et al. in press). Chapter 3 investigates reproductive compatibility in a series of crosses between T.
californicus populations of increasing genetic and geographic divergence. Survivorship, family sex ratios,
and male morphology are documented in non-recombinant backcross hybrids; hybrids were also genotyped
at population–diagnostic loci to search for loci exhibiting segregation distortion and/or involved in epistatic
incompatibilities and to map QTL for morphometric traits and for sex.
Polygenic sex determination has long been predicted to be unstable because it is vulnerable to the inva-
sion of strong sex–determiner (Rice 1986). Additionally, the evolution of a sex chromosome can be reduced
to a single–step process if sex–limited achiasmy already exists because recombination is already suppressed
across the entire genome (Wright et al. 2016). In species with polygenic sex determination, allopatric
populations on different evolutionary trajectories may have divergent sex–determining loci. Therefore, hy-
bridization may disrupt a carefully calibrated sex determining system, allowing for the invasion of a strong
sex determiner. Experimental hybrid swarms offer an approach to test these hypotheses. Previous work
tracked the genetic and phenotypic trajectories of replicate experimental hybrid populations created by com-
bining equal numbers of gravid females from two highly divergentT. californicus populations (Santa Cruz,
CA and San Diego, CA) for 21 months (Pritchard and Edmands 2012, Pritchard et al. 2013). Genotyping of
population–diagnostic markers in a limited number of individuals at 3–month intervals revealed that of the
18
three replicates, two had nearly gone to fixation for a single (and opposite) population’s genotype across the
majority of the genome, which matched the mitochondrial genotype that became fixed in those populations.
Chapter 4 continues this work by re-examining the SC–biased hybrid swarm population (the only extant
one), where a single chromosome had previously been shown to have a large effect on sex (Pritchard et al.
2013). More extensive population–diagnostic genotyping was conducted after 4 years of admixture, and
family sex ratio assays and full–genome re-sequencing were performed for the hybrid population (after 7
years of admixture) and its parental progenitors (from fresh collections).
Many fundamental questions remain about how and under what conditions speciation occurs, and they
are of increasing urgency in a world where anthropogenically–mediated hybridization and extinctions are
occurring at increased rates (Allendorf et al. 2001, Wolff et al. 2014, Ceballos et al. 2015). Answering
these questions will require an understanding of these processes that is based on information from diverse
taxa, which necessitates a broadening of the focus of traditional speciation genetics away from conventional
model organisms. It will especially require an understanding of how, and why, speciation processes may
differ in organisms who do not have sex chromosomes, as sex chromosomes are such an important fea-
ture of traditional characterizations of speciation genetics but are far from a ubiquitous sex determination
mechanism.
19
References
Alexander, H., J. Richardson, and B. Anholt. 2014. Multigenerational response to artificial selection for
biased clutch sex ratios inTigriopuscalifornicus populations. Journal of Evolutionary Biology27:1921–
1929.
Alexander, H., J. Richardson, S. Edmands, and B. Anholt. 2015. Sex without sex chromosomes: genetic
architecture of multiple loci independently segregating to determine sex ratios in the copepod Tigriopus
californicus. Journal of Evolutionary Biology28:2196–2207.
Allendorf, F., R. Leary, P. Spruell, and J. Wenburg. 2001. The problems with hybrids: setting conservation
guidelines. Trends in Ecology and Evolution16:613–622.
Anderson, J. L., A. R. Mari, I. Braasch, A. A. P. Hohenlohe, P. Batzel, and J. H. Postlethwait. 2012.
Multiple sex–associated regions and a putative sex chromosome in zebrafish revealed by RAD mapping
and population genomics. PLoS One7:e40701.
Ar-Rushdi, A. H. 1958. The polygenic basis of sex-ratio in Tigriopus. Paper presented at the Xth Interna-
tional Congress of Genetics, McGill University, Montreal, Canada.
Ar-Rushdi, A. H. 1963. The cytology of achiasmatic meiosis in the female Tigriopus (Copepoda). Chro-
mosoma13:1754–1763.
Bull, J. 1983. Evolution of Sex Determining Mechanisms. Benjamin/Cummings, Menlo Park, CA.
Burton, R. 1990a. Hybrid breakdown in developmental time in the copepodTigriopuscalifornicus. Evolu-
tion44:1814–1822.
Burton, R. 1990b. Hybrid breakdown in physiological response: A mechanistic approach. Evolution
44:1806–1813.
Burton, R., M. Feldman, and S. Swisher. 1981. Linkage relationships among 5 enzyme-coding gene loci in
the copepodTigriopuscalifornicus: A genetic confirmation of achiasmatic meiosis. Biochemical Genetics
19:1237–1245.
Burton, R., P. Rawson, and S. Edmands. 1999. Genetic architecture of physiological phenotypes: Empirical
evidence for coadapted gene complexes. American Zoologist39:451–462.
Burton, R. S., and M. W. Feldman. 1981. Population genetics of Tigriopus californicus: II. Differentiation
among neighboring populations. Evolution35:1192–1205.
Burton, R. S., and S. G. Swisher. 1984. Population structure of the intertidal copepodTigriopuscalifornicus
as revealed by field manipulation of allele frequencies. Oecologia65:108–111.
Ceballos, G., P. R. Ehrlich, A. D. Barnosky, A. Garcia, R. M. Pringle, and T. M. Palmer. 2015. Accel-
erated modern human–induced species losses: Entering the sixth mass extinction. Science Advances
1:e1400253.
Chalker-Scott, L. 1995. Survival and sex-ratios of the intertidal copepod,Tigriopuscalifornicus, following
ultraviolet-B (290-320 nm) radiation exposure. Marine Biology123:799–804.
Coyne, J., and H. Orr. 1997. “Patterns of speciation inDrosophila” revisited. Evolution51:295–303.
Coyne, J. A., and H. A. Orr. 1989a. Patterns of speciation inDrosophila. Evolution43:362–381.
20
Coyne, J. A., and H. A. Orr, 1989b. Speciation and Its Consequences, Chapter: Two rules of speciation,
pages 180–207 . Sinauer, Sunderland, MA.
Coyne, J. A., and H. A. Orr. 2004. Speciation. Sinauer, Sunderland, MA.
Dethier, M. N. 1980. Tidepools as refuges: predation and the limits of the harpacticoid copepod Tigriopus
californicus (Baker). Journal of Experimental Marine Biology and Ecology42:99–111.
Edmands, S. 1999. Heterosis and outbreeding depression in interpopulation crosses spanning a wide range
of divergence. Evolution53:1757–1768.
Edmands, S. 2001. Phylogeography of the intertidal copepod Tigriopus californicus reveals substantially
reduced population differentiation at northern latitudes. Molecular Ecology10:1743–1750.
Edmands, S., S. Northrup, and A. Hwang. 2009. Maladapted gene complexes within populations of the
intertidal copepodTigriopuscalifornicus? Evolution63:2184–2192.
Ellegren, H. 2009. Genomic evidence for a large-Z effect. Proceedings of the Royal Society B: Biological
Sciences276:361.
Ellison, C. K., and R. S. Burton. 2006. Disruption of mitochondrial function in interpopulation hybrids of
Tigriopuscalifornicus. Evolution60:1382–1391.
Ellison, C. K., and R. S. Burton. 2008. Interpopulation hybrid breakdown maps to the mitochondrial
genome. Evolution62:631–638.
Ellison, C. K., and R. S. Burton. 2010. Cytonuclear conflict in interpopulation hybrids: The role of RNA
polymerase in mtDNA transcription and replication. Journal of Evolutionary Biology23:528–538.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, and S. Edmands. 2013. Postzygotic isolation involves
strong mitochondrial and sex–specific effects in Tigriopus californicus, a species lacking heteromorphic
sex chromosomes. Heredity111:391–401.
Ganz, H., and R. Burton. 1995. Genetic differentiation and reproductive incompatibility among Baja
California populations of the copepodTigriopuscalifornicus. Marine Biology123:821–827.
Graves, J. A. M. in press. Did sex chromosome turnover promote divergence of the major mammal groups?
BioEssays38.
Haldane, J. 1922. Sex ratio and unisexual sterility in hybrid animals. Journal of Genetics12:101–109.
Harrison, J., and R. Burton. 2006. Tracing hybrid incompatibilities to single amino acid substitutions.
Molecular Biology and Evolution23:559–564.
Harrison, J. S., and S. Edmands. 2006. Chromosomal basis of viability differences inTigriopuscalifornicus
interpopulation hybrids. Journal of Evolutionary Biology19:2040–2051.
Hausdorf, B. 2011. Progress toward a general species concept. Evolution65:923–931.
Kitano, J., and C. L. Peichel. 2012. Turnover of sex chromosomes and speciation in fishes. Environmental
Biology of Fishes94:549–558.
Kitano, J., J. A. Ross, S. Mori, M. Kume, F. C. Jones, Y . F. Chan, D. M. Absher, J. Grimwood, J. Schmutz,
R. M. Myers, D. M. Kingsley, and C. L. Peichel. 2009. A role for a neo-sex chromosome in stickleback
speciation. Nature461:1079–1083.
21
Lazzaretto, I., and A. Libertini. 1985. Karyological comparisons among three species of the genusTigriopus
(Copepods, Harpacticoida). Atti. Acc. Lincei Rend. fis.126:377–381.
Lenormand, T., and J. Dutheil. 2005. Recombination difference between sexes: a role for haploid selection.
PLoS Biology3:e63.
Liew, W., R. Bartfai, Z. Lim, R. Sreenivasan, K. Siegfried, and L. Orban. 2012. Polygenic sex determination
system in zebrafish. PLoS One7:e34397.
Lijtmaer, D., B. Mahler, and P. Tubaro. 2003. Hybridization and postzygotic isolation patterns in pigeons
and doves. Evolution57:1411–1418.
Locey, K. J., and J. T. Lennon. 2016. Scaling laws predict global microbial diversity. Proceedings of the
National Academy of Sciences113:5970–5975.
Masly, J. P., and D. C. Presgraves. 2007. High-resolution genome-wide dissection of the two rules of
speciation inDrosophila. PLoS Biology5:1890–1898.
Mayr, E. 1942. Systematics and the Origin of Species. Columbia University Press, New York.
Mora, C., D. P. Tittensor, S. Adl, A. G. B. Simpson, and B. Worm. 2011. How many species are there on
Earth and in the ocean? PLoS Biology9:e1001127.
Naisbit, R. E., C. D. Jiggins, M. Linares, C. Salazar, and J. Mallet. 2002. Hybrid sterility, Haldane’s rule
and speciation inHeliconiuscydno andH.melpomene. Genetics161:1517–1526.
Palmer, C., and S. Edmands. 2000. Mate choice in the face of both inbreeding and outbreeding depression
in the intertidal copepodTigriopuscalifornicus. Marine Biology136:693–698.
Peterson, D., K. Kubow, M. Connolly, L. Kaplan, M. Wetkowski, W. Leong, B. Phillips, and S. Edmands.
2013. Reproductive and phylogenetic divergence of tidepool copepod populations across a narrow geo-
graphical boundary in Baja California. Journal of Biogeography40:1664–1675.
Powlik, J., 1996. Ecology of Tigriopus californicus (Copepoda, Harpacticoida) in Barkley Sound, British
Columbia. Ph.D. thesis, University of British Columbia.
Presgraves, D. C. 2002. Patterns of postzygotic isolation in Lepidoptera. Evolution56:1168–83.
Presgraves, D. C. 2008. Sex chromosomes and speciation inDrosophila. Trends in Genetics24:336–343.
Presgraves, D. C. 2010. The molecular evolutionary basis of species formation. Nature Reviews Genetics
11:175–180.
Presgraves, D. C., and H. A. Orr. 1998. Haldane’s rule in taxa lacking a hemizygous X. Science282:952–
954.
Price, T., and M. Bouvier. 2002. The evolution of F
1
postzygotic incompatibilities in birds. Evolution
56:2083–2089.
Pritchard, V ., L. Dimond, J. Harrison, C. Velasquez, J. Zieba, R. Burton, and S. Edmands. 2011. Interpopu-
lation hybridization results in widespread viability selection across the genome inTigriopuscalifornicus.
BMC Genetics12:54.
Pritchard, V ., and S. Edmands. 2012. The genomic trajectory of hybrid swarms: Outcomes of repeated
crosses between populations ofTigriopuscalifornicus. Evolution67:774–791.
22
Pritchard, V ., V . Knutson, M. Lee, J. Zeiba, and S. Edmands. 2013. Fitness and morphological outcomes of
many generations of hybridization in the copepodTigriopuscalifornicus. Journal of Evolutionary Biology
26:416–433.
Rice, W. 1986. On the instability of polygenic sex determination: The effect of sex-specific selection.
Evolution40:633–639.
Roberts, R. B., J. R. Ser, and T. D. Kocher. 2009. Sexual conflict resolved by invasion of a novel sex
determiner in Lake Malawi cichlid fishes. Science326:998–1001.
Rodrigues, N., C. Betto-Colliard, H. Jourdan-Pineau, and N. Perrin. 2013. Within-population polymorphism
of sex-determination systems in the common frog (Rana temporaria). Journal of Evolutionary Biology
26:1569–1577.
Rodrigues, N., Y . Vuille, A. Brelsford, J. Merila, and N. Perrin. 2016. The genetic contribution to sex deter-
mination and number of sex chromosomes vary among populations of common frogs (Ranatemporaria).
Heredity117:25–32.
Rodrigues, N., Y . Vuille, and N. Perrin. 2015. Sex-chromosome differentiation and ‘sex races’ in the com-
mon frog (Ranatemporaria). Proceedings of the Royal Society B: Biological Sciences282:20142726.
Rothfels, K. 1989. Speciation in black flies. Genome32:500–509.
Sasa, M., P. Chippindale, and N. A. Johnson. 1998. Patterns of postzygotic isolation in frogs. Evolution
52:1811–1820.
Ser, J., R. Roberts, and T. Kocher. 2010. Multiple interacting loci control sex determination in Lake Malawi
cichlid fish. Evolution64:486–501.
Shields, G. F. 2013. The significance of sex–linked chromosomal inversions in the speciation process of
the Simulium arcticum complex of black flies (Diptera: Simuliidae). Monographs of the Western North
American Naturalist6:64–86.
The Tree of Sex Consortium. 2014. Sex determination: Why so many ways of doing it? PLoS Biology
12:e1001899.
Tittensor, D. P., C. Mora, W. Jetz, H. K. Lotze, D. Ricard, E. V . Berghe, and B. Worm. 2010. Global patterns
and predictors of marine biodiversity across taxa. Nature466:1098–1101.
Tubaro, P., and D. Lijtmaer. 2002. Hybridization patterns and the evolution of reproductive isolation in
ducks. Biological Journal of the Linnean Society77:193–200.
Turelli, M., and L. C. Moyle. 2007. Asymmetric postmating isolation: Darwin’s corollary to Haldane’s rule.
Genetics176:1059–1088.
Vacquier, V . 1962. Hydrostatic pressure has a selective effect on the copepodTigriopus. Science135:724–
725.
Vacquier, V ., and W. Belser. 1965. Sex conversion induced by hydrostatic pressure in the marine copepod
Tigriopuscalifornicus. Science150:1619–1621.
van Doorn, G., and M. Kirkpatrick. 2007. Turnover of sex chromosomes induced by sexual conflict. Nature
449:909–912.
23
van Doorn, G., and M. Kirkpatrick. 2010. Transitions between male and female heterogamety caused by
sex–antagonistic selection. Genetics186:629–645.
Vittor, B. A., 1971. Effects of the environment on fitness-related life history characters inTigriopuscalifor-
nicus. Ph.D. thesis, University of Oregon.
V oordouw, M., and B. Anholt. 2002a. Environmental sex determination in a splash pool copepod. Biological
Journal of the Linnean Society76:511–520.
V oordouw, M., and B. Anholt. 2002b. Heritability of sex tendency in a harpacticoid copepod, Tigriopus
californicus. Evolution56:1754–1763.
V oordouw, M., H. Robinson, and B. Anholt. 2005a. Paternal inheritance of the primary sex ratio in a
copepod. Journal of Evolutionary Biology18:1304–1314.
V oordouw, M., H. Robinson, G. Stebbins, A. Albert, and B. Anholt. 2005b. Larval density and the Charnov-
Bull model of adaptive environmental sex determination in a copepod. Canadian Journal of Zoology
83:943–954.
V oordouw, M. J., G. Stebbins, H. E. Robinson, M.-J. Perrot-Minnot, T. Rigaud, and B. R. Anholt. 2008.
Genetic variation in the primary sex ratio in populations of the intertidal copepod,Tigriopuscalifornicus,
is widespread on Vancouver Island. Evolutionary Ecology Research10:1007–1023.
Willett, C., and R. Burton. 2001. Viability of cytochrome C genotypes depends on cytoplasmic backgrounds
inTigriopuscalifornicus. Evolution55:1592–1599.
Willett, C. S. 2008. No evidence for faster male hybrid sterility in population crosses of an intertidal copepod
(Tigriopuscalifornicus). Genetica133:129–136.
Willett, C. S., and J. N. Berkowitz. 2007. Viability effects and not meoitic drive cause dramatic departures
from Mendelian inheritance for malic enzyme in hybrids of Tigriopus californicus populations. Journal
of Evolutionary Biology20:1196–1205.
Willett, C. S., T. G. Lima, I. Kovaleva, and L. Hatfield. in press. Chromosome-wide impacts on the
expression of incompatibilities in hybrids ofTigriopuscalifornicus. G3: Genes, Genomes, Genetics .
Wilson, C. A., S. K. High, B. M. McCluskey, A. Amores, Y . lin Yan, T. A. Titus, J. L. Anderson, P. Batzel,
M. J. C. III, M. Schartl, and J. H. Postlethwait. 2014. Wild sex in zebrafish: Loss of the natural sex
determinant in domesticated strains. Genetics198:1291–1308.
Wolff, J., E. Ladoukakis, J. Enriquez, and D. Dowling. 2014. Mitonuclear interactions: evolutionary con-
sequences over multiple biological scales. Philosophical Transactions of the Royal Society B: Biological
Sciences369:20130443.
Wright, A. E., R. Dean, F. Zimmer, and J. E. Mank. 2016. How to make a sex chromosome. Nature
Communications7:12087.
Wu, C.-I. 1992. A note on Haldane’s rule: Hybrid inviability versus hybrid sterility. Evolution 46:1584–
1587.
24
Chapter 1: Does the speciation clock tick more slowly in the absence of
heteromorphic sex chromosomes?
Barret C. Phillips and Suzanne Edmands
Introduction
The genetic basis of postzygotic isolation–one path through which speciation proceeds–has been one
of the most active areas of recent evolutionary research. Perhaps the most universal pattern thus far re-
vealed is that genetic interactions driving speciation frequently involve sex chromosomes. However, the
speciation process in the absence of sex chromosomes remains understudied, despite the many taxa without
chromosomal sex-determination (Bull 1983). Though some aspects of postzygotic isolation in organisms
with ’alternative’ sex-determination mechanisms were recently reviewed (Schilthuizen et al. 2011), many
specific predictions regarding the evolution of postzygotic isolation in organisms lacking sex chromosomes
remain untested. We synthesize these predictions and discuss appropriate methods for testing them, with a
primary focus on the prediction that sex chromosomes influence speciation rates (Rieseberg 2001). Our goal
here is to generate discussion and motivate tests of these hypotheses, rather than to present in-depth tests of
the hypotheses themselves.
Postzygotic isolation is attributable to genetic interactions
If species are groups of organisms reproductively isolated from other such groups (Mayr 1942), then
speciation is the process whereby one such group diverges into distinct, non-interbreeding lineages. As iso-
lated lineages diverge, accumulated genetic differences eventually cause postzygotic isolation, the sterility
and/or inviability of hybrid offspring. Postzygotic isolation is most commonly attributed to Dobzhansky-
Muller interactions (DMIs), where novel allelic combinations cause reduced fitness of hybrid offspring as
a result of epistatic (gene x gene) interactions (Coyne and Orr 2004, Presgraves 2010). Many DMIs are
attributable to interactions between sex chromosomes and autosomes. Because the so-called “rules” of spe-
ciation (Haldane’s rule, the large X-effect, and Darwin’s corollary; see below) can be attributed wholly or in
part to the action of sex chromosomes (Coyne and Orr 1989b, Turelli and Moyle 2007), sex chromosomes
are considered to play a special role in speciation (Presgraves 2008). Additionally, because sex chromo-
25
somes are apparently so prone to involvement in DMIs, it follows that they might also affect other aspects
of speciation, such as the rate at which it proceeds.
These “rules” of speciation stem largely from work on Drosophila and a few other groups with morpho-
logically differentiated (heteromorphic) sex chromosomes (DSCs). Although DSCs have arisen indepen-
dently and repeatedly in many lineages (Matsubara et al. 2006), they are far from a universal characteristic
of sexually reproducing eukaryotes (Bull 1983).
Sex determination mechanisms vary widely
Mechanisms determining sex in plants and animals, the two groups most commonly possessing discrete
sexes, are diverse and range widely over short phylogenetic distances (Bull 1983). Sex-determination mech-
anisms should be viewed as an evolutionary continuum. Although the known mechanisms can be grouped
into distinct categories–for instance, genetic/non-genetic, chromosomal/non-chromosomal, homogametic/
heterogametic–they grade into each other, as when homogametic sex chromosomes “degenerate” into het-
erogametic sex chromosomes. Among vertebrates, heterogamety evolved independently in birds (ZW),
snakes (ZW) and mammals (XY) (Matsubara et al. 2006), while the mechanisms of sex-determination in
amphibians and in other reptilian groups are more variable, including both homomorphic sex chromosomes
and environmental mechanisms (Bull 1983, Pokorn´ a and Kratochv´ ıl 2009). Invertebrates possess diverse
and often poorly characterized sex determining mechanisms. Sex-determination in insects is known to be
particularly variable–for instance, lepidopterans have DSCs (ZW), hymenopterans are haplodiploid, and
many–but not all–flies have DSCs (XY or ZW) (Bull 1983). Most plant species are hermaphroditic or mo-
noecious, although dioecy has evolved many times. DSCs are present in four angiosperm families, and
although most of the species within these families are not heterogametic, DSCs have evolved repeatedly
within two families (Charlesworth and Mank 2010). Thus the question of whether sex chromosomes influ-
ence speciation rates involves evolutionary processes across a wide range of multicellular organisms.
26
The “rules” of speciation invoke conflict between sex chromosomes and auto-
somes
Haldane’srule: The first “rule” of speciation is that when one sex is more severely affected by hybridiza-
tion (e.g., absent, numerically reduced, or less classically fit), it is usually the heterogametic sex (Haldane
1922). This depends on the sex chromosomes themselves, rather than on some sex-specific phenomenon—it
has been documented in cases of both male (XY) and female (ZW) heterogamety. Haldane’s rule holds for
both sterility and inviability (Coyne and Orr 1989b, Haldane 1922, Wu and Davis 1993) and in both plants
and animals (Brothers and Delph 2010).
Despite the near-ubiquity of Haldane’s rule, it appears to have no single cause. The leading explanation
is the dominance hypothesis, which implicates recessive alleles on the X (Z) chromosome that are masked
in the homozygous sex (XX/ZZ) but exposed in the hemizygous sex (XY/ZW) (Turelli and Orr 1995). A
second explanation with support is the faster-male hypothesis, which proposes that male sterility arises
faster than inviability in XY systems, perhaps because spermatogenesis is especially sensitive to disruption
or because male fertility genes evolve more quickly (Tao and Hartl 2003). However, this explanation has
received support only for sterility and is only applicable in XY systems (Wu and Davis 1993). The faster-
male phenomenon could conceivably be found in organisms without sex chromosomes, as discussed in a
recent review (Schilthuizen et al. 2011), but it is important to distinguish this from true manifestations of
Haldane’s rule, which by definition require heterogamety. A third hypothesis with some support is that
Haldane’s rule is attributable to sex-ratio meiotic drive (biased transmission of one sex chromosome over
the other) (McDermott and Noor 2010), a situation that may result in faster evolution in the heterogametic
sex (Tao and Hartl 2003).
The large X-effect: The second “rule” involves the tendency of genes contributing to hybrid breakdown
to localize disproportionately to the X (or Z) chromosome. The X (Z) chromosome has been shown to
impact hybrid fertility and/or viability much more than individual autosomes in many crosses, even when
the X is neither physically larger nor more gene-dense than the autosomes (Coyne and Orr 1989b). Though
initially proposed to explain both hybrid sterility and hybrid inviability, it appears to hold true more often for
hybrid male sterility than for hybrid female sterility or for hybrid viability in either sex (Presgraves 2008).
The large X-effect has been more controversial than Haldane’s rule, and its causes are less well under-
stood. Hypotheses currently favored invoke faster evolution of the X chromosome relative to autosomes
27
(faster-X), meiotic drive, and the sensitivity of spermatogenesis to the disruption of X chromosome inacti-
vation (Presgraves 2008).
Darwin’s corollary: The third “rule” is the observation that many crosses demonstrate asymmetric fitness
(hybrids with an A mother and a B father may be more severely affected than those with a B mother and
an A father). This could be explained by conflicts between a number of uniparentally-inherited factors
and the nuclear genome, including X-autosome interactions, nuclear-mitochondrial conflicts, and maternal
effects. In plants, additional potential sources of asymmetric dysfunction exist, including nuclear-plastid
conflicts, gametophyte-sporophyte interactions, and conflicts between the haploid male and diploid female
contributions to triploid endosperm function (Turelli and Moyle 2007).
Are speciation rates influenced by the presence of sex chromosomes?
In species with DSCs, the first manifestations of postzygotic isolation appear nearly always to involve
sex chromosome-autosome conflicts. These interactions underlie both Haldane’s rule and the large X-effect,
and Darwin’s corollary is also partially explained by them (Coyne and Orr 1989b, Turelli and Moyle 2007).
A logical prediction of the dominance theory (Haldane’s rule) is that postzygotic isolation should evolve
more slowly in taxa with small X chromosomes than in those with large X chromosomes (Turelli and Orr
1995), which is supported by evidence from Drosophila (Turelli and Begun 1997). An extension of this
logic is that postzygotic isolation should evolve even more slowly in taxa lacking DSCs (Rieseberg 2001).
Though proposed a decade ago, this prediction remains empirically unexamined.
Proper tests of this hypothesis will involve replicated, phylogenetically grounded comparisons of speci-
ation rates in sister clades with and without DSCs. Few, if any, datasets of the appropriate breadth and depth
have been compiled, andposthoc comparisons of the results of different studies are difficult because of the
variety of methods used to calculate speciation rates.
Measurements of speciation rates rely upon a “speciation clock”—that is, the assumption that repro-
ductive isolation accumulates at a constant (“clocklike”) rate (Coyne and Orr 1989a). The most accurate
calculations, therefore, are derived from controlled crosses that quantify intertaxon reproductive isolation,
which can be incorporated into measures of biological speciation intervals (BSIs). As studies of this sort are
impracticable for many groups and also for examining large numbers of taxa simultaneously, a less rigorous
metric, the net diversification interval (NDI), is often calculated instead. NDIs use the number of extant
28
taxa within a clade and its age to extrapolate a speciation rate (Coyne and Orr 2004). Numerous methods of
calculating NDIs have been devised that may or may not incorporate, for example, estimates of extinction
rates. However, NDIs calculated with the same methodology and with comparably accurate estimates of
clade age and diversity may in the future provide insight into the question of whether sex chromosomes
influence speciation rates.
Data from a recent study provide limited evidence for the hypothesis that speciation occurs more rapidly
in clades possessing DSCs. NDIs calculated for 28 families of reptiles and birds (Eo and DeWoody 2010)
show that, in general, lizards and snakes (squamates)—where DSCs are common—have speciated more
quickly than turtles and crocodilians–where DSCs are rare and absent, respectively (Figure ??) . Further-
more, within groups, the same pattern can be seen at the family level. In snakes (range: 5.4–16.8 MY),
Typhlopidae (16.8 MY) has the slowest NDI and is the only family lacking DSCs. In turtles (range: 12.7–
68.3 MY), Geoemydidae (12.7MY), the only family in which DSCs have been described, has the shortest
NDI. In lizards (range: 9.7–17.5 MY), the only family examined that consistently possesses DSCs (Am-
phisbaenidae: 9.7 MY) has the fastest NDI, and the one family examined that consistently lacks DSCs
(Agamidae: 13.3 MY) has the second-lowest NDI (Pokorn´ a and Kratochv´ ıl 2009, Vitt and Caldwell 2009).
However, this pattern is contradicted by birds, which have NDIs comparable to turtles and crocodilians but
universally possess ZW sex chromosomes (Bull 1983).
What genetic conflicts might play roles in postzygotic isolation when DSCs
are absent?
In the absence of sex chromosome-autosome interactions, other genetic conflicts must necessarily play
greater roles in postzygotic isolation. Cytonuclear conflicts—those between a uniparentally inherited cy-
toplasmic factor, such as mitochondria, and the nuclear genome—seem to play large roles in taxa with
alternative mechanisms of sex-determination, such as yeast (haploid mating types), the waspNasonia (hap-
lodiploidy), and the copepod Tigriopus (polygenic sex determination) (Chou and Leu 2010, Ellison and
Burton 2006, Ellison et al. 2008, Willett 2008). Though also known to exist in groups possessing DSCs,
cytonuclear conflicts are not among the primary incompatibilities identified in these organisms (Presgraves
2010). Beyond cytonuclear conflicts, other genetic conflicts that might be expected to play important roles
in postzygotic isolation include autosome-autosome incompatibilities and chromosomal rearrangements.
29
Figure 1: Net diversification intervals (NDIs) in reptiles and birds. Mean NDIs ( standard error) are mapped onto a
phylogenetic tree. Heteromorphic sex chromosomes are present in birds, common in squamates (lizards and snakes),
rare in turtles, and absent in crocodilians. Sex determination data are from (Bull 1983, Eo and DeWoody 2010, Pokorn´ a
and Kratochv´ ıl 2009). Phylogeny and NDI data are from (Eo and DeWoody 2010).
Indeed, postzygotic isolation attributable to chromosomal rearrangements (“chromosomal speciation”) ap-
pears to play a larger role in plants than in animals, possibly because plants often lack DSCs which would
otherwise be the dominant source of incompatibility (Rieseberg and Blackman 2010). Further genetic map-
ping of incompatibilities in non-DSC species will provide increased understanding of the number and nature
of these conflicts.
Do sterility and inviability accumulate at different rates in the absence of
DSCs?
InDrosophila and Lepidoptera, hybrid sterility in the heterogametic sex evolves faster than hybrid steril-
ity in the homogametic sex and also faster than hybrid inviability in either sex (Presgraves 2002, Tao and
Hartl 2003, Wu and Davis 1993). In birds, hybrid sterility arises before hybrid inviability (Price and Bouvier
2002). If this pattern is driven by rapidly evolving gametogenesis genes in the heterogametic sex (Tao and
30
Hartl 2003), then taxa lacking a heterogametic sex should accumulate inviability and sterility at comparable
rates in both sexes. In Tigriopus (polygenic sex determination) males do not accumulate hybrid sterility
faster than females (Willett 2008). Studies using tomatoes also support this idea, suggesting both that male
and female sterility may accumulate at the same rate (Moyle and Graham 2005, Moyle and Nakazato 2010)
and that similar numbers of QTL underlie pollen and seed sterility in two of three crosses (Moyle and
Nakazato 2010). Additional explicit tests of this question are needed, however, in order to characterize this
as a pattern in non-DSC taxa.
Conclusions and Prospects
Sex chromosomes play a large role in the process of speciation as we currently understand it. The im-
pacts of Haldane’s rule and the large X-effect are both profound and widespread, implying that sex chromo-
somes evolve incompatibilities more rapidly than do autosomes. Thus, we might expect sex chromosomes
to accelerate the rate of speciation. Indeed, some of the available data supports this. However, it is difficult
to comment on the universality of this pattern when data are available for so few groups.
The best evidence for an influence of sex chromosomes on speciation rates will come from comparative
studies in phylogenetically well-defined groups in which DSCs have clearly been gained and lost. Squamates
may be a particularly attractive group for such studies because of the repeated evolution of heterogamety
(both XY and ZW), as well as an apparently large number of taxa with environmental sex-determination.
Clearly, knowledge of the genetics of speciation for taxa with alternative mechanisms of sex-determination
remains woefully behind that for those with DSCs. However, this is changing rapidly with the development
of species lacking DSCs as alternative models of speciation (e.g., Mimulus, Helianthus, Solanum, Nasonia,
Tigriopus). Emerging evidence from these taxa suggests that, in comparison with DSC-possessing species,
speciation in the absence of sex chromosomes may proceed very differently indeed.
Acknowledgements
Thanks to an anonymous reviewer and to Wai Leong, Brad Foley, and Jacob Cram for comments and
suggestions that have improved this manuscript. This work was supported by an NSF Graduate Research
Fellowship to BCP and by NSF grant DEB-0743472 to SE.
31
References
Brothers, A. N., and L. F. Delph. 2010. Haldane’s rule is extended to plants with sex chromosomes.
Evolution 64:3643–3648.
Bull, J. 1983. Evolution of Sex Determining Mechanisms. Benjamin/Cummings.
Charlesworth, D., and J. E. Mank. 2010. The birds and the bees and the flowers and the trees: Lessons from
genetic mapping of sex determination in plants and animals. Genetics 186:9–31.
Chou, J.-Y ., and J.-Y . Leu. 2010. Speciation through cytonuclear incompatibility: Insights from yeast and
implications for higher eukaryotes. BioEssays 32:401–411.
Coyne, J., and H. Orr. 1989a. Patterns of speciation inDrosophila. Evolution 43:362–381.
Coyne, J. A., and H. A. Orr, 1989b. Speciation and Its Consequences, Chapter: Two rules of speciation,
pages 180–207 . Sinauer Associates.
Coyne, J. A., and H. A. Orr. 2004. Speciation. Sinauer Associates.
Ellison, C. K., and R. S. Burton. 2006. Disruption of mitochondrial function in interpopulation hybrids of
Tigriopuscalifornicus. Evolution 60:1382–1391.
Ellison, C. K., O. Niehuis, and J. Gadau. 2008. Hybrid breakdown and mitochondrial dysfunction in hybrids
ofNasonia parasitoid wasps. Journal of Evolutionary Biology 21:1844–1851.
Eo, S. H., and J. A. DeWoody. 2010. Evolutionary rates of mitochondrial genomes correspond to diversifi-
cation rates and to contemporary species richness in birds and reptiles. Proceedings of the Royal Society
B: Biological Sciences 277:3587–92.
Haldane, J. 1922. Sex ratio and unisexual sterility in hybrid animals. Journal of Genetics 12:101–109.
Matsubara, K., H. Tarui, M. Toriba, K. Yamada, C. Nishida-Umehara, K. Agata, and Y . Matsuda. 2006.
Evidence for different origin of sex chromosomes in snakes, birds, and mammals and step-wise differen-
tiation of snake sex chromosomes. Proceedings of the National Academy of Sciences 103:18190–18195.
Mayr, E. 1942. Systematics and the Origin of Species. Columbia University Press, New York.
McDermott, S., and M. Noor. 2010. The role of meiotic drive in hybrid male sterility. Philosophical
Transactions of the Royal Society B: Biological Sciences 365:1265–1272.
Moyle, L., and E. Graham. 2005. Genetics of hybrid incompatibility betweenLycopersiconesculentum and
L.hirsutum. Genetics 169:355–373.
Moyle, L. C., and T. Nakazato. 2010. Hybrid incompatibility “snowballs” betweenSolanum species. Science
329:1521–3.
Pokorn´ a, M., and L. Kratochv´ ıl. 2009. Phylogeny of sex-determining mechanisms in squamate reptiles:
Are sex chromosomes an evolutionary trap? Zoological Journal of the Linnean Society 156:168–183.
Presgraves, D. 2002. Patterns of postzygotic isolation in Lepidoptera. Evolution 56:1168–1183.
Presgraves, D. 2010. The molecular evolutionary basis of species formation. Nature Reviews Genetics
11:175–180.
32
Presgraves, D. C. 2008. Sex chromosomes and speciation inDrosophila. Trends in Genetics 24:336–343.
Price, T., and M. Bouvier. 2002. The evolution of F1 postzygotic incompatibilities in birds. Evolution
56:2083–2089.
Rieseberg, L. 2001. Chromosomal rearrangements and speciation. Trends in Ecology and Evolution
16:351–358.
Rieseberg, L. H., and B. K. Blackman. 2010. Speciation genes in plants. Annals of Botany pages 1–17.
Schilthuizen, M., M. C. W. G. Giesbers, and L. W. Beukeboom. 2011. Haldane’s rule in the 21st century.
Heredity 107:95–102.
Tao, Y ., and D. Hartl. 2003. Genetic dissection of hybrid incompatibilities between Drosophila simulans
and D. mauritiana: III. Heterogeneous accumulation of hybrid incompatibilities, degree of dominance,
and implications for Haldane’s rule. Evolution 57:2580–2598.
Turelli, M., and D. Begun. 1997. Haldane’s rule and X-chromosome size inDrosophila. Genetics 147:1799–
1815.
Turelli, M., and L. C. Moyle. 2007. Asymmetric postmating isolation: Darwin’s corollary to Haldane’s rule.
Genetics 176:1059–1088.
Turelli, M., and H. Orr. 1995. The dominance theory of Haldane’s rule. Genetics 140:389–402.
Vitt, L. J., and J. P. Caldwell. 2009. Herpetology. Third edition edition. Academic Press, New York.
Willett, C. S. 2008. No evidence for faster male hybrid sterility in population crosses of an intertidal copepod
(Tigriopuscalifornicus). Genetica 133:129–136.
Wu, C., and A. Davis. 1993. Evolution of postmating reproductive isolation: The composite nature of
Haldane’s rule and its genetic bases. The American Naturalist 142:187–212.
33
Chapter 2: Phylogeography ofTigriopuscalifornicus, revisited: Highly
divergent populations in Baja California, Mexico constitute a novel species
Barret C. Phillips and Suzanne Edmands
Abstract
The marine intertidal copepod Tigriopus californicus Baker 1912 demonstrates remarkable genetic di-
vergence across its range on the Pacific coast of North America. Previous studies, however, have focused on
a single mitochondrial locus and have failed to resolve relationships both between highly divergent regional
clades and within northern populations, which show reduced differentiation, probably as a consequence of
post-Pleistocene expansion. Despite this remarkable genetic divergence, most populations remain interfer-
tile under laboratory conditions. Populations from the southern extreme of the species range, however, are
unable to hybridize with those occurring further north. Here, we resolve relationships between and within
regional clades using both nuclear and mitochondrial loci. We demonstrate that populations generally show
a pattern of isolation-by-distance, confirm the previously observed reduced differentiation at northern lat-
itudes, and find evidence for a species-level break between geographically continuous but reproductively
isolated populations.
Introduction
Speciation processes are complex and often involve opposing forces–gene flow homogenizes popula-
tions, while local adaptation and genetic drift may drive them apart. Coastal environments might be ex-
pected to be especially impacted by gene flow, as an open ocean and strong currents are potential sources
of homogenization over long distances. On the other hand, the patchy nature of specialized coastal habitat,
such as tide pools, could promote the accumulation of genetic divergence in taxa lacking a planktonic larval
stage.
The harpacticoid copepod Tigriopus californicus inhabits supralittoral splash pools along the Pacific
coast of North America. Its apparent restriction to the upper regions of the tidal zone has resulted in high,
and temporally stable, genetic differentiation between even adjacent outcrops (Burton and Feldman 1981,
Burton 1997, Edmands 2001, Handschumacher et al. 2010). Since the original description of “Tisbe cali-
34
fornica” from tide pools in Orange County, California (Baker 1912), the same species has been documented
from southern Alaska to Baja California Sur, Mexico (Dethier 1980, Peterson et al. 2013). Previous in-
vestigations of even geographically proximate T. californicus populations have revealed astonishing levels
of genetic differentiation (Burton et al. 1979, Edmands 2001, Peterson et al. 2013). This remarkable level
of genetic divergence is not, however, accompanied by complete reproductive isolation–even populations
with 20% mitochondrial divergence can produce hybrids in the laboratory (Burton et al. 2006, Pritchard
and Edmands 2012, Chapter 3, Chapter 4). ] However, the exception to this pattern of interpopulation
compatibility occurs at the edge of the species range. Ganz and Burton (1995) first described a single case
of extreme genetic divergence and reproductive incompatibility in a T. californicus population from central
Baja California. The PA (Playa Altimira) population was extremely genetically divergent from otherT.cali-
fornicus populations and shared no allozyme alleles (at 7 loci) with the nearest other sampledT.californicus
population, PBJ (Punta Baja), some 230km away. Additionally, most crosses between animals from the
PA population and animals from populations in northern Baja California or southern California failed to
produce F
1
or F
2
offspring, leading Ganz and Burton (1995) to suggest that the PA population might belong
to a semi–species or a sibling species. More extensive sampling of the region identified a 12km boundary
across which populations are nearly completely reproductively isolated and found that mitochondrial COI
sequences support a phylogenetic break in the same place (Peterson et al. 2013).
To date, phylogeographic investigations of the full species range of T. californicus have been restricted
to the mitochondrial COI locus (Edmands 2001, Peterson et al. 2013). Interpopulation divergence as high as
26% has been documented within T. californicus at COI, and phylogenetic analysis with this locus fails to
resolve many interpopulation relationships, due to both a dearth of divergent sites at high latitudes–possibly
as a result of post-glacial population expansion, and to saturation of the locus at higher divergence levels.
Additional loci are particularly necessary to resolve the species status of the southern populations below the
phylogeographic break in Baja California, which have previously been suggested to be a semi–species or a
sibling species toT.californicus (Ganz and Burton 1995) or an incipient species (Peterson et al. 2013).
Here we sought to further investigate the phylogeography of T. californicus across the greater part of
the species range, using both nuclear and mitochondrial gene sequences. Our results support previously
established patterns of divergence; clarify relationships between and within regional clades; and demonstrate
the presence of a cryptic species in the southern portion of the range.
35
Materials and Methods
Collection of study populations
Populations were collected from the San Juan Islands, WA to Los Angeles, CA in September 2011
and from San Diego, CA to Baja California Sur, Mexico in October 2011 (Table 1). Live animals were
collected into 1L bottles and brought into the laboratory for culture. Thirty-two animals per population were
rinsed in distilled water, dried, and frozen (80

C) within one week of arriving in the laboratory. Outgroup
samples, from congenersT.brevicornis andT.japonicus, were collected in Iceland and Hong Kong in 2011,
maintained in the laboratory, and sampled in January 2012.
Amplification and sequencing of nuclear and mitochondrial loci
DNA was extracted from frozen individuals using a homemade lysis buffer (10mM Tris pH 8.3, 50mM
KCl, 0:5% Tween-20; 50uL per individual) supplemented with 200 ug/mL proteinase K. Samples were
incubated at65

C for one hour, followed by a 15-minute heat inactivation of proteinase K at95

C. Extracted
DNA was subsequently stored at20

C and used as needed.
Mitochondrial 12S ribosomal DNA primers were designed based on alignments of full mitochondrial
genomes from 3 populations (Burton et al. 2007). Primers to amplify nuclear 18S ribosomal DNA, ITS1, and
5.8S ribosomal DNA were designed based on primer and genetic sequences in Burton et al. (2005), Huys
et al. (2009). Two linkage maps based on population-diagnostic SNPs between the San Diego (SD) and
Santa Cruz (SC) populations were published previously (Pritchard et al. 2011, Foley et al. 2011). We also
assayed this panel of 190 SNPs on all extant cultures in our lab in 2010, including populations ofTigriopus
japonicus and T. brevicornis (Phillips, Leong, and Edmands, unpublished data). A subset of these loci
were successfully genotyped in these populations and were deemed likely to be amplifiable using common
primers. We used the source transcriptome assembly from SC and SD (Barreto et al. 2011) to design primers
that would amplify a larger region of the transcript and tested their amplification in extractions of 10 pooled
individuals in multiple populations. Those that successfully amplified in allT.californicus populations were
subsequently used for PCR and sequencing with individual animal extractions (Table 2).
PCRs were performed using Choice-Taq polymerase (Denville Scientific). Amplification conditions
were 94

C/5 minutes; 94

C/1 minute, annealing 45 seconds, 72

C/extension time for 40 cycles; 72

C/2
minutes. PCR products were assessed by gel electrophoresis and then SAP-treated and Sanger sequenced
36
by Retrogen, Inc. (San Diego, CA).
Sequences were trimmed by hand in Sequencher (Gene Codes Corporation), and consensus forward and
reverse contigs were exported. Alignments were performed in MUSCLE Edgar (2004) and checked by hand
using ClustalX Larkin et al. (2007). Because ITS1 and 5.8S congener sequences were so divergent, only T.
californicus sequences were initially aligned in MUSCLE, and each congener was sequentially appended to
the existing alignment in ClustalX.
Analysis
Appropriate nucleotide substitution models for each locus were determined using jModelTest 2 (Guin-
don and Gascuel 2003, Darriba et al. 2012). Inter- and intrapopulation population genetic statistics were
assessed in MEGA6, where a maximum parsimony consensus tree was also constructed for the full con-
catenated set of loci Tamura et al. (2011). Genetic distances were calculated as between group average dis-
tances, under the Tamura 3-parameter model with gamma-distributed between-site rate variation for mt12S
sequences and the Kimura 2-paramater model with gamma-distributed between-site rate variation for nu-
clear sequences. Genetic diversity metrics were calculated using the same models as for genetic distance, as
was the coefficient of differentiation (G
ST
). For all measures, missing data were removed with pairwise dele-
tions, and variance was estimated by bootstrapping the datasets 1000 times. Geographic distances between
populations were calculated using the fossil package in R, and all other statistical analyses were conducted
in R (R Development Core Team 2011, Vavrek Palaeontologia Electronica).
Bayesian phylogenetic trees were constructed using MrBayes 3.2.2 Hulsenbeck et al. (2001). Trees were
constructed for the each locus individually and for the full concatenated gene set, which was partitioned by
locus. Mixed substitution models and an invariant + gamma site distribution were used for each run with
unlinked state frequencies, substitution rates, gamma shape parameters and proportions of invariant sites.
MCMC runs were performed using 4 chains, sampling every 500 generations, and until all simultaneous
walks achieved convergence (y< 0:01; minimum of 1,000,000 generations). The first 25% of generations
were discarded as burn-in during parameter estimation and consensus tree generation.
37
Table 1: Sampling locations forTigriopus populations
Species Collection location Abbreviation Latitude Longitude
Tigriopuscalifornicus
Friday Harbor, WA FHL 48.55

N 123.01

W
Sunset Beach, WA SUN 48.50

N 122.70

W
Yaquina Head, OR YH 44.83

N 124.07

W
Strawberry Hill Wayside, OR SH 44.25

N 124.12

W
Shelter Cove, CA SHC 40.02

N 124.07

W
Mendocino Head, CA MDH 39.31

N 123.81

W
Bodega Bay, CA BB 38.32

N 123.07

W
Santa Cruz, CA SC 36.95

N 122.05

W
Point Lobos, CA PL 36.52

N 121.95

W
Leo Carillo State Beach, Malibu, CA LC 34.03

N 118.68

W
Santa Cruz Island, CA SCI 34.02

N 119. 68

W
Point Dume, CA PD 34.00

N 118.81

W
Flat Rock, CA FR 33.80

N 118.41

W
Abalone Cove, CA AB 33.73

N 118.37

W
Royal Palms, Palos Verdes, CA RP 33.72

N 118.32

W
Laguna Beach, CA LB 33.54

N 117.79

W
Santa Catalina Island, CA CAT 33.45

N 118.49

W
La Jolla, CA LJ 32.83

N 117.27

W
Bird Rock, CA BR 32.82

N 117.27

W
San Diego, CA SD 32.75

N 117.26

W
Punta Morro, BCN PMO 31.86

N 116.67

W
Punta Banda, BCN PBN 31.73

N 116.73

W
Los Moros Colorados, BCN LMC 29.42

N 115.09

W
Los Moros Colorados South 3, BCN LMCS 29.39

N 115.01

W
El Cuervito, BCN ECV 29.31

N 114.91

W
Playa Altamira, BCN PA 28.53

N 114.10

W
Bahia Asuncion, BCS BA 27.14

N 114.31

W
T.japonicus
Shek O, Hong Kong SO 22.23

N 114.25

E
Cape D’Aguilar, Hong Kong CDA 22.21

N 114.14

E
T.brevicornis
Borgjarnes, Iceland BG 64.54

N 21.93

W
38
Table 2: Loci utilized
Locus Primers Best BLAST hit
Name Type Alignment length Name Sequence Gene Species E-value
12S Mitochondrial 687 12S–F CCAAGATACTTTAGGGATAACAGC 12S ribosomal RNA T.californicus 2e-133
12S–R CTGTTCTATCAAAACAGCCCT
18S Nuclear 1664 18Sf TACCTGGTTGATCCTGCCAG 18S ribosomal RNA T.californicus 0
18S–614r TCCAACTACGAGCTTTTTAACC
18S–554f AAGTCTGGTGCCAGCAGCCGC
18S–1282r TCACTCCACCAACTAAGAACGGC
18S–1150f(p2) ATTGACGGAAGGGCACCACCAG
18Sr TAATGATCCTTCCGCAGGTTCAC
ITS1/5.8S* Nuclear 549/178 ITS–F GAATTCCCAGTAAGCGCAAG Internal transcribed spacer 1 T.californicus 0
ITS–R GCTTAAATTCGGCGGGTAAT 5.8S ribosomal RNA T.californicus 1e-30
Tc22152 Nuclear (chr 2) 313 Tc22152–F ATGGCCACAAGCTGAAGAGT TBC1 domain family 22a Daniorerio 2e-32
Tc22152–R GAATCCTTCCAACTCCGACA
Tc25225 Nuclear (chr 12) 410 Tc25225–F–2 CGAGCTCTCATGACCAACAA Discs, large homolog 5 Homosapiens 8e-12
Tc25225–R TGATTTGAGCGCACATTAGC
Tc30567 Nuclear (chr 10) 393 Tc30567–F GGCCTTTCTTTGAGTCAGG OPA-3-like protein Calligusrogercresseyi 1e-39
Tc30567–R TTCGCACCTATGTCTGCATC
Tc28684 Nuclear (chr 9) 459 Tc28684–F TGTCACAACCACCGTCATCT Twitchin Caenorhabditiselegans 9e-35
Tc28684–R TGGTGGCAGTGAATTAACCA
*Amplicon encompasses the end of the 18S gene, ITS1, the 5.8S gene, ITS2, and part of the 28S gene; loci were separated according to the boundaries delineated in Burton et al. 2005
39
0.5
BB-2
SHC-3
MDH-1
BA-12
LJ-2
AB-6
AB-8
PA-4
PA-2
SHC-7
PBN-4
AB-4
FHL-9
PL-2
FHL-5
SCI-11
LMCS-8
BB-1
PA-8
PL-1
LMC-2
ECV-3
ECV-4
LC-5
AB-9
FHL-3
SHC-9
SC-1
LB-17
ECV-6
SH-10
BB-6
LC-1
YH-2
SHC-5
CAT-18
LMC-3
LMC-4
LJ-6
FHL-7
PD-1
BA-13
LMC-9
BR-6
SUN-1
SH-9
LMCS-4
SD-16
SC-13
CAT-16
PA-10
FHL-10
SCI-14
SCI-15
ECV-1
ECV-7
LB-14
CDA-1
SH-6
SD-11
PBN-5
CAT-12
RP-1
LMC-5
BA-15
SC-18
SH-2
MDH-2
LC-9
BB-9
SD-14
BA-17
SH-1
LMCS-1
SC-2
BB-10
BB-7
LMC-1
SD-17
PBN-7
YH-1
SCI-13
PBN-2
AB-2
AB-10
BA-14
PBN-6
PA-3
CAT-13
PA-1
LJ-3
LB-12
PA-9
SHC-1
CAT-14
SHC-10
SCI-18
SH-8
LMC-10
PA-7
SH-7
BB-5
AB-7
LC-4
PD-2
PBN-10
LMCS-2
FHL-6
BB-3
BR-4
FHL-1
SHC-4
CAT-17
LMCS-5
LMC-8
SCI-16
AB-3
RP-2
ECV-10
SHC-8
PBN-8
SUN-2
CAT-1
SD-13
SH-5
CAT-11
FHL-2
LC-3
LMCS-7
PBN-9
LC-6
BB-8
AB-5
LB-13
LMC-6
PA-5
SD-15
BR-9
FHL-8
BA-16
PBN-1
LC-7
LJ-5
CDA-2
SHC-2
ECV-2
BR-8
SHC-6
BB-4
SCI-12
LB-16
SD-12
LJ-1
BR-5
BR-7
FHL-4
CAT-15
LC-2
PA-6
LMC-7
BR-3
AB-1
1
1
1
1
1
1
1
0.92
1
1
1
1
1
1
1
0.83
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.99
1
1
1
1
1
1
1
1
0.73
1
1
1
1
1
1
1
1
1
0.82
1
1
1
1
1
1
1
1
0.96
0.6
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.7
1
1
1
1
1
1
0.77
1
1
1
1
0.95
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.98
1
1
1
1
1
1
1
1
1
0.61
0.59
0.84
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.79
0.84
1
0.64
1
1
1
1
1
1
1
1
1
0.82
1
1
1
1
0.95
1
1
1
1
1
1
1
1
1
1
1
1
0.65
1
1
1
1
1
1
1
1
1
1
0.7
1
1
1
0.95
0.97
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Washington
Oregon
Humboldt County, CA
Sonoma County, CA
Santa Cruz Island, CA
Malibu, CA
Palos Verdes, CA
La Jolla, CA
Northern Baja California, Mexico
Central Baja California, Mexico
Central Baja California, Mexico
Baja California Sur, Mexico
Tigriopus japonicus
Monterey Bay, CA
San Diego, CA
Orange County, CA
Santa Catalina Island, CA
Mendocino County, CA
South Central Baja California, Mexico
1
1
1
1
0.98
0.79
1
1
1
1
1
1
1
0.92
0.84
1
0.99
0.95
0.84
0.61
0.96
0.77
0.65
0.60
1
1
1
1
1
1
0.95
0.59
0.82
0.71
0.71
1
0.64
0.83
0.73
0.95
1
1
1
A
B
C
D
Figure 1: Mitochondrial 12S Bayesian consensus tree for 25 populations ofT.californicus and one popula-
tion of T. japonicus. Branches and individuals are colored by regional clades, which are designated by the
letters A–D (T. californicus), or by species (T. japonicus). Branch numbers indicate posterior probabilities.
MCMC was run for 2,500,000 generations.
40
Table 3: mt12S mean pairwise genetic distance (standard error).
AB BA BB BR CAT CDA ECV FHL LB LC LJ LMC LMCS MDH PA PBN PD PL RP SC SCI SD SH SHC SUN YH
AB 0.269 (0.033) 0.090 (0.013) 0.065 (0.011) 0.062 (0.011) 0.328 (0.035) 0.234 (0.029) 0.098 (0.014) 0.071 (0.011) 0.057 (0.010) 0.070 (0.011) 0.120 (0.016) 0.120 (0.016) 0.093 (0.014) 0.233 (0.029) 0.098 (0.015) 0.055 (0.010) 0.101 (0.015) 0.006 (0.003) 0.102 (0.015) 0.086 (0.013) 0.072 (0.011) 0.097 (0.014) 0.087 (0.013) 0.098 (0.014) 0.097 (0.014)
BA 0.273 (0.032) 0.267 (0.033) 0.278 (0.034) 0.427 (0.046) 0.076 (0.012) 0.285 (0.035) 0.276 (0.033) 0.281 (0.035) 0.278 (0.034) 0.295 (0.034) 0.295 (0.034) 0.277 (0.034) 0.071 (0.011) 0.323 (0.038) 0.279 (0.034) 0.299 (0.036) 0.267 (0.032) 0.301 (0.036) 0.296 (0.035) 0.274 (0.033) 0.278 (0.034) 0.275 (0.034) 0.285 (0.035) 0.278 (0.034)
BB 0.106 (0.016) 0.097 (0.014) 0.331 (0.035) 0.252 (0.030) 0.037 (0.008) 0.105 (0.015) 0.093 (0.013) 0.110 (0.016) 0.108 (0.015) 0.108 (0.015) 0.038 (0.008) 0.241 (0.028) 0.107(0.015) 0.089 (0.013) 0.063 (0.010) 0.092 (0.014) 0.064 (0.010) 0.100 (0.014) 0.103 (0.014) 0.040 (0.008) 0.037 (0.008) 0.037 (0.008) 0.04 (0.008)
BR 0.060 (0.011) 0.337 (0.035) 0.261 (0.032) 0.108 (0.015) 0.029 (0.007) 0.06 (0.011) 0.005 (0.002) 0.116 (0.016) 0.116 (0.016) 0.100 (0.015) 0.260 (0.032) 0.107 (0.016) 0.063 (0.011) 0.106 (0.016) 0.064 (0.010) 0.102 (0.016) 0.089 (0.013) 0.030 (0.007) 0.105 (0.015) 0.097 (0.014) 0.108 (0.015) 0.105 (0.015)
CAT 0.334 (0.035) 0.260 (0.031) 0.094 (0.014) 0.067 (0.011) 0.012 (0.004) 0.063 (0.011) 0.108 (0.015) 0.108 (0.015) 0.087 (0.013) 0.258 (0.031) 0.090 (0.014) 0.016 (0.005) 0.095 (0.014) 0.060 (0.011) 0.096 (0.014) 0.078 (0.012) 0.068 (0.011) 0.092 (0.013) 0.086 (0.013) 0.094 (0.014) 0.092 (0.013)
CDA 0.396 (0.042) 0.313 (0.033) 0.338 (0.035) 0.334 (0.035) 0.345 (0.035) 0.348 (0.036) 0.348 (0.036) 0.318 (0.034) 0.409 (0.044) 0.332 (0.035) 0.324 (0.033) 0.352 (0.037) 0.334 (0.036) 0.350 (0.037) 0.347 (0.037) 0.342 (0.035) 0.316 (0.033) 0.316 (0.034) 0.313 (0.033) 0.316 (0.033)
ECV 0.267 (0.032) 0.270 (0.032) 0.256 (0.031) 0.268 (0.032) 0.302 (0.036) 0.302 (0.036) 0.259 (0.031) 0.015 (0.005) 0.274 (0.033) 0.250 (0.030) 0.275 (0.033) 0.233 (0.029) 0.277 (0.033) 0.280 (0.032) 0.267 (0.032) 0.260 (0.031) 0.261 (0.031) 0.267 (0.032) 0.260 (0.031)
FHL 0.104 (0.014) 0.092 (0.013) 0.111 (0.016) 0.114 (0.016) 0.114 (0.016) 0.013 (0.004) 0.255 (0.031) 0.122 (0.017) 0.092 (0.013) 0.055 (0.010) 0.099 (0.014) 0.056 (0.010) 0.117 (0.015) 0.102 (0.014) 0.015 (0.005) 0.013 (0.004) 0.000 (0.000) 0.015 (0.005)
LB 0.067 (0.011) 0.031 (0.007) 0.111 (0.016) 0.111 (0.016) 0.097 (0.014) 0.268 (0.032) 0.109 (0.016) 0.071 (0.012) 0.109 (0.016) 0.069 (0.011) 0.105 (0.015) 0.091 (0.013) 0.002 (0.002) 0.100 (0.014) 0.094 (0.013) 0.104 (0.014) 0.100 (0.014)
LC 0.063 (0.011) 0.108 (0.015) 0.108 (0.015) 0.087 (0.013) 0.255 (0.031) 0.086 (0.013) 0.006 (0.003) 0.095 (0.014) 0.055 (0.01) 0.096 (0.014) 0.071 (0.011) 0.068 (0.011) 0.090 (0.013) 0.084 (0.012) 0.092 (0.013) 0.090 (0.013)
LJ 0.120 (0.016) 0.120 (0.016) 0.104 (0.015) 0.266 (0.032) 0.112 (0.016) 0.067 (0.011) 0.110 (0.016) 0.069 (0.011) 0.106 (0.016) 0.094 (0.013) 0.032 (0.007) 0.109 (0.015) 0.101 (0.014) 0.111 (0.016) 0.109 (0.015)
LMC 0.000 (0.000) 0.111 (0.015) 0.287 (0.033) 0.049 (0.009) 0.114 (0.015) 0.119 (0.016) 0.117 (0.016) 0.120 (0.016) 0.133 (0.018) 0.109 (0.016) 0.116 (0.016) 0.110 (0.015) 0.114 (0.016) 0.116 (0.016)
LMCS 0.111 (0.015) 0.287 (0.033) 0.049 (0.009) 0.114 (0.015) 0.119 (0.016) 0.117 (0.016) 0.120 (0.016) 0.133 (0.018) 0.109 (0.016) 0.116 (0.016) 0.110 (0.015) 0.114 (0.016) 0.116 (0.016)
MDH 0.247 (0.030) 0.117 (0.016) 0.087 (0.013) 0.056 (0.010) 0.094 (0.014) 0.057 (0.010) 0.114 (0.015) 0.095 (0.013) 0.007 (0.003) 0.004 (0.002) 0.013 (0.004) 0.007 (0.003)
PA 0.281 (0.034) 0.249 (0.030) 0.26 (0.031) 0.232 (0.029) 0.261 (0.032) 0.272 (0.031) 0.265 (0.032) 0.248 (0.030) 0.249 (0.030) 0.255 (0.031) 0.248 (0.030)
PBN 0.088 (0.014) 0.120 (0.017) 0.095 (0.014) 0.117 (0.017) 0.109 (0.015) 0.107 (0.016) 0.122 (0.017) 0.116 (0.016) 0.122 (0.017) 0.122 (0.017)
PD 0.095 (0.014) 0.056 (0.010) 0.097 (0.014) 0.074 (0.011) 0.072 (0.012) 0.090 (0.013) 0.084 (0.012) 0.092 (0.013) 0.090 (0.013)
PL 0.103 (0.015) 0.005 (0.003) 0.109 (0.015) 0.110 (0.016) 0.057 (0.011) 0.055 (0.010) 0.055 (0.010) 0.057 (0.011)
RP 0.104 (0.015) 0.087 (0.013) 0.070 (0.011) 0.099 (0.014) 0.089 (0.013) 0.099 (0.014) 0.099 (0.014)
SC 0.111 (0.015) 0.107 (0.016) 0.058 (0.011) 0.056 (0.010) 0.056 (0.010) 0.058 (0.011)
SCI 0.088 (0.013) 0.118 (0.016) 0.110 (0.015) 0.117 (0.015) 0.118 (0.016)
SD 0.097 (0.014) 0.092 (0.013) 0.102 (0.014) 0.097 (0.014)
SH 0.008 (0.003) 0.015 (0.005) 0.000 (0.000)
SHC 0.013 (0.004) 0.008 (0.003)
SUN 0.015 (0.005)
YH
41
Table 4: Population genetic summary statistics for mt12S and 18S ribosomal DNA (standard error)
Locus
within

between
D
xy
G
ST
mt12S 0.001 (0.000) 0.132 (0.011) 0.132 (0.011) 0.995 (0.001)
18S 0.000 (0.000) 0.009 (0.001) 0.009 (0.001) 0.995 (0.002)
Results
Mitochondrial divergence
At the mitochondrial 12S locus, 687bp were sequenced and aligned from 162 animals from 26Tigriopus
populations (2–10 animals per population). Interpopulation genetic distance ranged from 0.000 (0.000) to
0.323 (0.038) within T. californicus and from 0.313 (0.033) to 0.427 (0.046) between T. californicus
and T. japonicus (Table 3).
between
was equivalent to D
xy
, and was almost entirely attributable to inter-
population diversity—
within
was very low, butG
ST
was nearly fixed (Table 4). Intraspecific interpopulation
divergences ranged from 0 (0.000) to 0.133 (0.018) in populations from the US to central Baja California,
Mexico, but were universally higher when these populations were compared to the reproductively isolated
populations in Baja California (ECV , PA, and BA)—0.232 (0.029) to 0.323 (0.037). This differentiation
from the Baja populations is comparable to the differentiation between US populations of T. californicus
and T. japonicus (0.313 (0.033) to 0.351 (0.037)). Differentiation between the Baja populations and T.
japonicus is even higher, with genetic distances ranging from 0.395 (0.043) to 0.426 (0.047).
Populations exhibit a general isolation-by-distance pattern, but phylogenetic analysis also revealed the
presence of four regional clades within T. californicus (Figure 1). Clade A encompasses populations from
Washington to central California, Clade B from Malibu to San Diego in southern California, and Clade C
from northern to central Baja California. Clade D is composed of the reproductively isolated populations
from central Baja California, and branches separately from all otherT.californicus populations. The support
for a split between Clade D and Clades A, B, and C is strong, with a posterior probability of 1. Support for
the assignment of Clade C is also strong (posterior probability = 1). Posterior probabilities for Clade B (0.83)
and Clade A (0.61) are lower, but most internal branches within these clades are extremely well-supported
(posterior probabilities 0.95–1 in most cases). Genetic diversity within populations appears to be extremely
low (Table 4). Mitochondrial distance between Clades A, B, C was less than twice that between those clades
42
Table 5: Mean genetic distance (standard error) between regional clades. Above diagonal: nuclear loci;
below diagonal: mitochondrial 12S.
Clade A Clade B Clade C Clade D T.japonicus
Clade A 0.017 (0.001) 0.019 (0.001) 0.046 (0.003) 0.089 (0.007)
Clade B 0.099 (0.012) 0.014 (0.001) 0.039 (0.003) 0.082 (0.006)
Clade C 0.114 (0.014) 0.110 (0.013) 0.045 (0.004) 0.091 (0.007)
Clade D 0.262 (0.030) 0.262 (0.029) 0.292 (0.033) 0.097 (0.007)
T.japonicus 0.323 (0.033) 0.337 (0.033) 0.342 (0.035) 0.410 (0.043)
Table 6: Mean (standard error) mt12S interpopulation genetic distance (standard error) and geographic
distance between sampled populations within regional clades.
Clade Genetic distance Geographic distance (km) Ratio
A 0.033 (0.010) 634 (122) 5:2x10
5
B 0.058 (0.011) 108 (11) 5:4x10
4
C 0.032 (0.013) 206 (52) 1:6x10
4
D 0.054 (0.015) 174 (20) 3:1x10
4
and Clade D, though even larger genetic distances were measured between T. japonicus and Clades A–D
(Table 5). Genetic distance relative to geographic distance was reduced by an order of magnitude in Clade
A, relative to the Clades B, C, and D (Table 6). However, overall, genetic distance is strongly positively
correlated to geographic distance for mt12S (R
2
= 0.69 and 0.21, p < 0:001 for inter- and intraspecific
comparisons, respectively); for nuclear loci, genetic and geographic distance are positively correlated in
intraspecific comparisons, though the relationship is weak (R
2
= 0.019, p = 0.036) and are not correlated in
interspecific comparisons (R
2
= 0.019, p = 0.097; Figure 2).
Species-level boundaries inTigriopus
For the 18S ribosomal gene, 1664bp were sequenced and aligned for 241 individuals from 29Tigriopus
populations (2–10 animals per population). In most populations, intrapopulation variation was absent, and
the overall mean genetic distance was much lower than in mt12S (Table 4). However, despite the low
overall divergence, the populations are still highly differentiated, with G
ST
nearly fixed. These patterns
are also reflected in the phylogenetic tree produced by this locus (Figure 3). 18S sequences show little
divergence from Washington, USA to mid-Baja California, but populations south of the break in Baja form
43
Figure 2: Pairwise genetic divergence increases with geographic distance in both interspecific (left) and
interspecific (right) comparisons for the 12S mitochondrial locus (top) but not for the concatenated set of
nuclear loci (bottom). Line represents best-fit linear regression for genetic and geographic distance for inter-
and intraspecific comparisons. Comparisons are colored by type—red: interspecific; gray: intraspecifc.
44
0.4
PA-5
LB-6
PA-2
BB-1
SH-2
SCI-9
LC-7
LB-1
BB-2
SHC-4
SCI-8
LJ-4
BG-5
AB-3
RP-1
PA-3
CDA-4
PBN-10
FHL-5
SH-6
ECV-6
BA-6
ECV-8
RP-9
SHC-2
LMCS-7
PBN-3
PBN-1
SHC-6
SCI10
PA-7
SHC-8
FHL-4
SD-7
CDA-6
LB-4
RP-6
SCI-1
BA-9
FHL-8
PA-4
SD-9
SC-5
RP-7
AB-9
RP-5
ECV-9
LMC-8
BG-4
SH-7
BA-3
SD-5
LC-8
LC-2
CDA-10
BG-10
CDA-7
SHC-10
RP-2
PA-8
SC-10
LJ-2
ECV-10
SC-4
SH-3
PBN-6
SD-2
SHC-3
LC-9
LMCS-2
LMC-2
LC-10
SH-10
FHL-1
AB-10
AB-1
BR-11
SH-8
AB-4
CDA-9
SCI-6
SH-9
RP-8
BB-11
LB-2
CAT-4
CDA-3
BG-1
FHL-7
FR-10
LMC-4
ECV-1
LMCS-3
LMC-5
AB-6
PBN-7
LJ-8
SC-6
LJ-9
ECV-4
CDA-1
BA-5
SD-6
PA-9
BG-9
CAT-3
PBN-9
PA-6
AB-2
SCI-5
BG-3
BG-6
FHL-2
ECV10
LC-5
SD-1
LMC-9
PBN-5
LC-6
PA-1
BA-10
RP-3
SH-4
LB-8
RP-10
LMC-6
SC-3
LMC-7
BR-1
SC-2
BR-2
LJ-10
PBN-4
LC-1
LMCS-9
SHC-7
SHC-1
ECV-7
BA-7
PBN-2
AB-7
PBN-8
PA-10
CAT-9
SC-8
LJ-1
CDA-2
SH-5
LMC-1
SC-1
CAT-1
ECV-2
BR-12
CDA-5
BG-2
FHL-6
RP-4
SD-10
LMCS-8
BB-12
SD-4
LC-4
LMC-10
BA-4
SC-9
LJ-6
CDA-8
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.8155
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.9004
1
Clade A
Clade B
Clade C
Clade D
Tigriopus japonicus
Tigriopus brevicornis
Figure 3: Bayesian consensus phylogeny for nuclear 18S ribosomal RNA gene. Individuals are colored
by mitochondrial regional clade (T. californicus), as identified in Figure 1 or species (T. japonicus and T.
brevicornis). Branch numbers indicate posterior probabilities. MCMC was run for 2,500,000 generations.
45
0.007
BB-b
LMCS-b
PBN-a
ECV-b
BB-a
CDA-b
SCI-a
ECV-a
BA-a
SH-b
SHC-a
BR-a
BR-b
CAT-a
AB-a
SCI-b
BA-b
FHL-b
RP-b
LB-b
PA-a
CAT-b
LMCS-a
RP-a
PA-b
LC-a
SD-a
FHL-a
LJ-b
LC-b
SC-a
LB-a
SH-a
SC-b
LMC-b
SHC-b
CDA-a
PBN-b
AB-b
LJ-a
SD-b
LMC-a
1
1
1
1
0.52
1
0.99
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.83
1
1
1
1
0.96
1
1
1
1
1
1
1
1
1
1
1
1
1
0.89
1
1
1
1
1
1
1
1
0.55
0.73
1
1
0.88
1
1
1
1
1
1
1
1
1
0.99
1
1
0.83
1
1
1
1
1
1
1
1
Clade A
Clade B
Clade C
Clade D
Tigriopus japonicus
Figure 4: Bayesian consensus tree for concatenated set of 7 nuclear loci. Numbers on branches indicate
posterior probabilities. MCMC was run for 1,000,000 generations. Colors correspond to regional clades
identified in Figure 1.
a monophyletic sister group to all otherT.californicus populations (Figure 3). This split is highly supported
(posterior probability = 1) and clearly represents a species-level break. The inclusion of T. brevicornis
samples provides strong support for the position of T. japonicus as the sister species to T. californicus and
the undescribed Baja species (posterior probability = 1).
46
Table 7: Number (percentage) of total, variable, and parsimony–informative sites per locus
Locus Total Variable Parsimony–informative
mt12S 687 260 (38%) 249 (36%)
18S 1664 122 (7%) 117 (7%)
ITS1 552 392 (71%) 357 (65%)
5.8S 178 85 (48%) 29 (16%)
Tc22152 313 33 (11%) 21 (7%)
Tc25225 410 128 (31%) 78 (19%)
Tc30567 393 30 (8%) 28 (7%)
Tc28684 459 47 (10%) 45 (10%)
Total 4,656 835 (18%) 666 (14%)
47
A
B
C
D
0.02
CDA-a
LB-b
SCI-a
SC-a
CAT-b
LJ-a
LMC-a
BB-a
SD-b
PA-b
LC-b
LB-a
LC-a
LJ-b
SH-a
PA-a
SHC-b
FHL-a
AB-b
BR-a
ECV-a
BR-b
BA-a
LMCS-b
PBN-b
AB-a
CAT-a
PBN-a
SCI-b
RP-a
ECV-b
CDA-b
SHC-a
RP-b
LMC-b
FHL-b
SH-b
SD-a
SC-b
BA-b
LMCS-a
BB-b
0.8
0.98
0.99
0.94
0.52
0.99
0.89
0.02
CDA-a
LB-b
SCI-a
SC-a
CAT-b
LJ-a
LMC-a
BB-a
SD-b
PA-b
LC-b
LB-a
LC-a
LJ-b
SH-a
PA-a
SHC-b
FHL-a
AB-b
BR-a
ECV-a
BR-b
BA-a
LMCS-b
PBN-b
AB-a
CAT-a
PBN-a
SCI-b
RP-a
ECV-b
CDA-b
SHC-a
RP-b
LMC-b
FHL-b
SH-b
SD-a
SC-b
BA-b
LMCS-a
BB-b
0.02
CDA-a
LB-b
SCI-a
SC-a
CAT-b
LJ-a
LMC-a
BB-a
SD-b
PA-b
LC-b
LB-a
LC-a
LJ-b
SH-a
PA-a
SHC-b
FHL-a
AB-b
BR-a
ECV-a
BR-b
BA-a
LMCS-b
PBN-b
AB-a
CAT-a
PBN-a
SCI-b
RP-a
ECV-b
CDA-b
SHC-a
RP-b
LMC-b
FHL-b
SH-b
SD-a
SC-b
BA-b
LMCS-a
BB-b
0.99
0.02
CDA-a
LB-b
SCI-a
SC-a
CAT-b
LJ-a
LMC-a
BB-a
SD-b
PA-b
LC-b
LB-a
LC-a
LJ-b
SH-a
PA-a
SHC-b
FHL-a
AB-b
BR-a
ECV-a
BR-b
BA-a
LMCS-b
PBN-b
AB-a
CAT-a
PBN-a
SCI-b
RP-a
ECV-b
CDA-b
SHC-a
RP-b
LMC-b
FHL-b
SH-b
SD-a
SC-b
BA-b
LMCS-a
BB-b
0.02
CDA-a
LB-b
SCI-a
SC-a
CAT-b
LJ-a
LMC-a
BB-a
SD-b
PA-b
LC-b
LB-a
LC-a
LJ-b
SH-a
PA-a
SHC-b
FHL-a
AB-b
BR-a
ECV-a
BR-b
BA-a
LMCS-b
PBN-b
AB-a
CAT-a
PBN-a
SCI-b
RP-a
ECV-b
CDA-b
SHC-a
RP-b
LMC-b
FHL-b
SH-b
SD-a
SC-b
BA-b
LMCS-a
BB-b
0.02
CDA-a
LB-b
SCI-a
SC-a
CAT-b
LJ-a
LMC-a
BB-a
SD-b
PA-b
LC-b
LB-a
LC-a
LJ-b
SH-a
PA-a
SHC-b
FHL-a
AB-b
BR-a
ECV-a
BR-b
BA-a
LMCS-b
PBN-b
AB-a
CAT-a
PBN-a
SCI-b
RP-a
ECV-b
CDA-b
SHC-a
RP-b
LMC-b
FHL-b
SH-b
SD-a
SC-b
BA-b
LMCS-a
BB-b
FHL
SH
SHC
BB
SC
SD
BR, LJ
AB, RP
CAT
LC, (PD)
SCI
ECV
LMC,
LMCS
PA
BA
PBN
(MDH)
(PL)
(SUN)
(YH)
Figure 5: Bayesian consensus phylogeny of 1 mitochondrial and 7 nuclear loci and map of sampling loca-
tions. Numbers on branches indicate posterior probabilities (only values< 1 are shown). MCMC was run
for 2,500,000 generations. Colors correspond to regional clades identified in Figure 1. Populations on the
map designated in parentheses were sequenced at the mt12S locus but not nuclear loci.
Mito-nuclear phylogenetic concordance
In addition to mt12S and 18S, we sequenced the ITS1 region, the 5.8S ribosomal DNA gene, and 4
unlinked genes from the transcriptome in 2–10 animals per population, for a total of 4,656bp (Table 7). We
concatenated two sets of loci per population to generate multi-locus trees with and without the mt12S. We
recovered topologically similar trees with both approaches, both of which recover Clades A, B, C, and D
with strong support (Figures 4, 5). Genetic divergence is much higher in mt12S than in the concatenated
nuclear set (0.132 versus 0.027), and no polytomies occur in the tree derived from the combined nuclear and
mitochondrial sequences.
The proportion of variable and parsimony–informative sites varied widely among loci (Table 7). Max-
48
Figure 6: Consensus maximum parsimony tree for 1 mitochondrial and 7 nuclear loci (7 best trees were
found). Numbers on branches indicate support (only branches present in > 50% of trees are resolved).
Branches are colored by clade as in Figure 1.
imum parsiomony analysis for the concatenated set of loci found 7 best trees, and the consensus tree for
all loci is similar to the Bayesian analysis (Figure 6. The consistency index was 0.78 (0.75 for parsimony–
informative sites), and the retention index was 0.91 for all sites and for parsimony–informative sites.
Discussion
Regional clades withinT.californicus
Previous phylogeographic studies of T. californicus that sampled populations from most or all of the
species range also found that the mitochondrial COI locus provides evidence for the presence of regional
clades but failed to resolve the phylogenetic relationships between them (Edmands 2001, Peterson et al.
49
2013). We found evidence for the presence of 4 regional clades (Figures 1, 5) that are broadly similar to
those recovered from COI sequences in both the mitochondrial and nuclear genomes of T. californicus and
were able to place them phylogenetically. This is likely because substitutions in COI but not mt12S are
saturated within T. californicus. For instance, with COI, mean sequence divergence between the southern
California/northern Baja clade (Clade C in Peterson et al. 2013, analogous to our Clades B and C) and the
northernmost clade (Clade A in Peterson et al. 2013, analogous to our Clade A) is 0.269, more than double
the mt12S divergence between these clades (Table 5).
Interpopulation divergence is high but does not always scale with divergence
Populations of T. californicus show a general pattern of isolation-by-distance within and between re-
gional clades. These patterns have also been demonstrated inT.brevicornis andT.japonicus (Ki et al. 2009,
Handschumacher et al. 2010). As seen in at the mtCOI locus in both T. californicus and T. brevicornis
(Edmands 2001, Handschumacher et al. 2010), genetic differentiation between populations of T. californi-
cus from northern latitudes is reduced, a pattern which could be explained by a recent range expansion, or
by multiple range expansions and contractions coinciding with glacial retreats and expansions. Similarly,
we found shorter branch lengths within Clade A in the mt12S phylogeny and a polytomy within Clade B
with the concatenated nuclear loci set. Divergence within Clade A is comparable to that in Clade C, though
sampling sites are on average approximately 3x further apart in Clade A, and the ratio of average genetic
distance to average geographic distance in Clade A is an order of magnitude lower than in any of the other
regional clades (Table 6).
Genetic and geographic distances are correlated for mtCOI (Edmands 2001). We detected a similar
pattern for genetic and geographic distance at mt12S for both interspecific comparisons (Figure 2). However,
for nuclear loci we found only a very weak relationship for intraspecific comparisons and no relationship
for interspecific comparisons.
Speciation on the edge: Populations at the range edge are a separate species
Unusually extreme reproductive and genetic isolation in T. californicus populations in Baja California
were first documented two decades ago (Ganz and Burton 1995), but subsequent studies, relying on a single
mitochondrial marker, have failed to resolve Clade D’s placement relative to otherT.californicus populations
50
(Edmands 2001, Peterson et al. 2013). Our results, however, clearly demonstrate that this southern clade is
actually a separate species, rather than a semi–species or incipient species. This is indicated by the extreme
18S divergence and by the congruence of all of our nuclear and mitochondrial single-gene trees placing
Clade D as a sister group to all other populations ofT.californicus.
The designation of southern populations as a novel species is further supported by the reproductive isola-
tion documented betweenT.californicus and Clade D. All tested crosses between US-derivedT.californicus
and Clade D fail to produce offspring beyond the first generation, though some, but not all, crosses between
T.californicus populations from northern Baja and Clade D can produce second-generation offspring (Ganz
and Burton 1995, Peterson et al. 2013).
Previous studies have demonstrated that describedTigriopus species often exhibit complete reproductive
isolation, with geographically proximate populations generally able to interbreed while those from different
regions or oceans (e.g., eastern and western Pacific, or eastern Pacific and northern Atlantic) cannot. In
fact, this designation has been used to challenge the species-level status of a recently described Tigriopus
species from Thailand–though morphologically differentiated from T. japonicus, T. thailandensis produced
interfertile offspring when crossed to populations ofT.japonicus from Hong Kong, South Korea, and Japan
(Lee et al. 2012). However, some crosses produced only a limited number of generations, suggesting that
the speciation process is underway, though incomplete in these taxa.
The designation of Clade D as a separate species is remarkable given its geographic continuity with T.
californicus. The species break occurs between the LMCS and ECV populations, which are separated by
approximately 12 kilometers of coastline on a remote stretch of the Baja California coast (Peterson et al.
2013). Phylogenetic breaks occur for many species on the Baja California peninsula, but most of these
species are terrestrial, and it is unclear whether these patterns may be attributed to geological or ecological
processes, or stochastic genetic processes (Leach´ e et al. 2007). Phylogeographic breaks may also occur in
the absence of geographic or ecological barriers in species where genetic drift is strong and migration is low
(Irwin 2002). In the future, genomic studies may help to clarify when and how speciation occurred in these
populations.
51
Acknowledgments
Thanks to Wai Leong, Helen B. Foley, and Patrick Sun for help with collections; to Shiven Chaudry,
Mindy Guo, and Roxana Aslan for assisting in the sequencing work. This work was supported by a grant to
the National Science Foundation to SE.
52
References
Baker, C. F., 1912. First Annual Report of the Laguna Marine Laboratory at Laguna Beach, Orange County,
California, Chapter: Notes on crustacea of laguna beach . Claremont College.
Barreto, F. S., G. W. Moy, and R. S. Burton. 2011. Interpopulation patterns of divergence and selection
across the transcriptome of the copepodTigriopuscalifornicus. Molecular Ecology 20:560–72.
Burton, R. 1997. Genetic evidence for long term persistence of marine invertebrate populations in an
ephemeral environment. Evolution 51:993–998.
Burton, R., M. Feldman, and J. Curtsinger. 1979. Population genetics ofTigriopuscalifornicus (Copepoda,
Harpacticoida) 1. Population structure along the central California coast. Mar Ecol Prog Ser 1:29–39.
Burton, R., E. Metz, J. Flowers, and C. Willet. 2005. Unusual structure of ribosomal DNA in the copepod
Tigriopuscalifornicus: intergenic spacer sequences lack internal subrepeats. Gene 344:105–113.
Burton, R. S., R. J. Byrne, and P. D. Rawson. 2007. Three divergent mitochondrial genomes from California
populations of the copepodTigriopuscalifornicus. Gene 403:53–59.
Burton, R. S., C. K. Ellison, and J. S. Harrison. 2006. The sorry state of F-2 hybrids: Consequences of rapid
mitochondrial DNA evolution in allopatric populations. American Naturalist 168:S14–S24.
Burton, R. S., and M. W. Feldman. 1981. Population genetics of Tigriopus californicus: II. Differentiation
among neighboring populations. Evolution 35:1192–1205.
Darriba, D., G. Taboada, R. Doallo, and D. Posada. 2012. jModelTest2: More models, new heuristics and
parallel computing. Nature Methods 9:772.
Dethier, M. N. 1980. Tidepools as refuges: predation and the limits of the harpacticoid copepod Tigriopus
californicus (Baker). Journal of Experimental Marine Biology and Ecology 42:99–111.
Edgar, R. C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput.
Nucleic Acids Research 32:1792–1797.
Edmands, S. 2001. Phylogeography of the intertidal copepod Tigriopus californicus reveals substantially
reduced population differentiation at northern latitudes. Molecular Ecology 10:1743–1750.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, G. W. Moy, R. S. Burton, and S. Edmands. 2011.
A gene-based SNP resource and linkage map for the copepod Tigriopus californicus. BMC Genomics
12:568.
Ganz, H., and R. Burton. 1995. Genetic differentiation and reproductive incompatibility among Baja
California populations of the copepodTigriopuscalifornicus. Marine Biology 123:821–827.
Guindon, S., and O. Gascuel. 2003. A simple, fast and accurate method to estimate large phylogenies by
maximum-likelihood. Systematic Biology 52:696–704.
Handschumacher, L., M. B. Steinarsdottir, S. Edmands, and A. Ingolfsson. 2010. Phylogeography of the
rock-pool copepodTigriopusbrevicornis (Harpacticoida) in the northern North Atlantic, and its relation-
ship to other species of the genus. Marine Biology 157:1357–1366.
Hulsenbeck, J. P., F. Ronquist, R. Nielsen, and J. P. Bollback. 2001. Bayesian inference of phylogeny and
its impact on evolutionary biology. Science 294:2310–2314.
53
Huys, R., J. Mackenzie-Dodds, and J. Llewellyn-Hughes. 2009. Cancrincolidae (Copepoda, Harpacticoida)
associated with land crabs: A semiterrestrial leaf of the ameirid tree. Molecular Phylogenetics and Evo-
lution 51:143–156.
Irwin, D. E. 2002. Phylogeographic breaks without geographic barriers to gene flow. Evolution 56:2283–
2394.
Ki, J.-S., K.-W. Lee, H. G. Park, S. Chullasorn, H.-U. Dahms, and J.-S. Lee. 2009. Phylogeography of the
copepodTigriopusjaponicus along the Northwest Pacific rim. J Plankton Res 31:209–221.
Larkin, M., G. Blackshields, N. Brown, R. Chenna, P. McGettigan, H. McWilliam, F. Valentin, I. Wallace,
A. Wilm, R. Lopez, J. Thompson, T. Gibson, and D. Higgins. 2007. Clustal W and Clustal X version 2.0.
Bioinformatics 23:2947–2948.
Leach´ e, A. D., S. C. Crews, and M. J. Hickerson. 2007. Two waves of diversification in mammals and
reptiles of Baja California revealed by hierarchical Bayesian analysis. Biology Letters 3:646–650.
Lee, K.-W., H.-U. Dahms, and J.-S. Lee. 2012. Multigenerational life history traits of the copepodTigriopus
from the Northwest Pacific rim in crossbreeding experiments. Journal of Experimental Marine Biology
and Ecology 436:56–62.
Peterson, D., K. Kubow, M. Connolly, L. Kaplan, M. Wetkowski, W. Leong, B. Phillips, and S. Edmands.
2013. Reproductive and phylogenetic divergence of tidepool copepod populations across a narrow geo-
graphical boundary in Baja California. Journal of Biogeography 40:1664–1675.
Pritchard, V ., L. Dimond, J. Harrison, C. Velasquez, J. Zieba, R. Burton, and S. Edmands. 2011. Interpopu-
lation hybridization results in widespread viability selection across the genome inTigriopuscalifornicus.
BMC Genetics 12:54.
Pritchard, V ., and S. Edmands. 2012. The genomic trajectory of hybrid swarms: Outcomes of repeated
crosses between populations ofTigriopuscalifornicus. Evolution 67:774–791.
R Development Core Team, 2011. R: a language and environment for statistical computing.
Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and S. Kumar. 2011. MEGA5: Molecular evo-
lutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony
methods. Molecular Biology and Evolution 28:2731–2739.
Vavrek, M. J. Palaeontologia Electronica. fossil: Palaeoecological and palaeogeographical analysis tools.
2011 14:1.
54
Chapter 3: Hybrid breakdown, transgressive segregation, and sex–specific
effects of hybridization in non-recombinant backcrosses in the copepod
Tigriopus californicus
Barret C. Phillips and Suzanne Edmands
Abstract
Postzygotic reproductive isolation may accumulate gradually over time, but not necessarily in the same
way between each diverging set of populations. Traditionally, investigations of the genetic basis of reproduc-
tive isolation have focused on a limited number of crosses. Here, we utilize non-recombinant backcrosses
between populations of the marine copepodTigriopuscalifornicus to examine the genetic incompatibilities
underlying hybrid breakdown. Fitness measures included hatch number, survivorship, and family sex ratio,
all of which varied dramatically between crosses, and even between different backcrosses crosses for the
same population pair. In our most divergent cross—which spanned the majority of the species range—near-
complete mortality occurred at or shortly after the time when larvae hatched from eggs, but only one cross
direction, indicating that the asymmetry of incompatibilities may not level out over time. We genotyped the
progeny of 7 crosses at population-diagnostic loci, and we found that in different crosses, incompatibilities
are underlain by both shared and unique genetic combinations. In the two crosses where we genotyped both
sexes, we detected both common and unique loci associated with sex. These loci may be involved in poly-
genic sex determination inT.californicus, although these patterns could also be due to sex-specific viability
differences. Male hybrids consistently fared worse than females in our crosses—female-biased sex ratios
suggest sex-specific mortality, and more segregation distortion was detected in males than in females. Sev-
eral crosses exhibited transgressive morphological phenotypes, but we found little support for morphometric
QTL. Though we detected the most genetic incompatibilities in crosses between populations of intermediate
divergence, we found one of the most extreme patterns of hybrid breakdown documented to date within T.
californicus in one direction of our most divergent cross. This study supports the idea that some, but not
all, hybrid incompatibility may be predictable based on knowledge of phylogenetic relationships, and that
polygenic sex determination loci may have a common genetic basis in multiple populations.
55
Introduction
Though species are discrete entities, the process by which they arise is continuous (Hendry et al. 2009,
Shaw and Mullen 2014). As lineages diverge, they may accumulate behavioral, ecological, morphological,
and genetic differences that eventually create reproductive isolation. While prezygotic isolation may emerge
as a direct consequence of the former three types of changes (e.g., changes in mating preferences, mating
season, genital morphology), the evolution of postzygotic isolation poses a conundrum: how can a mutation
that makes an organism incompatible with others of its type be maintained? Postzygotic reproductive isola-
tion, defined as reduced fertility and/or viability of hybrid offspring, must instead emerge as a byproduct of
evolutionary divergence, for by definition, any factor that prevents an organism from breeding with others
will reduce that organism’s fitness relative to those that are unimpeded. Dobzhansky (Dobzhansky 1937)
and Muller (Muller 1940) worked around this problem by proposing that in fact postzygotic isolation must
be caused by multi-locus interactions. Dobzhansky-Muller incompatibilities (DMIs) are thus epistatic (gene
x gene) interactions where novel allelic combinations fail to function properly. This could be due to novel
allele(s) arising in each of two diverging lineages or to the sequential evolution of novel alleles within a
single lineage, while each remains unchanged in the other lineage.
In species with heteromorphic sex chromosomes, DMIs map disproportionately to the X (or Z) chromo-
some (large X/Z effect), and the hemizygous sex generally suffers more hybrid breakdown (Haldane’s rule)
(Haldane 1922, Presgraves 2008, Ellegren 2009). Both of these so-called “rules” of speciation are functions
of the peculiar properties of sex chromosomes, which have lower effective population sizes than autosomes
and expose recessives in the hemizygous sex (Schilthuizen et al. 2011). Most studies of the genetic basis
of postzygotic reproductive isolation have been conducted in species with sex chromosomes, and most of
the “speciation genes” identified to date are located on the X chromosome (Presgraves 2010). However, sex
chromosomes are highly evolutionarily labile, and many species with separate sexes actually lack chromo-
somal sex determination (Mank and Avise 2009, The Tree of Sex Consortium 2014). Thus, in the absence
of sex chromosomes, speciation may proceed at a slower rate (Phillips and Edmands 2012, Lima 2014),
and postzygotic isolation is therefore most likely underlain by either nuclear-nuclear or cyto-nuclear barri-
ers. Like sex chromosomes, mitochondrial genomes have lower effective population sizes than the nuclear
genome, and animal mitochondrial genomes often evolve faster than nuclear genomes . Additionally, many
proteins and RNA molecules that function in the mitochondria are encoded by nuclear genes, and there-
56
fore, disruption of mitochondrial function in hybrids could play an especially important role in postzygotic
isolation, especially in the absence of sex chromosomes (Burton et al. 2006, Burton and Barreto 2012).
The intertidal copepod Tigriopus californicus provides an ideal system in which to examine how re-
productive isolation evolves because it consists of many allopatric populations spanning the Pacific coast
of North America (Baja California Sur, Mexico to Alaska, USA). Despite high genetic divergence between
many populations, most are interfertile when crossed in the laboratory (Edmands 2001, Peterson et al. 2013).
Populations may be maintained indefinitely in the laboratory, and generation time is 23 days when cultured
at 20

C (Burton 1987). Females mate only once and use stored sperm to fertilize multiple clutches, fa-
cilitating the production of isofemale lines (Vittor 1971, Burton 1985). Sex determination in Tigriopus is
genetic, but no sex chromosomes (homomorphic or heteromorphic) are present; rather, under polygenic sex
determination, multiple, mostly unlinked loci contribute to sex (Ar-Rushdi 1958, Lazzaretto and Libertini
1985, Alexander et al. 2015). Meiotic recombination is absent in females, facilitating the production of non-
recombinant backcrosses, in which entire chromosomes are inherited intact (Burton et al. 1981, Harrison
and Edmands 2006).
The evolution of reproductive isolation has been studied extensively inT.californicus. Intraspecifically,
prezygotic isolation and hybrid sterility are generally absent (Palmer and Edmands 2000, Willett 2008a),
and F
1
hybrids generally exhibit comparable or heterotic fitness relative to their parents (Edmands 1999).
However, many crosses demonstrate hybrid breakdown in later generations, though this can sometimes be
recovered through backcrosses or long-term admixture (Edmands 1999, Burton et al. 2006, Ellison and Bur-
ton 2008b, Hwang et al. 2011, 2012). However, most studies investigating patterns of postzygotic isolation
in T. californicus to date have (i) measured hybrid fitness without examining the genetic basis of incom-
patibility (e.g., Edmands 1999, Ellison and Burton 2008b); (ii) focused on mapping incompatibilities in a
smaller set of crosses (e.g., Edmands et al. 2009, Pritchard et al. 2011, Foley et al. 2013); (iii) mapped
incompatibilities in multiple, but non-reciprocal crosses (e.g., Harrison and Edmands 2006); or (iv) focused
on the interactions of candidate incompatibility genes (primarily cytonuclear interactions, e.g., Willett and
Burton 2001, 2003, Willett 2006, Willett and Berkowitz 2007, Willett 2008b, Ellison and Burton 2008a,
2010, Willett 2011).
In this study, we performed non-recombinant backcrosses between 4 populations ofT.californicus span-
ning a wide range of geographic and genetic divergence, in which reciprocal F
1
s were backcrossed to their
paternal population, which maximizes the potential for cytonuclear mismatches that could lead to hybrid
57
Table 1: Populations, isofemale lines, and crosses
Date of Date line Crossed to males from
Population Latitude Longitude Line collection established AB BR FHL SD
Abalone Cove, CA (AB) 33:73

N 118:37

W AB-8 6/2011 10/2011 ! ! !
Bird Rock, CA (BR) 32:82

N 117:27

W BR-4 9/2012 11/2012 ! ! !
Friday Harbor, WA (FHL) 48:55

N 123

W FHL-8 9/2011 11/2012 ! !
San Diego, CA (SD) 32:75

N 117:26

W SD-202 8/2006 8/2006 ! ! ! !
dysfunction (these crosses are hereafter termed reciprocal). We examined 3 metrics of hybrid fitness–hatch
number, survivorship, and family sex ratio–in addition to morphometric measurements, and we conducted
population-diagnostic genotyping to search for patterns of segregation distortion and two-locus epistatic
genetic incompatibilities. We thus examine the relationship between genetic and geographic divergence
and reproductive isolation, and whether reproductive isolation accumulates similarly within crosses on mis-
matched cytonuclear backgrounds.
Materials and methods
Interpopulation crosses
Isofemale lines established from single ovigerous females from 5 populations (Figure 1) were inbred
for at least six months before crosses commenced (Table 1). Cultures were initially maintained in 60mL
Petri dishes with 3x 250um filtered seawater (from Wrigley Marine Science Center, Catalina Island, CA)
and 0.1g/L each Spirulina (Nutraceutical Sciences Institute) and Tetramin (Tetra) fish food and cultured at
15

C. Isofemale lines were subsequently transferred to 1L beakers, where they were cultured with a 12/12
light/dark cycle at 20

C in 3x 250um filtered seawater, 0.1g/L each Spirulina and Tetramin, and 100mL
each live Platymonas and Isochrysis algal cultures (Carolina Biological), with water changes as needed. All
crosses were performed in 15mL Petri dishes with autoclaved, 3x 250um filtered seawater, 0.1g/L Spirulina
and Tetramin, with weekly feeding, rehydration, and dish randomization within a single incubator.
Females mate once and store sperm that they use to fertilize up to twelve clutches (Vittor 1971). Males
have chelate antennules with which they grasp females for as much as a week before they are mature enough
to be inseminated (Burton 1985). Virgin females and mature males are thus easily obtained for crossing stud-
ies by extracting these clasped pairs from culture. For F
1
crosses, clasped pairs were removed from culture,
58
split apart under a dissecting microscope, and randomly allocated to either hybrid or parental line dishes (one
male and one immature female per dish). Dishes were checked periodically, and upon observation of an egg
sac on the female, the male was removed from the dish. After at least two clutches had hatched, parental
females were transferred to a fresh dish to allow further production of offspring while maintaining discrete
age classes (clutches usually hatch 3-5 days apart). Offspring were monitored for pair formation, and when
F1 pairs were observed, they were removed, split apart, and the females placed into fresh dishes with a male
from an F1 dish of the appropriate parental line to produce non-recombinant backcross offspring. Parental
lines were crossed by pair-splitting and random allocation to avoid inbreeding while providing generation-
appropriate controls. Males were removed from backcross dishes when egg sacs were observed on females.
Upon hatch of backcross offspring (or generation-matched control cross offspring), females were moved
to a fresh dish (first clutch) or discarded (second clutch). Freshly hatched nauplii were counted and pipetted
to fresh dishes in sets of 10 to prevent crowding. After 28 days, survivors were counted and sexed. Counts
from multiple dishes originating from a single clutch were pooled, giving a single count per clutch per
family, with up to two replicate measures per family. Males were photographed and frozen for later genetic
analysis. Hatch number, survivorship, and family sex ratios were all calculated for the non-recombinant
backcross generation—that is, from the offspring produced by F
1
females. Between 9 and 111 families per
cross or parental line were assayed.
Morphometrics
Live males were placed into a small drop of seawater on a slide and photographed on a Leica MZ12
stereomicroscope at 32x magnification. Images were acquired using SPOT imaging software with a SPOT
Idea camera (Diagnostic Instruments, Inc.). Image processing was done in ImageJ, where cephalothorax
length and width, urosome length and width, and left and right clasper length were measured using the
ObjectJ module (Schneider et al. 2012, https://sils.fnwi.uva.nl/bcb/objectj/index.html).
Genotyping
Because meiotic recombination is absent in T. californicus females (Ar-Rushdi 1963, Burton and Feld-
man 1981, Harrison and Edmands 2006, Foley et al. 2013), maternal chromosomes are transmitted intact
to offspring. Thus, in non-recombinant backcrosses, where F
1
females are mated to parental population
59
males, only one population-diagnostic marker is required per chromosome to assess the genetic composi-
tion of offspring. Diagnostic SNPs for all crosses were developed using the T. californicus draft genome
(Barreto et al. in prep.) and next-generation sequencing reads from all crossed populations (Barreto et al.
in prep.; Chapter 4). A subset of scaffolds in the draft reference genome (comprising approximately half of
the assembled bases) have been assigned to chromosomes by BLASTing a set of 190 population–diagnostic
SNP loci previously mapped in F
2
larvae (Foley et al. 2011). Read sets from all populations were input into
DiscoSNP++, a reference-free, kmer-based SNP discovery program (Uricaru et al. 2015), with a kmer size
of 31 and only one SNP allowed per graph. The resulting contigs were mapped to the assembled 12 chro-
mosomes with bwa (Li and Durbin 2009) and filtered for invariant contigs longer than 100bp on each side of
the SNP. Raw read sets from all five populations were then mapped back to a subset of candidate SNPs and
examined visually in IGV (Robinson et al. 2011) to ensure that the SNPs found were fixed, diagnostic, and
surrounded by invariant sequence. Primer design and SNP pool assignment were performed at the Roswell
Park Cancer Institute (Table 2).
Previously frozen samples were extracted by adding 50uL lysis buffer (10mM Tris pH 8.3, 50mM KCl,
0:05% Tween 20, 200g/mL proteinase K) to each individually frozen copepod and incubated at 65

C for 1
hour, followed by 15 minutes at100

C. A 20uL aliquot of each extracted sample was evaporated by incubat-
ing samples at65

C in a thermal cycler until completely dried (2-4 hours per plate) and sent to Roswell Park
Cancer Center, where they were genotyped in Sequenom MassArray iPlex Gold Assays. Genotyping was
performed using a set of 30 population-diagnostic SNPs with at least one diagnostic locus per chromosome
and mitochondrial genome per cross (Tables 2, 3). F
1
and pure parental line individuals were initially
genotyped to validate the diagnostic capacity of each SNP, and parental line animals were genotyped on the
same chips as the backcross animals.
60
FHL
AB
BR
SD
A B
C
Pop A Pop B Pop B Pop B
F
1
Pop B
NRBC
Figure 1: Phylogeny and map of populations used in the crosses. (A) Population-consensus phylogeny from
1 mitochondrial and 7 nuclear loci. Data from Chapter 2. (B) Sampling sites are indicated by stars. Colors
as in A. (C) Schematic of cross design. F
1
females were crossed to generation-controlled males from the
paternal population to produce non-recombinant backcross (NRBC) offspring.
61
Table 2: Population-diagnostic SNPs genotyped
Allele
Chromosome SNP ID AB BR FHL SD PCR Primer 1 PCR Primer 2 Extension Primer Pool
1 A586812 C T C C ACGTTGGATGAAATTGGTGGGACCGAAGAG ACGTTGGATGTCCCTTTTGCGATGGCATTG GCAAACTTCTTGCATAAGC 1
1 A30250 C T C T ACGTTGGATGGAACCTGGTGAACTGAAGAG ACGTTGGATGTGTCGAAGACTGTGACAACG GAACTGAAGAGATGGGG 3
2 A178530 T C T C ACGTTGGATGTGCAAGGATAAAGTGGTGGC ACGTTGGATGCCAGAAACCCTTTCAGTTCG ACTGTTTGAAGGTAGAAGG 1
2 A368970 C T C C ACGTTGGATGAGTACACCTCCCAATGTCTG ACGTTGGATGTGTCTTTGTTGACCGTGGTG GAACTTCTTGATGGAGAACTG 2
3 A110311 A C A C ACGTTGGATGACCCGATCAAATATCGTCCG ACGTTGGATGGTCTCCCAGAAATTGAAGCC TCGTCCGTCGTTCTCT 2
3 A522015 A G A A ACGTTGGATGGGAGGTTTCTTGCTGTGAAC ACGTTGGATGTCAAGGCACGGAACCAAATG AGAACGAAACTACTACAATAC 3
4 A102107 A G A G ACGTTGGATGAATGGTTTCTCTCCAGTGTG ACGTTGGATGAGCGATGACAAATGCTGTCC TTTTGCTGTTGATGAGGAGCCC 2
4* A73407* C A C C ACGTTGGATGACTATGGAAGACAGACACGC ACGTTGGATGCCAAACCTTGAGTTCCTTTC GTTCCGGAGGCCATTCT 3
5 A295659 T C T C ACGTTGGATGGACGAGACTCTTGTCTCTAC ACGTTGGATGGCCTTCATTGGCTTTCTACC CAACCAATCTACCAGCCT 1
5 A358372 C C C C ACGTTGGATGCCAAACTGACTTGGAAGCTC ACGTTGGATGTCACATGAGTCAATGGAGGG GGTGGCGGCCAAAATTGACATGACGTT 1
5 A575811 T C T T ACGTTGGATGCAATGGGAATGCGATCCTTG ACGTTGGATGGCCCAACATGTTCCAACAAC TTGACTCCCCTTTGTTCCTCTAACTC 1
6 A251428 G G G T ACGTTGGATGCCATCCTCTGTGGGTTCCAA ACGTTGGATGTTTTCTCCTCCATCGAGGTG GTGGGTTCCAATCACAT 1
6 A395869 C C T C ACGTTGGATGAAGAAGCACGAAACCTGCTC ACGTTGGATGAAGTATCCCTCTTCTTCGGC ATCCATCGGCCATCGTCGACTC 1
6 A166226 G A G G ACGTTGGATGAATCCAGAAGTTGGGCTTCG ACGTTGGATGAAAAGGGACCCTCGTATCAC TCGTAAAACTAACGGCTGG 2
7 A438880 G A G A ACGTTGGATGATTTACTACACCCGATCCCC ACGTTGGATGGAATTTCCCGTTTTACTCGC CTCCACTTTGTGACCA 1
7 A500766 A G A A ACGTTGGATGGCCGAGATCATGAGAATGAC ACGTTGGATGCTTCAATGCACAATGACGGC ATGACATCCTTTGAACATCGACC 1
8 A338005 C T C C ACGTTGGATGCATGTGTGGCACTCACTTAC ACGTTGGATGTGGGAAAACCAACATGCCAC TACACAAAGTGACAACACTCGTA 2
8 A460601 T T T C ACGTTGGATGTTTTTAGCACACTGCCAGCC ACGTTGGATGGCTCTTAGGCTTCTCTACTG TGGCTCTTAATGGAATGAATCTTGGC 2
9 A331643 T A T A ACGTTGGATGCTTGATCGCTTAGTCTTTGG ACGTTGGATGATACAAAGACTCCTGGGGAC AAATATGGGGGTAATTTTCTGCTCTCTC 1
9 A377671 G G G A ACGTTGGATGACACTACTCGAAGTGTGGAC ACGTTGGATGTCGCAGCCAAGATGATGAAG CATGGCACTGACTCC 1
10 A263251 G A G G ACGTTGGATGCATGAGCAACCTGAGAATCG ACGTTGGATGTTCCCCTTGTGCTTTCGAAG GCGTTGGCACCTTATCT 1
10 A542807 C G C G ACGTTGGATGAACTGGACTTCTTGAAGGGC ACGTTGGATGCTCCACCAAGAAGTAAGAGC TTGAAGGGCTCTTAATGG 1
11 A296010 T A T A ACGTTGGATGTTTTCTGCCATGTCAGGTCC ACGTTGGATGACCTCGACTCATCTTGACAG GGAACAAACACGCCT 2
11 A435999 A G A A ACGTTGGATGAGACCTCTGAGGATGAGAAG ACGTTGGATGCCAACAATTGGCCTGTTGTC ATTGTGGAGATCTCGC 2
12 A141149 A A G A ACGTTGGATGACTCCGCTTGTTCCAACTTG ACGTTGGATGCCTTTGAGCAACTGTGCATC CGCTTGTTCCAACTTGAAGATGAACTT 1
12 A449207 G A G G ACGTTGGATGAGGCATGAAACCGAAACACG ACGTTGGATGATTGGGTCATAACGGGTCAG GGGCAGGATTAGGTAATAACCACG 1
12 A355994 A T T T ACGTTGGATGAACTTTTGGAGTTCCTCCCC ACGTTGGATGCCAGGTTTGTGAATAAGTCG TCCTCCCCCGCTCCC 2
12 A273470 T T G T ACGTTGGATGGGTGGGTTGGATGATCCTTC ACGTTGGATGAGCGTGAACTACAGGAGAAG TTCGCCACTTTCCCG 3
Mt Mito 9128 C A G A ACGTTGGATGCCTGAATGACCTCGATGTTG ACGTTGGATGCGTCGATCTTAACTCAAATC CCTCGATGTTGAATTAAGAAACCTTCTA 2
Mt Mito 9163 A A A G ACGTTGGATGCGTCGATCTTAACTCAAATC ACGTTGGATGCCTGAATGACCTCGATGTTG TTAACTCAAATCATGTAAGAAAATA 3
*A73407 is located on scaffold 22 in theT.californicus draft genome, which contains marker Tc39888 from (Foley et al. 2011). The scaffold containing this marker was mapped to chromosome 4 in the original SD x SC cross (Foley et al. 2011),
but mapped to chromosome 1 in both the AB x BR and BR x SD crosses. We have treated it as assigned to chromosome 1 in our analyses.
62
Table 3: Diagnostic loci per chromosome per cross
Chromosome
Cross 1 2 3 4 5 6 7 8 9 10 11 12 Mt
BR x SD A586812 A368970 A522015 – A575811 A251428 A500766 A338005 A377671 A263251 A435999 A449207 Mt 9163
A73407
AB x BR A30250 A178530 A110311 A102107 A295659 A166226 A438880 A338005 A331643 A263251 A296010 A355994 Mt 9128
A586812 A368970 A522015 A500766
A73407
AB x SD A30250 A178530 A110311 A102107 A295659 A166226 A438880 A460601 A331643 A542807 A296010 A355994 Mt 9128
A251428 A377671 Mt 9163
SD-B x FHL A30250 A178530 A110311 A102107 A295659 A395869 A438880 A460601 A331643 A542807 A296010 A141149 Mt 9128
A166226 A377671 A273470 Mt 9163
A251428
63
Statistical analysis
Fitness and morphometric analysis
All statistical analyses were conducted in R 3.2.2 (R Core Team 2015). Parental populations were
assessed for homogeneity of variance using Levene’s test, as implemented in the car package (Fox and
Weisberg 2011). Differences between populations in survivorship, family sex ratios, and morphometric
measures were tested using independent two-sample t-tests with an assumption of equal variances and were
corrected for multiple testing using and false discovery rate (FDR) correction with an FDR of 5% (Benjamini
and Hochberg 1995). ANOV A and planned contrast tests were used to compare hybrids to the three-quarter
parent, which is the expected genetic makeup of a backcross. Because of the overlapping nature of the cross
design, all crosses and parental populations were compared in a single ANOV A calculation, followed by
contrast tests using the gmodels package for each cross conducted (Warnes et al. 2015).
Genotyping
Males from 7 crosses and females from 2 crosses were genotyped at both informative and uninformative
loci, but only informative loci were retained for analysis. Parental animals per cross were genotyped in each
384 animal-set. The genotype calls of these parental animals were compared to the expected genotypes of
parental animals, based on the next-generation sequences used in assay design. This comparison was used
to assign each parental animal to membership in 1 of the 4 populations used in these crosses. For each
pairwise interpopulation cross, SNPs were considered informative (diagnostic) when alternate alleles were
called in the two parental populations.
Linkage mapping was performed separately for each cross direction in JoinMap 4.1 (Ooijen 2006). Both
maximum likelihood and regression mapping (Kosambi function) were conducted and yielded concordant
maps. Markers found to be discordant with their originally assigned chromosome were subsequently as-
sessed based on their new map assignment.
Statistical analyses of segregation distortion and epistatic incompatibilities were conducted in R. For
each cross, each informative nuclear locus was examined for deviations from an expected 1:1 heterozygote:
paternal homozygote Mendelian ratio (segregation distortion) using chi-square tests, and (FDR) correction
was conducted for each cross with an FDR of 5%. Chi-square tests were also used to look for pairwise
epistatic interactions. Observed two-locus genotype frequencies for all unlinked loci were compared to
64
expected two-locus genotypes that were calculated from single-locus allele frequencies, and FDR correction
was applied.
Replacement of SC line by SD animals
Two crosses in this study (SC x BR and SC x FHL) utilized the SC-106 line that was used to create the
T.californicus linkage map and to map SC-SD incompatibilities (Foley et al. 2011, 2013). However, initial
genotyping (48-animal validation plate) revealed that at all SC-SD diagnostic loci, the SC line contained
only SD alleles, suggesting that SD animals had been mislabeled as SC. Levene’s tests for homogeneity of
variance and independent two-sample t-tests revealed no significant differences between the two lines for
survivorship, family sex ratio, or any morphometric feature, nor were any differences detected for the two
sets of hybrid crosses ((BR x SC) x SC versus (BR x SD) x SD and (SC x BR) x BR versus (SD x BR) x
BR), excepting left and right clasper lengths. However, our genetic assay did not include any intrapopulation
SNPs, and thus we cannot say with confidence that the SC-106 line has been replaced by animals from the
SD-202 line, so we have not combined the BR x SD and BR x SC crosses, or the fitness data from the SC
and SD parental lines. Fitness and morphometric data from crosses that utilized this contaminated line are
labeled “SD-B”, while those that used the original SD-202 line are labeled “SD-A”. Only SD-A line animals
were crossed to AB, only animals from the SD-B line were crossed to FHL, and only BR crosses to the
SD-A line were genotyped. Consequently, alleles matching either the SD-A or the SD-B parental line are
all identified simply as SD.
QTL mapping
Quantitative trait locus mapping was conducted with the R package qtl (Broman et al. 2003). QTL map-
ping was performed using Haley-Knott regressions under a single-locus model with the function scanone
and a two-locus model with the function scantwo, with significance thresholds for each trait and cross set
using 200,000 and 1,000 permutations, respectively. Additionally, for each trait in each cross, the stepwise-
qtl function, with a maximum of 6 possible QTL, was implemented to determine the best model via stepwise
model selection, with penalties based on the scantwo permutations. In the two crosses where both males
and females were genotyped, QTL mapping of sex was performed as in morphometric measurements, but
under a binary, rather than a normal, model, and with 12 possible QTL during stepwise model selection.
65
Table 4: Differences between parental lines for survivorship, sex ratio, and morphometric measures. Devia-
tions from equal values were calculated using independent two-sample t-tests (p-values are shown). Bolded
values remain significant after FDR correction for multiple tests. For significant differences, the population
with larger hatch number, higher survivorship, or more male–biased family sex ratio is in parentheses.
Population 1 Population 2 Hatch number Survivorship Family sex ratio
AB BR <0.001 (AB) <0.001 (AB) 0.002 (BR)
SD-A 0.842 0.006 (AB) 0.064
BR SD-B 0.327 0.482 0.174
SD-A 0.003 (SD-A) 0.832 0.895
FHL SD-B 0.463 0.173 0.261
Results
Fitness measures
Differences between parental lines
Generation-controlled parental lines were assayed for hatch number, survivorship, and family sex ratio,
but parental lines were compared only to the other lines to which they were crossed. Relative to the AB and
SD lines, significantly fewer BR animals hatched per clutch. Survivorship in AB was significantly higher
than in either BR or SD, and family sex ratios were significantly more female-biased in AB than in BR. No
other significant differences were seen in hatching number, survivorship, or family sex ratios between the
crossed parental lines (Table 4).
Hybrid fitness
Hatch number was significantly increased relative to the 3/4-parent expectation in 1 cross, (SD x BR) x
BR, and significantly decreased in 4 crosses: (AB x BR) x BR, (AB x SD) x SD, (SD x AB) x AB, and, most
dramatically, (FHL x SD-B) x SD-B (Table 5, Figure 2). The hatching success in the latter cross was so low
(mean number of animals per clutch = 1.84) that this cross direction was not included in further assays.
Survivorship demonstrated large deviations from the 3/4-parent in most crosses (31% in (BR x SD-
B) x SD-B to +21% in (SD-B x FHL) x FHL), but these deviations were only statistically significant in
the (BR x SD-B) x SD-B cross (Table 2). For 3 out of the 4 crosses with reciprocal data, the direction of
deviation from the 3/4-parent differed (Figure 2). In the fourth reciprocal cross, AB x SD, however, both
66
-100
-75
-50
-25
0
25
(BR x SD-A) x SD-A
(SD-A x BR) x BR
(BR x SD-B) x SD-B
(SD-B x BR) x BR
(AB x BR) x BR
(BR x AB) x AB
(AB x SD-A) x SD-A
(SD-A x AB) x AB
(SD-B x FHL) x FHL
(FHL x SD-B) x SD-B
% Devia(on from 3/4-parent
Hatch number
***
*
***
***
***
-35
-20
-5
10
25
(BR x SD-A) x SD-A
(SD-A x BR) x BR
(BR x SD-B) x SD-B
(SD-B x BR) x BR
(AB x BR) x BR
(BR x AB) x AB
(AB x SD-A) x SD-A
(SD-A x AB) x AB
(SD-B x FHL) x FHL
Survivorship
*
-40
-20
0
20
(BR x SD-A) x SD-A
(SD-A x BR) x BR
(BR x SD-B) x SD-B
(SD-B x BR) x BR
(AB x BR) x BR
(BR x AB) x AB
(AB x SD-A) x SD-A
(SD-A x AB) x AB
(SD-B x FHL) x FHL
Family sex ra(o
AB
BR
FHL
SD-B
SD-A
***
***
*
A B C
Figure 2: Comparison of backcross hatch number (A), survivorship (B) and family sex ratio (C) to 3/4-
parent expectations. Percent difference between observed and additive expected values are shown. For
hatch number, positive values indicate more backcross offspring per clutch, and negative values indicate
fewer backcross offspring per clutch. For survivorship, positive values indicate higher-than-expected sur-
vivorship, and negative values indicate lower-than-expected survivorship. For family sex ratio, positive
values indicate more males than expected (suggestive of female-biased mortality), and negative values in-
dicate more females than expected (suggestive of male-biased mortality). Crosses are colored according to
their expected genetic composition (red: AB; blue: BR; violet: FHL; green: SD-A; yellow: SD-B). As-
terisks above/below bars indicate statistically significant differences from the 3/4-parent expectation. *p<
0.05; **p< 0.01; ***p< 0.001.
cross directions exhibited slightly, though not statistically significantly, reduced survivorship.
Family sex ratios were more female-biased than expected in 6 out of the 7 crosses (Figure 2). This bias
was statistically significant in 3 crosses, including both crosses where F1s were backcrossed to AB (Table
5).
Morphometrics
Differences between parental lines
Morphometric measurements differed significantly between parental populations only in crosses in-
volving AB, which differed from SD in urosome length, and in left and right clasper lengths. AB also
significantly differed from BR in left, but not right, clasper length (Table 6).
67
Table 5: Hatch number, survivorship, and sex ratio measurements for parental lines and non-recombinant
backcrosses. Deviations from 3/4-parent expectations were calculated by ANOV A and planned contrast
tests. Statistically significant differences are bolded.
Hatch number Survivorship Family sex ratio
Cross n Mean ( Std. Dev.) p n Mean ( Std. Dev.) p n Mean ( Std. Dev.) p
(AB x AB) x AB 72 25.81 (8:33) – 72 0.728 (0:252) – 68 0.423 (0:210) –
(BR x BR) x BR 54 19.11 (7:91) – 54 0.467 (0:339) – 44 0.575 (0:313) –
(FHL x FHL) x FHL 9 17.67 (10:28) – 9 0.612 (0:394) – 7 0.293 (0:206) –
(SD-B x SD-B) x SD-B 21 17.14 (8:31) – 21 0.469 (0:332) – 17 0.391 (0:382) –
(SD-A x SD-A) x SD-A 53 25.34 (8:74) – 74 0.516 (0:373) – 58 0.505 (0:321) –
(BR x SD-A) x SD-A 65 22.71 (7:78) 0.442 65 0.452 (0:376) 0.215 46 0.476 (0:272) 0.109
(SD-A x BR) x BR 89 25.17 (10.62) <0.001 89 0.512 (0.336) 0.215 79 0.518 (0.276) 0.234
(BR x SD-B) x SD-B 64 19.66 (7:74) 0.254 64 0.324 (0.359) 0.031 36 0.371 (0.295) 0.318
(SD-Bx BR) x BR 51 17.59 (6:59) 0.504 51 0.481 (0:325) 0.821 42 0.423 (0:317) 0.045
(AB x BR) x BR 111 13.71 (9.22) <0.001 111 0.479 (0.315) 0.239 92 0.524 (0:293) 0.744
(BR x AB) x AB 83 23.28 (7:51) 0.487 83 0.744 (0:252) 0.079 77 0.322 (0.189) <0.001
(AB x SD-A) x SD-A 83 16.16 (7.43) <0.001 83 0.510 (0:291) 0.139 76 0.454 (0:273) 0.136
(SD-A x AB) x AB 88 22.72 (10.28) 0.014 88 0.610 (0:293) 0.125 83 0.322 (0.199) <0.001
(SD-B x FHL) x FHL 20 17.50 (7:68) 0.99 20 0.698 (0:304) 0.26 19 0.436 (0:272) 0.229
(FHL x SD-B) x SD-B 31 1.84 (0.52) <0.001 – – – – – –
Table 6: Differences between parental lines for morphometric measures. Deviations from equal values were
calculated using independent two-sample t-tests (p-values are shown). Bolded values remain significant after
FDR correction. For significant differences, the population with longer/wider cephalothoraxes, urosomes,
or claspers is in parentheses.
Cephalothorax Urosome Clasper length
Population 1 Population 2 Length Width Length Width Left Right
AB BR 0.531 0.319 0.510 0.577 0.003 (AB) 0.052
SD-A 0.002 (AB) 0.305 <0.001 (AB) 0.274 <0.001 (AB) <0.001 (AB)
BR SD-B 0.638 0.098 0.696 0.367 0.191 0.776
SD-A 0.025 0.198 0.191 0.707 0.549 0.476
SD-B FHL 0.394 0.133 0.441 0.067 0.891 0.641
68
Hybrid fitness
Significant deviations from the 3/4-parent expectation were detected in at least two crosses per morpho-
metric measure. Both populations backcrossed to AB (BR and SD) had significantly shorter cephalothoraxes
and urosomes than the 3/4-parent expectation. Indeed, the average cephalothorax lengths of both (BR x AB)
x AB and (SD x AB) x AB were shorter than either parental line’s mean length. Conversely, the (BR x SD)
x SD and (AB x BR) x BR crosses had significantly longer-than-expected urosomes, and the latter hybrids
had longer urosomes than either parental population. (SD x BR) x BR and (AB x BR) x BR had significantly
wider-than-expected cephalothoraxes, while those of (AB x SD) x SD hybrids were significantly narrower
than expected. (AB x SD) x SD also had narrower-than-expected urosomes, as did the reciprocal cross, (SD
x AB) x AB, and the (BR x AB) x AB cross. Both left and right claspers were significantly longer than
expected in (AB x BR) x BR and (AB x SD) x SD and significantly shorter than expected in the reciprocal
crosses, (BR x AB) x AB and (SD x AB) x AB (Figure 3, Table 7).
We tested whether morphometric differences in hybrids were related to the genetic distance between
the parental populations (Figure 4). Linear regression found a small but significant effect of increasing
morphometric divergence with increasing genetic divergence (R
2
= 0:09, p = 0:02), but morphometric
changes were most extreme at intermediate divergence, in the crosses involving AB.
69
-6
-3
0
3
% Devia(on from 3/4-parent
Cephalothorax length
-3
0
3
(BR x SD-A) x SD-A
(SD-A x BR) x BR
(BR x SD-B) x SD-B
(SD-B x BR) x BR
(AB x BR) x BR
(BR x AB) x AB
(AB x SD-A) x SD-A
(SD-A x AB) x AB
(SD-B x FHL) x FHL
% Devia(on from 3/4-parent
Cephalothorax width
-6
-3
0
3
Urosome length
-3
-2
-1
0
1
2
(BR x SD-A) x SD-A
(SD-A x BR) x BR
(BR x SD-B) x SD-B
(SD-B x BR) x BR
(AB x BR) x BR
(BR x AB) x AB
(AB x SD-A) x SD-A
(SD-A x AB) x AB
(SD-B x FHL) x FHL
Urosome width
-8
-4
0
4
8
12
16
Le> clasper length
-8
-4
0
4
8
12
16
(BR x SD-A) x SD-A
(SD-A x BR) x BR
(BR x SD-B) x SD-B
(SD-B x BR) x BR
(AB x BR) x BR
(BR x AB) x AB
(AB x SD-A) x SD-A
(SD-A x AB) x AB
(SD-B x FHL) x FHL
Right clasper length
***
***
***
*
***
**
***
*
***
*
***
**
**
***
**
***
***
***
***
**
***
***
***
***
D E
F
A B C
BR
FHL
SD-B
AB
SD-A
Figure 3: Comparison of morphometric measurements to 3/4-parent expectations. Percent difference be-
tween observed and additive expected (3/4-parent) values is shown. For cephalothorax length (A), urosome
length (C), and left (E) and right (F) clasper lengths, positive values indicate longer-than-expected body
parts in hybrids, and negative values indicate shorter-than-expected body parts. For cephalothorax width
(B) and urosome width (D), positive values indicate wider-than-expected body parts in hybrids, and neg-
ative values indicate narrower-than-expected body parts. Crosses are colored according to their expected
genetic composition (red: AB; blue: BR; violet: FHL; green: SD-A; yellow: SD-B). Asterisks above/below
bars indicate statistically significant differences from the 3/4-parent expectation. *p< 0.05; **p< 0.01;
***p< 0.001.
70
Table 7: Morphometric measurements for parental lines and non-recombinant backcrosses. Morphometric measurements are given in pixels. Devia-
tions from 3/4-parent expectations were calculated by ANOV A and planned contrast tests. Statistically significant differences are bolded.
Cephalothorax Urosome Clasper length
Length Width Length Width Left Right
Cross n Mean ( SD) p n Mean ( SD) p n Mean ( SD) p n Mean ( SD) p n Mean ( SD) p n Mean ( SD) p
(AB x AB) x AB 320 246.34 (14.49) – 319 167.93 (7.36) – 306 161.97 (16.84) – 306 57.67 (4.71) – 320 65.87 (5.35) – 318 65.75 (5.12) –
(BR x BR) x BR 47 249.04 (15.60) – 47 166.28 (8.72) – 47 157.45 (14.96) – 47 57.60 (4.05) – 46 62.36 (4.48) – 46 62.71 (6.60) –
(FHL x FHL) x FHL 5 251.56 (9.64) – 5 179.47 (3.27) – 5 153.44 (12.70) – 5 60.74 (1.81) – 60 60.00 (2.75) – 4 59.96 (3.12) –
(SD-A x SD-A) x SD-A 95 241.48 (12.93) – 95 169.06 (5.67) – 86 151.59 (13.62) – 86 56.68 (2.99) – 94 61.99 (3.20) – 95 62.13 (3.74) –
(SD-B x SD-B) x SD-B 22 241.88 (16.36) – 22 172.59 (6.36) – 19 156.09 (16.10) – 19 54.39 (4.44) – 22 59.07 (4.99) – 22 61.97 (4.73) –
(BR x SD-A) x SD-A 255 242.95 (16.03) 0.779 255 168.8 (8.29) 0.588 250 156.29 (14.28) 0.045 250 56.57 (3.38) 0.372 253 60.41 (4.04) 0.002 252 61.28 (3.88) 0.058
(SD-A x BR) x BR 264 247.38 (14.09) 0.899 264 172.97 (7.22) <0.001 260 158.61 (12.02) 0.169 263 57.04 (2.87) 0.464 255 61.27 (4.19) 0.121 259 61.64 (3.86) 0.15
(BR x SD-B) x SD-B 141 239.28 (14.50) 0.08 141 167.55 (7.2) 0.013 139 153.06 (18.37) 0.245 138 55.69 (3.60) 0.469 141 59.94 (3.77) 0.967 139 60.72 (3.80) 0.101
(SD-B x BR) x BR 173 245.83 (10.69) 0.486 173 172.56 (5.31) <0.001 167 157.91 (11.23) 0.71 167 56.11 (2.57) 0.173 168 60.32 (3.79) 0.092 171 60.21 (3.96) 0.001
(AB x BR) x BR 188 249.06 (14.00) 0.712 188 169.74 (8.13) 0.002 188 163.57 (14.26) 0.011 188 58.18 (3.46) 0.216 185 67.48 (4.75) <0.001 184 66.69 (4.98) <0.001
(BR x AB) x AB 188 235.37 (12.48) <0.001 188 167.46 (6.91) 0.941 178 153.28 (13.67) <0.001 178 56.59 (2.88) 0.001 185 60.95 (4.74) <0.001 186 60.85 (4.50) <0.001
(AB x SD-A) x SD-A 238 243.19 (14.88) 0.732 238 165.83 (9.14) <0.001 217 156.71 (15.74) 0.109 217 55.56 (3.04) <0.001 234 71.22 (8.04) <0.001 236 71.65 (7.57) <0.001
(SD-A x AB) x AB 247 240.11 (14.54) <0.001 247 168.89 (7.92) 0.266 212 154.57 (14.86) <0.001 212 56.58 (3.58) 0.003 240 61.13 (4.21) <0.001 241 61.51 (4.52) <0.001
(SD-B x FHL) x FHL 113 252.09 (14.93) 0.555 113 172.25 (7.53) 0.85 102 155.25 (12.02) 0.828 102 59.12 (3.23) 0.979 110 60.77 (5.78) 0.568 110 61.57 (6.12) 0.568
71
Figure 4: Linear regression of absolute value of percent deviation from the 3/4-parent for morphometric
traits versus pairwise genetic divergence (d
S
). Points are colored by morphometric measure.
Population-diagnostic genotyping
SNPtyping of the four crosses yielded 13 to 18 diagnostic SNPs per cross (Table 3). The nuclear genome
of T. californicus comprises 12 chromosomes (Ar-Rushdi 1963). We designed our population-diagnostic
markers based on reads that aligned to scaffolds containing markers that had previously been mapped in
freshly hatched F2 larvae (before segregation distortion is observed; Pritchard et al. 2011, Foley et al. 2011).
In order to ensure that loci segregated as expected based on a priori chromosomal assignment, we created a
new linkage map using all diagnostic nuclear loci for each cross. Only two loci demonstrated an unexpected
segregation pattern. During SNP assay design, locus A73407 was initially mapped to scaffold 22, which is
anchored to chromosome 4 in the draft genome by marker Tc39888 from (Foley et al. 2011). In an initial
validation run of parental and F1 samples, A73407 was confirmed to be diagnostic for the AB x BR and BR
x SD crosses. However, when genotyped in backcrossed animals, locus A73407 co-segregated strongly with
the 2 diagnostic loci on chromosome 1 in the AB x BR crosses (A30250 and A586812) and, in the BR x
SD crosses, with the single diagnostic locus for chromosome 1 (A586812). In both crosses, the LOD score
for linkage was found to be> 10. Subsequently, this locus was treated as being assigned to chromosome 1,
72
rather than chromosome 4, which left 1 diagnostic locus on chromosome 4 for the AB x BR cross, and no
diagnostic loci on chromosome 4 for the BR x SD cross.
Locus A522015 on chromosome 3 performed as expected in the parental lines but exhibited an unusual
segregation pattern in 1 of the 4 crosses for which it is diagnostic. In the (AB x BR) x BR cross, all indi-
viduals were genotyped as heterozygous at this locus. However, a second diagnostic locus on chromosome
3, A110311, did not deviate from expected Mendelian ratios, and neither locus deviated from expected
Mendelian ratios in the reciprocal cross. Linkage mapping successfully placed the two markers on the same
linkage group in the (BR x AB) x AB cross, but separated them with an independence LOD score of> 10
in the (AB x BR) x BR cross. Locus A522015 was therefore removed before further analysis for the AB x
BR crosses.
One pair of loci in the AB x BR crosses, A438880 and A500766 on chromosome 7, exhibited a re-
combination rate> 10x the overall average recombination rate for all other chromosomes (average for all
other chromosomes in both directions of the AB x BR crosses: 0:47%; chromosome 7 in (AB x BR) x BR:
15:6%, or 33x, higher; chromosome 7 in (BR x AB) x AB: 9:0%, or 19x, higher). The A500766 marker
also included an unusually high proportion of AB homozygote calls in the (AB x BR) x BR cross and of SD
homozygote calls in the (SD x BR) x BR cross, suggesting that there may be a SNP or deletion in BR or
that the marker may be amplifying non-specifically. The same pattern of unexpected SD homozygote calls
was seen for a different marker on chromosome 3, A110311 in both males and females from the (SD x AB)
x AB cross, lending further credence to the idea that there may be variations in copy number for this locus
between the AB, SD, and BR populations. In the T. californicus draft genome, these markers are located
approximately 50kb apart on scaffold 9.
Due to the unusually high recombination rates and the large numbers of individuals called as wrong-
parental homozygotes, locus A110311 on chromosome 3 in the AB x SD crosses, and locus A500766 on
chromosome 7 in the BR x SD and AB x SD crosses were removed from subsequent analyses. This left
diagnostic loci on 10 of 12 chromosomes in the BR x SD crosses, on 11 of 12 chromosomes in the AB x SD
crosses, and on all 12 chromosomes in the AB x BR and SD-B x FHL crosses.
In the (BR x SD) x SD cross, 25 animals from 5 families were found to have an SD mitochondrial geno-
type, rather than the expected BR. All animals within the same family had the same mitochondrial genotype
and appeared to be genuinely backcrossed to SD (e.g., possessed only heterozygote and SD homozygote
genotypes). Four of the 5 families shared a single set of maternal grandparents (that is, their mothers were
73
Figure 5: Frequency of potential recombination events versus genotyping error frequency. Points represent
chromosomes with at least two diagnostic SNPs; points are colored by cross. Dashed line indicates 1:1 ratio
of potential recombination events to genotyping errors. Points on or below the dashed line indicate that most
“recombination events” are likely attributable to genotyping errors for that chromosome. Points above the
dashed line may be indicative of the presence of actual recombination events or may indicate genotyping
bias for one or more loci on that chromosomes.
siblings), and the 5 genotyped families are the only offspring assayed from either of these grandparental
families. These individuals were thus designated as belonging to a third cross, (SD x BR) x SD, that was
not conducted in this study and were removed from subsequent analyses.
Meiotic recombination may occur at very low frequencies in females, or putative recombination events
may actually be genotyping errors (Harrison and Edmands 2006, Foley et al. 2013). In our samples, recom-
bination could lead to different genotype calls for linked markers but cannot a priori be distinguished from
genotyping errors. Wrong-parental homozygotes (e.g, SD homozygotes in animals backcrossed to AB),
on the other hand, are unequivocally genotyping errors. We examined both the frequency of genotyping
74
Table 8: Average ( std. dev.) genotyping error rate for each cross.
Cross Sex Loci Error rate
(BR x SD) x SD Males 12 0.45% (0.60%)
(SD x BR) x BR Males 12 0.44% (0.44%)
(AB x BR) x BR Males 16 0.14% (0.26%)
(BR x AB) x AB Males 16 1.49% (2.83%)
(AB x SD) x SD Males 13 0.57% (0.58%)
(SD x AB) x AB Males 13 0.27% (0.47%)
Females 13 0.41% (0.44%)
(SD x FHL) x FHL Males 16 2.08% (0.95%)
Females 16 1.16% (0.67%)
errors, and, where possible, the frequency of recombinant calls. Recombinant calls were those where linked
loci were assigned different genotype calls within an individual, while all wrong-parental homozygote calls
were designated genotyping errors. The rates of genotyping error overall were very low, averaging< 1%
of calls per locus in 4 of 7 crosses (Table 8). Each cross had a variable number of chromosomes with>1
diagnostic markers (Table 3). The putative recombination rate for the retained loci ranged from 0 to 2.97%
across all the crosses (Figure 5). In the two crosses where both males and females were genotyped–AB x
SD and SD-B x FHL–putative recombination rates were higher in females than in males (mean = 1.3% in
females, 0.4% in males; Welch’s two-sample t-test, p = 0.038). However, when putative recombination rates
are compared to genotyping error rates, most recombination frequencies are less than or equal to genotyping
error rates (Figure 5), suggesting that they are also attributable to errors in genotype calling rather than true
recombination. Four chromosomes in four crosses, however, display an excessive putative recombination
event frequency. These may represent actual rare recombination events, or they may indicate genotyping
bias.
Heterozygosity
On average, individual heterozygosity in backcrosses is expected to be 0.5. Deviations from this ex-
pected ratio were detected in both directions of the AB x BR cross, and in both males and females for the
(SD x AB) x AB cross (Table 9). In (AB x BR) x BR, individuals had higher-than-expected heterozygosity,
while in the reciprocal cross, (BR x AB) x AB, individuals had lower-than-expected heterozygosity. Both
75
Table 9: Average heterozygosity by cross and sex. Deviations from the expected proportion of 0.5 were
tested for with one-sample t-tests.
Males Females
Cross n Nuclear loci
a
n Indiv. Avg, het. p n Indiv. Avg, het. p
(BR x SD) x SD 12 191 0.50 (0.17) 0.754 – – –
(SD x BR) x BR 12 193 0.50 (0.16) 0.844 – – –
(AB x BR) x BR 16 174 0.55 (0.16) <0.001 – – –
(BR x AB) x AB 16 192 0.45 (0.18) <0.001 – – –
(AB x SD) x SD 14 192 0.50 (0.15) 0.976 – – –
(SD x AB) x AB 14 190 0.53 (0.15) 0.005 149 0.53 (0.15) 0.025
(SD-B x FHL) x FHL 16 100 0.51 (0.18) 0.455 99 0.51 (0.19) 0.837
a
All crosses were additionally genotyped at 1 diagnostic mitochondrial locus, which was not included in these results.
males and females from (SD x AB) x AB had higher heterozygosity than expected.
We also tested for differences in heterozygosity between males and females at individual loci with
Fisher’s exact tests. Six loci on 5 chromosomes differed in (SD x AB) x AB, as did 5 loci on 3 chromosomes
in (SD-B x FHL) x FHL (Table 10). Chromosomes 6 and 10 exhibited differences between males and
females in both crosses, while chromosomes 2, 4, and 8 only differed in (SD x AB) x AB and chromosome
11 differed only in (SD-B x FHL) x FHL. For all but chromosome 10 in (SD x AB) x AB, males were more
heterozygous than females. In (SD-B x FHL) x FHL, males were more heterozygous on chromosome 6, but
females were more heterozygous on chromosomes 10 and 11. In both crosses, the most extreme pattern was
seen on chromosome 10, and in both cases, females were more heterozygous than males.
Single-locus segregation distortion
We detected significant deviations from expected Mendelian ratios in 3 of 4 interpopulation crosses. In
the closest crosses, BR x SD, we genotyped a total of 384 males but detected no loci with non-Mendelian
ratios in either cross direction after FDR correction (Figure 6). The largest number of chromosomes ex-
hibiting segregation distortion were found in the intermediate divergence crosses, AB x BR and AB x SD
(Figure 7, Figure 8), while only two chromosomes were distorted in the SD-B x FHL cross, and only in
males (Figure 9).
In the AB x BR crosses, 366 males were genotyped (Table 9). For (AB x BR) x BR, three chromosomes
(chromosomes 7, 10, and 12) exhibited significant segregation distortion after FDR correction (Figure 7).
All three chromosomes exhibited fewer BR homozygotes (more heterozygotes) than expected. Chromosome
76
Table 10: Sex-specific allele frequency differences for individual loci. Differences were calculated using
Fisher’s Exact Test. Only statistically significant comparisons are shown.
Females Males
Cross Chromosome Locus n Prop. Het. n Prop. Het. p
(SD x AB) x AB 2 A178530 148 0.45 188 0.61 0.003
4 A102107 147 0.43 187 0.6 0.002
6 A251428 144 0.49 181 0.66 0.002
6 A166226 145 0.5 186 0.64 0.01
8 A460601 146 0.52 185 0.63 0.044
10 A542807 148 0.8 186 0.33 <0.001
(SD-B x FHL) x FHL 6 A251428 96 0.45 96 0.61 0.03
6 A395869 94 0.46 95 0.61 0.042
6 A166226 96 0.44 99 0.63 0.01
10 A542807 95 0.59 99 0.23 <0.001
11 A296010 97 0.62 99 0.43 0.011
10 was also distorted in the reciprocal cross, (BR x AB) x AB, as were chromosomes 1, 2, and 8. In this
cross, chromosomes 1 and 10 had an excess of AB homozygotes (fewer heterozygotes), while chromosomes
2 and 8 had excess heterozygosity. Across all samples on chromosome 10, therefore, BR alleles are missing
in males.
For the AB x SD crosses, we genotyped 149 females in one cross direction, in addition to 382 total males
from both cross directions, (Table 9). Fully half the chromosomes were distorted in males in both directions
of this cross, though only 4 chromosomes were distorted in both cross directions (Figure 8). In (AB x
SD) x SD, where only males were genotyped, we found that chromosomes 1, 4, 6, 8, 10, and 11 exhibited
significant deviations from Mendelian ratios. Chromosomes 1 and 10 were missing SD homozygotes (had
an excess of heterozygotes), while chromosomes 4, 6, 8, and 11 had excess SD homozygotes (were missing
heterozygotes). In (SD x AB) x AB males, chromosomes 2, 4, 5, 6, 8, and 10 were distorted. Chromosomes
2, 4, 5, 6, and 8 were missing AB homozygotes (had an excess of heterozygotes), while chromosome 10
displayed the opposite pattern, with an excess of AB homozygotes (a dearth of heterozygotes). In (SD x
AB) x AB females, only chromosome 10 was distorted. It displayed a contrary pattern to what was seen
in males, with an excess of heterozygotes. When male and female samples were combined, chromosomes
5, 6, and 8 were distorted, as seen in males. All three distorted chromosomes had fewer AB homozygotes
than expected (an excess of heterozygotes). Chromosome 10 was not significantly distorted in the combined
77
*
**
**
*
Figure 6: Genotype frequencies for BR x SD crosses. (A) (BR x SD) x SD males. (B) (SD x BR) x BR
males. BR homozygotes are blue, SD homozygotes are green, and heterozygotes are brown. Asterisks
indicate p-values for
2
tests for segregation distortion (deviation from a 1:1 heterozygote: homozygote
ratio): *p< 0.05, **p< 0.01, ***p< 0.001; black: significant after FDR correction; gray: not significant
after FDR correction.
sample, presumably because the divergent patterns in males and females canceled each other out.
In the most divergent cross, (SD x FHL) x FHL, we genotyped a smaller number of males and females
(100 and 99, respectively) and detected significant segregation distortion on chromosomes 2 and 10, but
only in males (Table 9, Figure 9). No loci survived FDR correction in females or in the combined sample.
Males at chromosome 10 were overwhelmingly FHL homozygotes, while the female ratio at this locus was
slightly, but not significantly, skewed towards heterozygotes.
Although AB is less genetically and geographically distant from SD than FHL, the number of distorted
chromosomes detected in this cross is much lower than in the AB x BR and AB x SD crosses. However,
because the number of genotyped individuals is smaller in this cross than in the others, we may have reduced
power to detect incompatibilities. We therefore performed 10,000 permutations of random downsampling of
genotyping data for the other crosses to 100 males, and, in (SD x AB) x AB, 99 females to examine whether
the small number of distorted chromosomes detected here might be due to sampling size limitations. With a
78
**
***
***
***
**
**
**
**
*
***
*
Figure 7: Genotype frequencies for AB x BR crosses. (A) (AB x BR) x BR males. (B) (BR x AB) x AB
males. AB homozygotes are red, BR homozygotes are blue, and heterozygotes are pink. Asterisks indicate
p-values for
2
tests for segregation distortion (deviation from a 1:1 heterozygote: homozygote ratio): *p
< 0.05, **p< 0.01, ***p< 0.001; black: significant after FDR correction; gray: not significant after FDR
correction.
sample size comparable to that in the SD x FHL cross, we would expect to detect 2 distorted chromosomes
(as seen in SD x FHL males) only in the AB x BR crosses, where we found 3–4 distorted chromosomes with
a larger dataset, suggesting that we have slightly reduced power to detect segregation distortion in this cross
(Table 11). However, the only crosses in the permuted sample set where no segregation distortion would
be expected to be detected (as in SD x FHL females) was in the BR x SD crosses, where no segregation
distortion was detected in the larger sample sets either, suggesting that females may not be under-sampled.
Additionally, segregation distortion was present on 1 chromosome in the full dataset of (SD x AB) x AB
females, and downsampling indicated that it would be extremely unlikely to detect no distorted loci when
only 99 females were sampled (Figure 10). Segregation distortion was also not detected when both sexes
were combined. The sample size for this dataset is comparable to the full set of sampled individuals in the
other crosses, where segregation distortion was not detected only in the closest crosses (BR x SD).
Even assuming that the (SD x FHL) x FHL cross actually has 3–4 distorted chromosomes, however, the
AB x SD crosses still exhibit the most segregation distortion, and thus segregation distortion peaks in an
79
***
**
*
***
**
**
***
*
*
***
**
**
*
***
***
***
***
*
***
**
**
**
Figure 8: Genotype frequencies for AB x SD crosses. (A) (AB x SD) x SD males. (B) (SD x AB) x AB
males (C) (SD x AB) x AB females. (D) (SD x AB) x AB males and females combined. SD homozygotes
are green, AB homozygotes are red, and heterozygotes are gold. Asterisks indicatep-values for
2
tests for
segregation distortion (deviation from a 1:1 heterozygote: homozygote ratio): *p< 0.05, **p< 0.01, ***p
< 0.001; black: significant after FDR correction; gray: not significant after FDR correction.
80
*
*
**
*
*
*
***
**
*
**
Figure 9: Genotype frequencies for (SD-B x FHL) x FHL cross. (A) Males. (B) Females. (C) Males
and females combined. SD homozygotes are green, FHL homozygotes are violet, and heterozygotes are
gray. Asterisks indicate p-values for
2
tests for segregation distortion (deviation from a 1:1 heterozygote:
homozygote ratio): *p< 0.05, **p< 0.01, ***p< 0.001; black: significant after FDR correction; gray: not
significant after FDR correction.
intermediate cross, rather than in the most genetically and geographically distant cross (Table 11).
Segregation distortion and geographic distance
Epistatic incompatibilities
Using
2
tests, we found no significant pairwise interactions for either direction of the BR x SD cross
(Table 12). In (AB x BR) x BR, we detected a single interaction, between chromosomes 5 and 10, but
with FDR, it only becomes statistically significant when nearly 50% of detected interactions are allowed
to be false positives. In the reciprocal cross, (BR x AB) x AB, we found two weak interactions, neither
of which is ever significant after FDR correction. The AB x SD cross differed greatly in the number of
detected interactions between cross directions. We detected none in (AB x SD) x SD but found 3 different
interactions with chromosome 10 in males (one of which is statistically significant even at an FDR of 5%)
and 1 in females (significant when FDR = 18%), and 3 in the combined sample (again, one of which is
statistically significant even at an FDR of 5%) in (SD x AB) x AB. In (SD-B x FHL) x FHL, we found one
interaction each in males and females, neither of which was significant after FDR correction. When males
and females were combined for the (SD-B x FHL) x FHL cross, no significant interactions were detected.
81
Table 11: Median, mean, and corrected mean number of distorted chromosomes per cross when downsam-
pled to SD x FHL sample sizes, from 10,000 permutations, and actual number of distorted chromosomes
observed in the full dataset. Not all chromosomes were genotyped in all crosses, so means are corrected
based on the proportion of chromosomes that were genotyped in both SD x FHL (all 12) and in each cross.
Cross Median Mean Corrected mean Observed
(BR x SD) x SD 0 0.23 0.28 0
(SD x BR) x BR 1 0.66 0.80 0
(AB x BR) x BR 2 1.72 1.72 3
(BR x AB) x AB 2 2.03 2.03 4
(AB x SD) x SD 3 3.50 3.80 6
(SD x AB) x AB males 3 3.21 3.49 6
(SD x AB) x AB females 1 1.48 1.61 1
(SD x AB) x AB combined 1 1.39 1.51 3
*
*
*
*
*
*
*
*
Figure 10: Number of chromosomes exhibiting segregation distortion per cross in 10,000 replicate resam-
plings of genotyped individuals at sample sizes equal to those genotyped for (SD x FHL) x FHL after FDR
correction. Histograms are colored by cross. Dashed lines: number of distortions detected in (SD x FHL)
x FHL males (dark purple), females (light blue), and males and females combined (light blue) after FDR
correction. Asterisks: the number of distorted chromosomes detected for each cross with the full dataset.
82
Table 12: Pairwise epistatic loci.
FDR Overrepresented
Cross Sex Interaction
2
p threshold
a
Genotypes
b
(BR x SD) x SD Males – – – –
(SD x BR) x BR Males – – – –
(AB x BR) x BR Males 5 x 10 6.74 x 10
-3
0.44 Mismatched
(BR x AB) x AB Males 2 x 10 0.034 n.s. Matched
1 x 6 0.044 n.s. Mismatched
(AB x SD) x SD Males – – – –
(SD x AB) x AB Males 2 x 10 6.68 x 10
-8
< 0:05 Matched
5 x 10 0.010 0.47 Mismatched
8 x 10 0.016 0.47 Mismatched
Females 5 x 10 0.002 0.18 Mismatched
Males and females 5 x 10 1.88 x 10
-4
< 0:05 Mismatched
3 x 4 0.013 0.44 Matched
1 x 4 0.015 0.47 Matched
(SD-B x FHL) x FHL Males 4 x 7 0.016 n.s. Mismatched
Females 2 x 4 0.047 n.s. Mismatched
Males and females – – – –
a
FDR level at which interaction is statistically significant. Values designated n.s. never reach significance with FDR correction.
b
Matched: more double homozygotes and double heterozygotes than expected; mismatched: fewer
double homozygotes and double heterozygotes than expected from single-locus genotype frequencies.
83
Out of 10 detected two-locus interactions, 7 had more mismatched genotypes (that is, a locus 1 heterozy-
gote/locus 2 homozygote genotype or vice versa) than expected based on single-locus genotype frequencies.
For the chromosome 5 x 10 interaction, mismatched genotypes were over-represented in (AB x BR) x BR
males and in males, females, and the combined sample. Conversely, in the chromosome 2 x 10 interac-
tion, seen in (BR x AB) x AB males and (SD x AB) x AB males, matched two-locus genotypes were
over-represented in both samples. No other two-locus interaction was detected more than once.
Morphometric QTL
In BR x SD, chromosome 11 was associated with cephalothorax length and width, but only in one cross
direction, (BR x SD) x SD, under a single-locus model (Figure 11). No loci were significant for any other
morphometric trait in (BR x SD) x SD, and none were significant for any trait in the reciprocal cross, (SD x
BR) x BR. Under a two-locus model, chromosome 2 was supported in an additive model with chromosomes
11 (LOD = 3.14) and 3 (LOD = 2.82) for cephalothorax width, but not length in (BR x SD) x SD, and no
significant loci were detected in (SD x BR) x BR. The stepwise model gave weak support for the presence
of a single QTL on chromosome 11 for cephalothorax length and width (penalized LOD = 0.359 and 0.241,
respectively) in (BR x SD) x SD, and no QTL for any other morphometric trait in either cross direction.
In (AB x BR) x BR, cephalothorax length and width were mapped to chromosome 3 (Figure 12), and no
significant QTL were detected in the (BR x AB) x AB cross under a single-locus model. Under a two-locus
model, we detected an additive effect of chromosomes 3 and 5 for cephalothorax length and width (additive
LOD = 4.24 and 4.41, respectively) in (AB x BR) x BR. Stepwise model selection found that an additive
model including chromosomes 3 and 5 best described cephalothorax length, but support was low (penalized
LOD = 0.60). For cephalothorax width, stepwise model selection favored a single QTL on chromosome 3
(penalized LOD = 1.102). No QTL were supported for urosome length or width or clasper length in (AB x
BR) x BR, and none were supported for any morphometric trait in (BR x AB) x AB using stepwise model
selection.
Under a single-locus model, we detected QTL for cephalothorax length and width on chromosomes
2, 8, and 10; for urosome width on chromosomes 2 and 10; and for right, but not left, clasper length on
chromosome 4 in (SD x AB) x AB but no QTL for any morphometric trait in (AB x SD) x SD (Figure 13).
Using a two-locus model, we detected additive QTL on chromosomes 2, 4, 6, 8, and 10 for cephalothorax
length (additive LOD 3.36 to 4.86) and width (additive LOD 5.03 to 10.60) in (SD x AB) x AB, and we
84
B A
D C
F E
Mt Mt
Mt Mt
Mt Mt
Figure 11: Morphometric QTL for the BR x SD crosses under a single-locus model. (A) Cephalothorax
length; (B) cephalothorax width; (C) urosome length; (D) urosome width; (E) left and (F) right clasper
length. Orange: (BR x SD) x SD; dark red: (SD x BR) x BR. Dashed lines indicate p = 0.05 significance
threshold as determined by permutation.
85
B A
D C
F E
Mt Mt
Mt Mt
Mt Mt
Figure 12: Morphometric QTL for the AB x BR crosses under a single-locus model. (A) Cephalothorax
length; (B) cephalothorax width; (C) urosome length; (D) urosome width; (E) left and (F) right clasper
length. Red: (AB x BR) x BR; green: (BR x AB) x AB. Dashed lines indicate p = 0.05 significance
threshold as determined by permutation.
86
B A
D C
F E
Mt Mt
Mt Mt
Mt Mt
Figure 13: Morphometric QTL for the AB x SD crosses under a single-locus model. (A) Cephalothorax
length; (B) cephalothorax width; (C) urosome length; (D) urosome width; (E) left and (F) right clasper
length. Dark green: (AB x SD) x SD; light blue: (SD x AB) x AB. Dashed lines indicatep = 0.05 significance
threshold as determined by permutation.
87
found evidence for a single pair of epistatic QTL on chromosomes 2 and 4 in (AB x SD) x SD (interaction
LOD = 3.45). With stepwise model selection, we found evidence for additive QTL on chromosomes 4
and 10 and for an epistatic interaction between them for cephalothorax length (penalized LOD = 1.64); for
additive QTL on chromosomes 8 and 10 for cephalothorax length (penalized LOD = 7.15) and for a single
QTL on chromosome 10 for urosome width (penalized LOD = 3.17) in (SD x AB) x AB. We also found
weak evidence of a single QTL on chromosome 4 for right clasper length (penalized LOD = 0.039) in the
same cross, and no evidence for any QTL for any morphometric trait in the (AB x SD) x SD cross.
In (SD-B x FHL) x FHL, we found no QTL for any morphometric trait under a single-locus, two-locus,
or stepwise model (Figure 14).
Sex QTL
In addition to males, females were genotyped in two crosses. In (SD x AB) x AB, we found support for
an effect on sex on chromosomes 2, 4, and 10 under a single-locus model (Figure 15). Using a two-locus
model, we found support for additive effects of chromosomes 2, 4, 5, 6, 8, and 10 (additive LOD scores 3.29
to 17.98) and for an interaction between chromosomes 2 and 10 (interactive LOD = 3.84). However, with
stepwise model selection, we found that the best-fit model includes additive effects only for chromosomes
2, 4, and 10 and a single epistatic interaction between chromosomes 2 and 10 (penalized LOD = 23.3).
In (SD-B x FHL) x FHL, under a single-locus model, only chromosome 10 was significantly associated
with sex (Figure 16). Under a two-locus model, we found an additional, additive effect of chromosome 6
(additive LOD = 7.67), which was retained under stepwise model selection (penalized LOD = 4.23), and no
evidence for any epistatic interactions.
Discussion
Fitness consequences differ between cross directions even at high genetic divergence
In this study, we assessed the effects of hybridization in pairs of populations that spanned a broad
range of geographic and phylogenetic divergence. We assayed 3 metrics of hybrid fitness: hatch number,
survivorship, and family sex ratio, all of which demonstrated reduced viability in some crosses. We detected
a signal of increased hybrid fitness, in terms of hatch number, only in one cross, (SD x BR) x BR, and only
using the SD-A line. These viability changes were accompanied by signals of genetic incompatibility in all
88
B A
D C
F E
Mt Mt
Mt Mt
Mt Mt
Figure 14: Morphometric QTL for the SD x FHL cross under a single-locus model. (A) Cephalothorax
length; (B) cephalothorax width; (C) urosome length; (D) urosome width; (E) left and (F) right clasper
length. Dashed line indicatesp = 0.05 significance threshold as determined by permutation.
89
A
Mt
C B D
Figure 15: QTL mapping of sex determination in the (SD x AB) x AB cross. (A) Single-locus scan for
sex QTL. Dashed line indicates p = 0.05 significance threshold as determined by permutation. (B-D) Effect
plots of two-locus genotypes: (B) chromosomes 2 x 10; (C) chromosomes 4 x 10; (D) chromosomes 4 x 2.
90
A
B
Mt
Figure 16: QTL mapping of sex determination in the (SD x FHL) x FHL cross. (A) Single-locus scan for
sex QTL. Dashed line indicates p = 0.05 significance threshold as determined by permutation. (B) Effect
plot of two-locus genotypes for chromosomes 6 and 10.
91
but the closest crosses (BR x SD). We also found numerous changes in hybrid morphology, relative to the
additive expectation, but found few replicable QTL for morphometric characteristics. The strongest QTL
that we did detect were for the genetic basis of polygenic sex determination, which found both common and
unique loci across two backcrosses involving a common maternal grandparental population.
In 8 of 10 crosses, hatch number in hybrids was reduced relative to the 3/4-parent expectation, with 4
crosses achieving statistical significance, while only 2 crosses had increased numbers of offspring at hatch
(1 was statistically significant). The near-complete inviability of (FHL x SD) x SD animals upon hatching
(dishes were checked every 1-2 days after egg sacs were observed) was the most extreme and most surpris-
ing manifestation of reduced hybrid fitness we observed in this study and is unusual inT.californicus, where
hybridization rarely leads to complete or near-complete inviability (Edmands 1999). All other observations
of reproductive isolation this extreme in T. californicus have come from crosses between populations from
the southern edge of the species range, in central Baja California, and populations from the US, but these
southern populations are a separate species (Peterson et al. 2013, Chapter 2). Additionally, the extreme
inviability observed in the (FHL x SD) x SD cross manifested unusually early. Numerous previous studies
have demonstrated that freshly hatched hybrid nauplii generally do not exhibit the extreme patterns of segre-
gation distortion seen in adult hybrids, either at specific loci (Willett and Burton 2001, Willett 2006, Willett
and Berkowitz 2007, Willett 2008b) or genomewide (Foley et al. 2011, 2013). However, some studies have
seen weak and/or stochastic signals of segregation distortion in F
2
nauplii, though they are not as strong
as those that emerge throughout development to be present in adults (Pritchard et al. 2011, Willett 2011).
Recently, in AB x SD F
2
hybrids, it was shown that segregation distortion was present at at least one locus
immediately after hatching and became more extreme, but in the reverse direction, within 2 days of hatching
(SD homozygotes go from overrepresented to underrepresented, and are essentially absent in adults; Willett
et al. in press). This, in combination with our results, suggests that lethal inviabilities may begin to act very
shortly after hatching.
Five of 9 crosses also demonstrated reduced 28-day survivorship relative to the 3/4-parent mean, though
only one was statistically significant. The isofemale lines utilized in this study were highly inbred, having
been maintained for between 6 months and 7 years before the commencement of these studies, and both
hatch numbers and survivorship of most of the parental lines were low relative to outbred populations (Vittor
1971, Edmands 1999). However, we also detected a positive effect of hybridization for hatch number in the
(SD-A x BR) x BR cross. Inbreeding depression reduces fecundity in T. californicus (Brown 1991), and
92
heterosis has sometimes been observed in crosses between closely related populations, including in one
recombinant inbred line from a cross between SD females and BR males (Edmands 1999, Pereira et al.
2014).
Because we only backcrossed F
1
hybrids to the paternal, and not also to the maternal, population, we
cannot determine whether the asymmetric incompatibilities we detected are the result of cytonuclear interac-
tions. Ellison and Burton (2008b) documented near-complete fitness recoveries of F
3
T.californicus hybrids
when backcrossed to the maternal population, while paternal backcrosses maintained fitness levels similar
to F
2
s and F
3
, including in both directions of AB x SD crosses. However, other studies have found that
many nuclear-nuclear interactions also contribute to hybrid breakdown, even in crosses where mitonuclear
interactions play a role (Willett 2006, Edmands et al. 2009, Pritchard et al. 2011, Foley et al. 2013).
Male hybrids fare worse than females
Eight of 9 crosses had family sex ratios that were more female-biased than expected (3 were statistically
significant), suggesting widespread male-biased mortality in the backcross hybrids. We also detected more
chromosomes with segregation distortion in males than in females in both crosses where we genotyped both
sexes, and we found more 2-locus interactions in males than in females in the (SD x AB) x AB cross. In many
species, sex-biased mortality occurs in the heterogametic sex, a manifestation of Haldane’s rule (Haldane
1922). However, sex determination inT.californicus is polygenic, rather than chromosomal (Alexander et al.
2015), and females, rather than males, lack meiotic recombination, a trait often associated with heterogamety
(Ar-Rushdi 1963, Burton et al. 1981, Harrison and Edmands 2006). Consequently, the pattern observed
here of greater hybrid breakdown in males must have an explanation other than X-autosome interactions.
Overall, sex-specific hybrid breakdown may be highly variable in Tigriopus, perhaps in part due to the
different sex determination dynamics occurring within each population. For example, Foley et al. (2013)
found more segregation distortion in males, while others have found that female hybrids exhibit equivalent
or more segregation distortion than males, either genomewide (Harrison and Edmands 2006) or at specific
loci (Willett 2006, Willett and Berkowitz 2007, Willett 2011). Like Foley et al. (2013), both of the crosses
where we genotyped females involved the SD population (and, in fact, the same isofemale line), and it is
possible that this pattern would not be seen if AB x BR females were genotyped, though we also found
evidence for male-biased mortality in this cross.
One potential cause of more severe hybrid breakdown in males is the mother’s curse hypothesis, which
93
posits that because mitochondria are uniparentally inherited in most taxa, male-harming mutations in the
mitochondrial genome are immune to selection so long as they do not also harm females. Thus, male-
detrimental mutations may accumulate in the mitochondrial genome (Gemmell et al. 2004). This phe-
nomenon has been invoked to explain reductions in male fertility and lifespan (Smith et al. 2010, Camus
et al. 2012). Because all of our backcrosses were to the paternal population, only heterozygotes retain nu-
clear alleles that match the mitochondrial genotype, so we cannot directly test for the presence of mother’s
curse, but it could provide a partial explanation for the consistently stronger detrimental effects of hy-
bridization we observed for males. The mitochondrial genome in T. californicus evolves much faster than
the nuclear genome, and many populations possess considerable divergence even across small geographic
distances (e.g., BR and SD are nearly 10% divergent across the mitochondrial genome but less than 1%
divergent for nuclear genes; Pereira et al. 2016).
The genetic architecture of hybrid incompatibility is both conserved across and unique to
different crosses
Genome-wide heterozygosity deviated from the expected average of 50% in 3 crosses but to different
effect: in the AB x BR crosses, heterozygotes were over-represented when AB was the 1/4-parent and under-
represented when AB was the 3/4-parent, suggesting selection for AB alleles in these hybrids; in contrast,
heterozygotes were over-represented in both sexes when AB was the 3/4-parent but not when AB was the
1/4-parent in the SD x AB cross, suggesting selection against AB alleles. Allele-specific fitness, then, may be
highly context-dependent. This pattern could be due to incompatibilities that have evolved on the SD branch,
or to population-specific processes, such as bottlenecks, that increase genetic load in populations, or to some
combination of both. An examination of transcriptomes from six T. californicus populations found that all
populations, but especially AB, BR, and SD, exhibited signatures of small effective population size and
elevated frequencies of deleterious mutations, most of which were fixed between populations (Pereira et al.
2016). Previous crosses between a different pair of T. californicus populations from Southern California
produced F
2
s with deficits of mitochondrially-matched parental homozygotes at several nuclear loci, which
they suggested could be a result of genetic or even epistatic load incurred by populations with low effective
population sizes and repeated genetic bottlenecks (Edmands et al. 2009).
We detected segregation distortion on two or more chromosomes in all but the genetically, geographi-
94
cally, and phylogenetically closest cross (BR x SD). Contrary to our initial predictions, however, far more
chromosomes were distorted in the AB x BR and AB x SD crosses than in the SD x FHL cross. This is most
likely due to the reduced sample size in the latter cross, however, rather than to the absence of single-locus
viability effects. Two previous studies of F
2
hybrids between SD and populations from the northern clade,
which extends from northern California to southern Alaska (Edmands 2001, Peterson et al. 2013, Chapter
2), found extensive segregation distortion (Foley et al. 2013, Alexander et al. 2015).
Despite detecting many chromosomes exhibiting segregation distortion, we found evidence for very few
epistatic incompatibilities, and our strongest candidate interaction, which was found only in (SD x AB) x
AB males, was also identified as a candidate epistatic locus pair involved in sex determination in this cross.
This failure to detect DMIs is likely due at least in part to the low power of the
2
test. A simulation
study examining the power to detect two-locus viability DMIs in yeast, whose haploid spore genotype
probabilities are analogous to those in a backcross, found that with sample sizes comparable to those used
here (200 spores), the ability to detect viability DMIs ranged from 77% when DMIs were few (8) and of
large effect (92% lethality each) to a mere 2.4% when DMIs were numerous (15) and of lesser effect (52%
lethality each) (Li et al. 2013). Our results thus suggest that though DMIs may be numerous in these crosses,
they are likely to have small effects that will require recombination to dissect.
Transgressive segregation for morphological traits
AB is the only parental line to differ from the others in morphometric traits. In spite of this, we found
evidence for numerous differences between hybrids and their parents, and between reciprocal hybrid crosses
for morphometric characteristics. Other studies have also found evidence of transgressive segregation for
morphometric traits in T. californicus crosses: Pritchard and Edmands (2012) found that egg sac size in-
creased in a cross between two highly divergent populations (PB, from Baja California, and SD), and Hwang
et al. (submitted) found evidence of transgressive segregation in cephalothorax length and egg sac size in a
cross between two populations of intermediate divergence (RP, which is close to AB, and SD). Transgressive
segregation has also been documented inT.californicus for thermal tolerance, but only in a low-divergence
cross (BR x SD), and only for some hybrid lines (Pereira et al. 2014). Transgressive segregation may be
attributable to novel combinations of alleles for traits with a complex genetic architecture (Rieseberg et al.
1999). However, we found low support for all the detected morphometric QTL, and none of the detected
QTL were found across reciprocal crosses. Furthermore, in the (SD x AB) x AB cross, all detected QTL
95
are on chromosomes that exhibit segregation distortion. This may be because most of the reported mor-
phometric changes in hybrids are relatively small, and there is considerable variation within crosses as well
as between them. A previous F
2
mapping study identified 0–5 QTL for cephalothorax length and width,
but in many cases the same loci were implicated in the same trait in multiple sexes and/or mitochondrial
backgrounds (Foley et al. 2013).
Locus-specific effects on sex are repeated across crosses
In both crosses where we genotyped females in addition to males, the hybrids were descended from SD
grandmothers. We found a large effect of chromosome 10 on sex in both crosses, where the segregation dis-
tortion seen within each sex was canceled out when the full dataset was combined, and we found evidence
that loci on other chromosomes have additive, in SD x FHL, and both additive and epistatic, in SD x AB,
effects on sex. This coincides well with a previous QTL mapping study between SD and a Canadian popu-
lation, which identified loci on five chromosomes, including all of those found in this study, as contributing
to sex determination (Alexander et al. 2015). Chromosome 10, especially has repeatedly been demonstrated
to exhibit a large effect on sex determination in several crosses involving SD. Alexander et al. (2015) found
evidence for two independent QTL on this chromosome, and Foley et al. (2013) also identified it as a poten-
tial site of sex-determining, but not meiotic drive, genes. Furthermore, in experimental hybrid populations
created by mixing SC and SD parental individuals, chromosome 10 repeatedly reached and then maintained
highly differentiated allele frequencies between males and females (where females, but not males, possessed
SD alleles) across years of admixture, even in the face of genetic swamping by SC (Pritchard and Edmands
(2012), Chapter 4). Taken in sum, these results suggest that chromosome 10 has emerged as a major con-
tributor to sex determination in SD, but perhaps not in other populations. Curiously, chromosome 10 is not
distorted in either direction of the BR x SD cross (Figure 6), but males from the AB x BR cross exhibit
segregation distortion in opposite directions in reciprocal backcrosses (Figure 7), such that BR alleles are
always underrepresented in AB x BR males, which is the same pattern seen in AB x SD males (Figure 8).
In sum, these results suggest that the major effect of chromosome 10 on sex may not be limited to SD but
rather may be regionally widespread (SD is from San Diego, CA, while BR is from La Jolla, a suburb of
San Diego; Figure 1). Unfortunately, because larvae are not sexually dimorphic, it is difficult to distinguish
between possible sex-determining and sex-specific viability effects for genetic markers. This blurred dis-
tinction is especially evident in the (SD x AB) x AB cross, where we detected both the strongest epistatic
96
viability effect and an effect on sex for chromosomes 2 x 10. Unraveling the genetic basis of polygenic
sex determination in T. californicus may therefore require the development of non-lethal genetic sampling
approaches, such that the genotypes of individual animals could be tracked across development.
Acknowledgements
Thanks to Mindy Guo for laboratory assistance with crosses and to Noel Reyes for performing morpho-
metric measurements. This work was supported by an NSF Graduate Research Fellowship to BCP and by
NSF grants DEB-1146520 and DEB-1355170 to SE.
97
References
Alexander, H., J. Richardson, S. Edmands, and B. Anholt. 2015. Sex without sex chromosomes: genetic
architecture of multiple loci independently segregating to determine sex ratios in the copepod Tigriopus
californicus. Journal of Evolutionary Biology 28:2196–2207.
Ar-Rushdi, A. H. 1958. The polygenic basis of sex-ratio in Tigriopus. Paper presented at the Xth Interna-
tional Congress of Genetics, McGill University, Montreal, Canada.
Ar-Rushdi, A. H. 1963. The cytology of achiasmatic meiosis in the female Tigriopus (Copepoda). Chro-
mosoma 13:1754–1763.
Benjamini, Y ., and Y . Hochberg. 1995. Controlling the false discovery rate: A practical and powerful
approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological) 57:289–
300.
Broman, K. W., H. Wu, S. Sen, and G. A. Churchill. 2003. R/qtl: QTL mapping in experimental crosses.
Bioinformatics 19:889–890.
Brown, A. F. 1991. Outbreeding depression as a cost of dispersal in the harpacticoid copepod, Tigriopus
californicus. Biological Bulletin 181:123–126.
Burton, R. 1987. Differentiation and integration of the genome in populations of the marine copepod
Tigriopuscalifornicus. Evolution 41:504–513.
Burton, R., and F. Barreto. 2012. A disproportionate role for mtDNA in Dobzhansky-Muller incompatibili-
ties? Molecular Ecology 20:4942–4957.
Burton, R., M. Feldman, and S. Swisher. 1981. Linkage relationships among 5 enzyme-coding gene loci in
the copepodTigriopuscalifornicus: A genetic confirmation of achiasmatic meiosis. Biochemical Genetics
19:1237–1245.
Burton, R. S. 1985. Mating system of the intertidal copepod Tigriopus californicus. Marine Biology
86:247–252.
Burton, R. S., C. K. Ellison, and J. S. Harrison. 2006. The sorry state of F
2
hybrids: Consequences of rapid
mitochondrial DNA evolution in allopatric populations. American Naturalist 168:S14–S24.
Burton, R. S., and M. W. Feldman. 1981. Population genetics of Tigriopus californicus: II. Differentiation
among neighboring populations. Evolution 35:1192–1205.
Camus, M. F., D. J. Clancy, and D. K. Dowling. 2012. Mitochondria, maternal inheritance, and male aging.
Current Biology 22:1717–1721.
Dobzhansky, T. 1937. Genetics and the Origin of Species. Columbia University Press, New York.
Edmands, S. 1999. Heterosis and outbreeding depression in interpopulation crosses spanning a wide range
of divergence. Evolution 53:1757–1768.
Edmands, S. 2001. Phylogeography of the intertidal copepod Tigriopus californicus reveals substantially
reduced population differentiation at northern latitudes. Molecular Ecology 10:1743–1750.
Edmands, S., S. Northrup, and A. Hwang. 2009. Maladapted gene complexes within populations of the
intertidal copepodTigriopuscalifornicus? Evolution 63:2184–2192.
98
Ellegren, H. 2009. Genomic evidence for a large-Z effect. Proceedings of the Royal Society B: Biological
Sciences 276:361.
Ellison, C. K., and R. S. Burton. 2008a. Genotype-dependent variation of mitochondrial transcriptional
profiles in interpopulation hybrids. Proceedings of the National Academy of Sciences of the United
States of America 105:15831–15836.
Ellison, C. K., and R. S. Burton. 2008b. Interpopulation hybrid breakdown maps to the mitochondrial
genome. Evolution 62:631–638.
Ellison, C. K., and R. S. Burton. 2010. Cytonuclear conflict in interpopulation hybrids: The role of RNA
polymerase in mtDNA transcription and replication. Journal of Evolutionary Biology 23:528–538.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, and S. Edmands. 2013. Postzygotic isolation involves
strong mitochondrial and sex–specific effects in Tigriopus californicus, a species lacking heteromorphic
sex chromosomes. Heredity 111:391–401.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, G. W. Moy, R. S. Burton, and S. Edmands. 2011.
A gene-based SNP resource and linkage map for the copepod Tigriopus californicus. BMC Genomics
12:568.
Fox, J., and S. Weisberg. 2011. An R Companion to Applied Regression. 2
nd
edition edition. Sage.
Gemmell, N. J., V . J. Metcalf, and F. W. Allendorf. 2004. Mother’s curse: the effect of mtDNA on individual
fitness and population viability. Trends in Ecology and Evolution 19:238–244.
Haldane, J. 1922. Sex ratio and unisexual sterility in hybrid animals. Journal of Genetics 12:101–109.
Harrison, J. S., and S. Edmands. 2006. Chromosomal basis of viability differences inTigriopuscalifornicus
interpopulation hybrids. Journal of Evolutionary Biology 19:2040–2051.
Hendry, A. P., D. I. Bolnick, D. Berner, and C. L. Peichel. 2009. Along the speciation continuum in
sticklebacks. Journal of Fish Biology 75:2000–2036.
Hwang, A., S. Northrup, J. Alexander, K. V o, and S. Edmands. 2011. Long-term experimental hybrid
swarms between moderately incompatible Tigriopus californicus populations: hybrid inferiority in early
generations yields to hybrid superiority in later generations. Conservation Genetics 12:895–909.
Hwang, A., S. Northrup, D. Peterson, Y . Kim, and S. Edmands. 2012. Long-term experimental hybrid
swarms between nearly incompatible Tigriopus californicus populations: persistent fitness problems and
assimilation by the superior population. Conservation Genetics 13:567–579.
Lazzaretto, I., and A. Libertini. 1985. Karyological comparisons among three species of the genusTigriopus
(Copepods, Harpacticoida). Atti. Acc. Lincei Rend. fis. 126:377–381.
Li, C., Z. Wang, and J. Zhang. 2013. Toward genome-wide identification of Bateson-Dobzhansky-Muller
incompatibilities in yeast: A simulation study. Genome Biology and Evolution 5:1261–1272.
Li, H., and R. Durbin. 2009. Fast and accurate short read alignment with Burrows–Wheeler transform.
Bioinformatics 25:1754–1760.
Lima, T. G. 2014. Higher levels of sex chromosome heteromorphism are associated with markedly stronger
reproductive isolation. Nature Communications 5:4743.
99
Mank, J., and J. Avise. 2009. Evolutionary Diversity and Turn-Over of Sex Determination in Teleost Fishes.
Sexual Development 3:60–67.
Muller, H. J., 1940. The New Systematics, Chapter: Bearing of theDrosophila work on systematics, pages
185–268 . Clarendon Press.
Ooijen, J. W. V ., 2006. Join Map 4, Software for the calculation of genetic linkage maps in experimental
populations. Kyazma B.V ., Wageningen, Netherlands.
Palmer, C., and S. Edmands. 2000. Mate choice in the face of both inbreeding and outbreeding depression
in the intertidal copepodTigriopuscalifornicus. Marine Biology 136:693–698.
Pereira, R. J., F. S. Barreto, and R. S. Burton. 2014. Ecological novelty by hybridization: Experimental
evidence for increased thermal tolerance by transgressive segregation inTigriopuscalifornicus. Evolution
68:204–215.
Pereira, R. J., F. S. Barreto, N. T. Pierce, M. Carneiro, and R. S. Burton. 2016. Transcriptome-wide patterns
of divergence during allopatric evolution. Molecular Ecology 25:1478–1493.
Peterson, D., K. Kubow, M. Connolly, L. Kaplan, M. Wetkowski, W. Leong, B. Phillips, and S. Edmands.
2013. Reproductive and phylogenetic divergence of tidepool copepod populations across a narrow geo-
graphical boundary in Baja California. Journal of Biogeography 40:1664–1675.
Phillips, B. C., and S. Edmands. 2012. Does the speciation clock tick more slowly in the absence of
heteromorphic sex chromosomes? BioEssays 34:166–169.
Presgraves, D. C. 2008. Sex chromosomes and speciation inDrosophila. Trends in Genetics 24:336–343.
Presgraves, D. C. 2010. The molecular evolutionary basis of species formation. Nature Reviews Genetics
11:175–180.
Pritchard, V ., L. Dimond, J. Harrison, C. Velasquez, J. Zieba, R. Burton, and S. Edmands. 2011. Interpopu-
lation hybridization results in widespread viability selection across the genome inTigriopuscalifornicus.
BMC Genetics 12:54.
Pritchard, V ., and S. Edmands. 2012. The genomic trajectory of hybrid swarms: Outcomes of repeated
crosses between populations ofTigriopuscalifornicus. Evolution 67:774–791.
R Core Team, 2015. R: a language and environment for statistical computing.
Rieseberg, L. H., M. A. Archer, and R. K. Wayne. 1999. Transgressive segregation, adaptation and specia-
tion. Heredity 83:363–372.
Robinson, J. T., H. Thorvaldsd´ ottir, W. Winckler, M. Guttman, E. S. Lander, G. Getz, and J. P. Mesirov.
2011. Integrative genomics viewer. Nature Biotechnology 29:24–26.
Schilthuizen, M., M. C. W. G. Giesbers, and L. W. Beukeboom. 2011. Haldane’s rule in the 21st century.
Heredity 107:95–102.
Schneider, C. A., W. S. Rasband, and K. W. Eliceiri. 2012. NIH Image to ImageJ: 25 years of image
analysis. Nature Methods 9:671–675.
Shaw, K. L., and S. P. Mullen. 2014. Speciation continuum. Journal of Heredity 105:741–742.
100
Smith, S., C. Turbill, and F. Suchentrunk. 2010. Introducing mother’s curse: low male fertility associated
with an imported mtDNA haplotype in a captive colony of brown hares. Molecular Ecology 19:36–43.
The Tree of Sex Consortium. 2014. Sex determination: Why so many ways of doing it? PLoS Biology
12:e1001899.
Uricaru, R., G. Rizk, V . Lacroix, E. Quillery, O. Plantard, R. Chikhi, C. Lemaitre, and P. Peterlongo. 2015.
Reference-free detection of isolated SNPs. Nucleic acids research 43:e11–e11.
Vittor, B. A., 1971. Effects of the environment on fitness-related life history characters inTigriopuscalifor-
nicus. Ph.D. thesis, University of Oregon.
Warnes, G. R., B. Bolker, T. Lumley, and R. C. Johnson, 2015. gmodels: Various R Programming Tools for
Model Fitting, r package version 2.16.2 edition.
Willett, C., and R. Burton. 2001. Viability of cytochrome C genotypes depends on cytoplasmic backgrounds
inTigriopuscalifornicus. Evolution 55:1592–1599.
Willett, C., and R. Burton. 2003. Environmental influences on epistatic interactions: Viabilities of cy-
tochrome c genotypes in interpopulation crosses. Evolution 57:2286–2292.
Willett, C. S. 2006. Deleterious epistatic interactions between electron transport system protein-coding loci
in the copepodTigriopuscalifornicus. Genetics 173:1465–1477.
Willett, C. S. 2008a. No evidence for faster male hybrid sterility in population crosses of an intertidal
copepod (Tigriopuscalifornicus). Genetica 133:129–136.
Willett, C. S. 2008b. Significant Variation for Fitness Impacts of ETS Loci in Hybrids between Populations
ofTigriopuscalifornicus. Journal of Heredity 99:56–65.
Willett, C. S. 2011. Complex deleterious interactions associated with malic enzyme may contribute to
reproductive isolation in the copepodTigriopuscalifornicus. PLoS One 6:e21177.
Willett, C. S., and J. N. Berkowitz. 2007. Viability effects and not meoitic drive cause dramatic departures
from Mendelian inheritance for malic enzyme in hybrids of Tigriopus californicus populations. Journal
of Evolutionary Biology 20:1196–1205.
Willett, C. S., T. G. Lima, I. Kovaleva, and L. Hatfield. in press. Chromosome-wide impacts on the
expression of incompatibilities in hybrids ofTigriopuscalifornicus. G3: Genes, Genomes, Genetics .
101
Chapter4: Emergenceofanovelsexchromosomeinanexperimentalhybrid
populationofaspecieswithpolygenicsexdetermination
Barret C. Phillips, Eric T. Watson, Victoria L. Pritchard, Suzanne Edmands
Abstract
Chromosomal sex determination has evolved repeatedly, and interactions between sex chromosomes
and autosomes frequently underlie postzygotic isolation. However, little is known about how postzygotic
isolation evolves in species lacking heteromorphic sex chromosomes. Where multiple, unlinked loci are
involved in sex determination (polygenic sex determination), the invasion of alleles of strong effect, coupled
with sexual antagonists, offers a potential route to the evolution of new sex chromosomes. Hybridization
in species with polygenic sex determination potentially involves the interaction of novel combinations of
sex-determining alleles—as well as genetic incompatibilities—and, if coupled with subsequent inbreeding,
may provide a tipping point for a transition to chromosomal sex determination. We document a transition
from polygenic sex determination towards chromosomal sex determination in the copepod Tigriopus cali-
fornicus, using an experimental hybrid swarm composed of two highly divergent populations (SC and SD).
Population–diagnostic SNP-typing revealed the sex–specific presence of SD alleles across chromosome 10,
which were maintained for>4 years (50 generations), despite genetic swamping of the SC population’s
alleles across most of the remainder of the genome early in the history of the hybrid population. Associ-
ation tests and principal components analysis identified this chromosome as strongly associated with sex
after 4 years of admixture. After >7 years (90 generations) of admixture, family sex ratios were not
binomially distributed in the hybrid population, as would be expected under fully chromosomal sex deter-
mination. Whole–genome re-sequencing of pooled samples revealed elevated allele frequency differences
between males and females on chromosome 10 in the hybrid population but not in the parental populations.
Sequencing of individual animals from the experimental hybrid population and its parental populations pro-
vided genome-wide confirmation that no degenerated sex chromosomes exist in any of the 3 sequenced
populations and revealed that genetic variance measurably contributes to sex in they hybrid population but
not in the parental populations. This study is the first to examine the genomes of individual T. californicus
copepods and, to our knowledge, is the first example of the emergence of a sex chromosome in a species
102
with polygenic sex determination.
Introduction
Sex chromosomes are ubiquitous, yet evolutionarily labile. Closely related species, or in some cases,
populations of the same species, may have different sex chromosomes, while many other species lack sex
chromosomes entirely (The Tree of Sex Consortium 2014, Rodrigues et al. 2015). Their frequent emergence
may be due to their ability to resolve sexually antagonistic genomic conflicts, or because they provide a
stable route of transmitting a balanced sex ratio, or both (Bull 1983). Theory suggests that population sex
ratios should evolve toward a stable equilibrium wherein the parental investment in each sex is equivalent
(Fisher 1930), and sex chromosomes are widely invoked as a stable mechanism by which to transmit a 1:1
sex ratio to offspring (Bull 1983, Charlesworth 1991).
Polygenic sex determination, where sex is determined by multiple, unlinked genes, differs from chromo-
somal sex determination in that the genetic basis of sex determination is a quantitative trait with a binomial
outcome. It is usually envisioned as a threshold system, whereby individuals possessing some minimum
number of maleness or femaleness factors will become male or female, and those below the threshold de-
velop into the opposite sex (Bulmer and Bull 1982). Polygenic sex determination has been described in an
increasing number of organisms in recent years, including frogs, flies, crustaceans, bivalves, and numer-
ous fish species (e.g., Ar-Rushdi 1958, Franco et al. 1982, Vandeputte et al. 2007, Yusa 2007, Ser et al.
2010, Ghiselli et al. 2012, Liew et al. 2012, Parnell and Streelman 2013, Rodrigues et al. 2013, Yusa et al.
2013). Reports of species with segregating sex chromosomes have also increased in number in recent years,
suggesting that there may be many natural opportunities to witness the earliest stages of sex chromosome
evolution (Franco et al. 1982, Green 1988, Green et al. 1993, Sharbel and Green 1998, Ser et al. 2010,
Parnell and Streelman 2013, Rodrigues et al. 2013, 2015).
Systems with polygenic sex determination are expected to be inherently unstable because they are easily
disrupted by interactions between sexually antagonistic alleles and novel sex determiners. Thus they have
often been envisioned as a transitory state between [presumably more stable] chromosomal systems (Bull
and Charnov 1977, Bull 1983, Rice 1986, van Doorn and Kirkpatrick 2010). However, more recent research
demonstrates that, under certain conditions, polygenic sex determination systems may be a state of stable
equilibrium (van Doorn and Kirkpatrick 2007, 2010).
103
Under traditional models, a sex chromosome may emerge when a sex determiner (inducing maleness
or femaleness) and a sexual antagonist (beneficial to one sex but detrimental to the other) become linked
(Bull 1983, Rice 1986, 1992). In this scenario, recombination will be selected against because it creates
unfavorable genotypes. If recombination is suppressed, the new X and Y chromosomes will be free to
diverge due to differing selective pressures and effective population sizes. Degeneration, like that seen in
the mammalian Y , is thus a direct consequence of the cessation of recombination (Charlesworth 1991).
This has been demonstrated experimentally—when autosomes were forced to act like sex chromosomes
in Drosophila, males displayed reduced fitness over the course of fewer than 40 generations, implying
that deleterious mutations had accumulated (Rice 1992, 1994). However, when recombination is absent
in one sex (usually the heterogametic sex, a pattern known as the Haldane–Huxley rule), the evolution of
a sex chromosome requires only the emergence of a strong sex determiner, and no linkage to a sexually
antagonistic locus is necessary (Wright et al. 2016).
Conditions under which sex chromosomes “turn over” (switch from one set to another) are also primarily
thought to involve the linkage of a sex determiner and a sexual antagonist or to involve segregation distortion
(Bull and Charnov 1977, van Doorn and Kirkpatrick 2007, 2010). For example, in cichlids from Lake
Malawi, both XY and ZW systems segregate within species (Ser et al. 2010, Parnell and Streelman 2013).
The ZW system is tightly linked to a sexually antagonistic color allele and is also thought to be the newer,
invasive system (Roberts et al. 2009). However, the same chromosomes act as XY and ZW in multiple
cichlid species within Lake Malawi, suggesting that polygenic sex determination has persisted in a complex
state across speciation events (Ser et al. 2010, Parnell and Streelman 2013).
Hybridization is another potential source of genomic disruption that might result in transitions between
sex chromosomes because it may result in novel combinations of sex determiners that throw into disarray
previously stable systems. True-breeding hybrid lines with novel sex chromosomes have been generated in
laboratory populations of Drosophila nasuta nasuta (2n = 8) x D. n. albomicans (2n = 6) hybrids where
an autosome fused to the X, creating “cytoraces” with 2n = 6 females and 2n = 7 males (Ramachandra and
Ranganath 1986, Tanuja et al. 1999, Yu et al. 1999). Likewise, a neo-X chromosome, created by the fusion
of an autosome to the Y chromosome, has been linked to speciation in marine sticklebacks because both
the original and neo-X chromosomes contain loci that directly contribute to reproductive isolation with the
ancestral, sympatric species (Kitano et al. 2009).
In species with polygenic sex determination, allopatric populations on divergent evolutionary tracks may
104
thus also possess divergent sex determination alleles that are well-calibrated within their own population but
which might not be compatible with sex–determining alleles in other populations. This could thus provide an
opportunity for a strong sex determiner to invade a polygenic system, as predicted by Rice (1986). Polygenic
sex determination is well–documented in the harpacticoid copepodTigriopuscalifornicus: family sex ratios
are heritable, respond to selection, and do not conform to a binomial distribution, and QTL mapping in
interpopulation crosses has identified multiple loci associated with sex on several chromosomes (V oordouw
and Anholt 2002, V oordouw et al. 2008, 2005, Alexander et al. 2014, 2015, Ar-Rushdi 1958, Chapter 3).
Additionally,T.californicus exclusively inhabits splash pools above the mean high tide line along the Pacific
coast of North America, from Baja California to Alaska, and populations are restricted to individual rock
outcrops and display high genetic divergence across the species range, often even between geographically
close sites (Peterson et al. 2013, Chapter 2). In spite of this unusually high level of genetic divergence,
interpopulation crosses readily produce hybrids in the laboratory, and hybrid populations may be maintained
indefinitely.
Here, we examine the sex-determining mechanisms in one hybrid and two natural populations of T.
californicus. Population–diagnostic genotyping of a long-term experimental hybrid population revealed a
segregating, female-specific pattern of heterozygosity across an entire chromosome in the face of genetic
swamping across the remainder of the nuclear and mitochondrial genome. Family sex ratio assays conducted
in both the parental populations and the experimental hybrid population demonstrated that family sex ratios
in all populations did not conform to a binomial distribution, as would be expected under chromosomal,
but not polygenic, sex determination. With whole–genome re-sequencing pooled individuals from all three
populations, we found that allele frequency differences were elevated between males and females across a
single chromosome in the hybrid population but not in the parental populations. By sequencing individual
animals from all 3 populations, we confirmed that degenerated chromosomes are not present inT.californi-
cus and that within natural populations, only a small proportion of the inter–individual variation present is
explained by sex. To our knowledge, this is the first experimental observation of the emergence a novel sex
chromosome in a species with polygenic sex determination.
105
Materialsandmethods
Collectionandmaintenanceofanimals
Experimentalhybridpopulation
As part of previous studies (Pritchard and Edmands 2013, Pritchard et al. 2013), eight hybrid swarms
were started from freshly collected animals from San Diego, CA (SD;32

45
0
N,117

15
0
W) and Santa Cruz,
CA (SC; 36

57
0
N, 122

03
0
W) in August 2006. 150 gravid females from each population were pooled into
1L beakers and maintained for 21 months at 20

C with a 12/12 light/dark cycle and regular water changes
and feeding. After 21 months, the extant swarms were maintained at 15

C with a 12/12 light/dark cycle
and occasional water changes and feeding. One replicate (designated E4 in Pritchard and Edmands 2013
and Pritchard et al. 2013 but hereafter simply referred to as the hybrid population), survived more than 48
months. This hybrid swarm was sampled at 3, 6, 9, 15, 21 months of admixture (11–19 females and 15–20
males per timepoint, detailed in Pritchard and Edmands (2013)), and again at 48 (40 males, 40 females),
and 51 months (50 males, 48 females) of admixture, with individuals frozen at80

C. After 80 months of
admixture, the single extant hybrid swarm (E4) was transferred to20

C and maintained under the same light
conditions with occasional water changes and feeding.
Naturalpopulations
Fresh collections of the hybrid swarm’s two wild progenitor populations (SD and SC) were made in
March 2014. Populations were maintained in 1L and 400mL beakers at20

C on a 12/12 light/dark cycle with
live algal cultures (Platymonas and Monochrysis, Carolina Biological), dried food (0.1 g/L each Tetramin
and Spirulina, companies), and 3x-filtered (37um) seawater obtained from the Wrigley Marine Science
Center (Catalina Island, CA).
SNP-typing
DNA extraction and genotyping of samples from the hybrid population from 3–21 months of admixture
is described in Pritchard and Edmands (2013). Briefly, individuals were frozen at80

C; DNA was later
extracted in 50uL of a homemade lysis buffer supplemented with proteinase K, with a 1-hour incubation at
65

C and a 15-minute proteinase K inactivation period at 95

C. 20uL of each sample was dried down for
106
4 hours at 65

C and shipped to the Roswell Park Cancer Institute (Buffalo, NY) for Sequenom iPLEX
Gold SNP-typing assays. Genotyping was conducted using 45 SC-SD population–diagnostic nuclear SNPs
that were mapped to chromosomes using F
2
crosses (originally in Pritchard et al. 2011 and updated in Foley
et al. 2011).
DNA extraction and genotyping were performed as described above on hybrid population individuals
collected at 48 and 51 months after population establishment but utilized a larger panel of 190 mapped,
SC-SD population–diagnostic SNPs used in several previous studies (Pritchard and Edmands 2013, Foley
et al. 2011, 2013). Loci that were genotyped in less than 90% of individuals and individuals genotyped at
less than 80% of loci were discarded as low quality (3 loci and 1 female), for a total of 39 females and 40
males sampled at 48 months of admixture, and 48 females and 50 males sampled at 51 months of admixture,
and 187 population–diagnostic loci.
All statistical analyses were conducted in R (Team 2015) unless otherwise noted. Fisher’s exact tests
were used to calculate allele frequency differences between males and females and between timepoints.
10% false discovery rate corrections were conducted using the R package fdrtool (Klaus and Strimmer
2015). Principal components analysis (PCA) was performed with the R package SNPRelate (Zheng et al.
2012). Association testing was performed in PLINK (Purcell et al. 2007) using the assoc function with
10,000 permutations (option mperm).
Familysexratioassays
Family sex ratios can be used to infer the sex determination mechanism in a population. If sex chromo-
somes are present, family sex ratios are expected to fit a binomial distribution; however, under polygenic
sex determination, family sex ratios will exhibit extra–binomial variation. Male and female Tigriopus cali-
fornicus are dimorphic, with males having geniculate antennules used in clasping immature females during
pre-copulatory mate-guarding. Females brood a single egg sac that is attached beneath the urosome for 3-5
days, until it hatches. To assay family sex ratios, clasped pairs were removed from culture into individual
60mm Petri dishes with 3x filtered seawater and dry food. Dishes were checked every 1-3 days. Parents
were removed and frozen (80

C) upon hatching of the first clutch of eggs, and offspring were pipetted in
sets of up to 10 animals to fresh 60mm Petri dishes. Offspring were fed and rehydrated weekly, and at 28-
days post-hatch, survivors were counted and individually assessed for sex. Up to 10 adults (equal numbers
of males and females where possible) were frozen (80

C) per family for later sequencing.
107
Family estimates of survivorship and sex ratio were obtained by pooling across replicate dishes from
the same clutch. Survivorship was measured as the total number of surviving individuals (adult males, adult
females, and copepodids) divided by the total number of larvae produced. Family sex ratios were calculated
as the number of males divided by the total number of surviving adults.
Populations with polygenic sex determination are expected to exhibit extra-binomial variation in family
sex ratios, while those with sex chromosomes are expected to produce families whose sex ratios fit a bino-
mial distribution. We performed a quantile-quantile regression on family sex ratio for each population, as
compared to an idealized binomial distribution sampled from 10,000 random draws for the median clutch
size with a 0.5 probability of success (maleness). A perfect correspondence (binomial distribution of family
sex ratios for a population) would thus fall on a line with a slope of 1 and a y-intercept of 0.
Whole–genomere-sequencing
DNAextraction
All previous investigations of sex determination mechanisms in T. californicus have utilized genetic
markers that are fixed between, and thus invariant within populations. Whole–genome re-sequencing of
the hybrid population and its parental progenitor populations thus would provide the first genomic insight
into how males and females within T. californicus populations differ. DNA was extracted from previously
frozenTigriopuscalifornicus by salt precipitation followed by a phenol-chloroform cleanup. For individual
extractions, single animals frozen in 1.5mL tubes were crushed with a mortar and pestle in 50uL homemade
lysis buffer. Molecular-grade glycogen (20mg/mL) and filter-sterilized sodium acetate (3M; pH 5.2) were
added, and DNA was precipitated with 95% ethanol, incubated on ice, spun down, and the supernatant
discarded. A 70% ethanol wash was performed, and samples were dried overnight before resuspension
in 20uL 0.5x TE (pH 8.0) and incubation at 37

C for at least one hour to ensure full resuspension. For
phenol-chloroform purification, total sample volume was brought up to 100uL before addition of 25:24:1
phenol:chloroform:isoamyl alcohol. Samples were mixed, spun down, and the aqueous layer removed. This
was ethanol-precipitated, spun down and the supernatant removed, then washed with 70% ethanol, dried
overnight, resuspended in 20uL 0.5x TE (pH 8.0), and incubated at 37

C for at least one hour to ensure full
resuspension.
T. californicus animals are small (1mm long) and yield little DNA per individual, making library
108
preparation of many individuals prohibitively expensive. Therefore, we utilized a pooled sequencing ap-
proach to acquire unbiased estimates of allele frequencies in males and females from the hybrid population
and the parental populations. For bulk extractions, individually-frozen copepods were pooled into a single
tube and phenol-chloroform extracted followed by an ethanol precipitation with glycogen. Ten of the 12
pools were composed of offspring from the family sex ratio assays, stratified by parental genotype on chro-
mosome 10 (3 male and 3 female pools from the hybrid swarm, and one male and one female pool from
each parental population, with 50 animals per pool). Because these pools included siblings and results could
thus possibly be confounded by population structure, we included two additional pools (48 animals each)
that were composed of copepods sampled randomly from the hybrid swarm population after 78 months of
admixture (approximately 13 months prior to the commencement of the family sex ratio assays).
Fragmentanalysis
For the hybrid swarm, inclusion of family sex ratio assay offspring in either a bulk extraction for pool-seq
or an individual library prep was based on parental genotype at chromosome 10, as assessed in a fragment
analysis assay. Family sex ratio assay parents were individually extracted as above and amplified at a single
locus on chromosome 10, designated scaf-76 because of its location in the T. californicus draft genome.
This locus contains a 15bp insertion that is present at a high frequency in the SC population (SD reads
are 100% ’C’; 85.71% of SC reads are ’CAAAGTTATGTTGTCA’, 14.29% are ’C’). For each parent, the
locus was PCR-amplified with the primers scaff76-F (TGTAAAACGACGGCCAGTCCTGCCCTTGGT-
CATAGCAA), scaff76-R (GGGCGATAATTGTGACTCCG) and an M13 forward primer that also included
the scaff76-F primer sequence and a 5’ 6-FAM fluorescent label (TGTAAAACGACGGCCAGT; MWG Eu-
rofins Genomics) as in Schuelke (2000). To ensure that all forward-primed molecules were incorporated into
fluorescently tagged fragments, scaff76-F was added at half the volume of M13 and scaff76-R. Reactions
were amplified using Choice Taq (Denville Scientific, Inc.) with an initial denaturing period of 94

C for
5 minutes; followed by 10 cycles of touch-down PCR with 94

C for 45 seconds, 55

C for 1 minute (de-
creasing by 0:5

C every cycle), 72

C for 2 minutes; 25 rounds of 94

C for 45 seconds, 50

C for 1 minute,
72

C for 2 minutes; and a final extension time of 72

C for 10 minutes. Fluorescently labeled PCR products
were checked for amplification with gel electrophoresis and then subjected to fragment analysis via capillary
electrophoresis on an ABI 3730 (Applied Biosystems) at the University of Arizona Genetics Core Facility
and analyzed using PeakScanner v1.0 (Thermo Fisher Scientific). Offspring were then grouped based on
109
Table 1: Chromosome 10 indel (scaf-76) genotype, survivorship, and sex ratios of families sequenced indi-
vidually
Population Family Maternal genotype, scaf-76y Paternal genotype, scaf-76y Survivorship Sex ratio Mortality-corrected sex ratioyy Family sex ratio bias
Hybrid E4-1 CC CC 0.86 0.35 0.44 unbiased
Hybrid E4-11 DD null 0.84 0.19*** 0.32* female-biased
Hybrid E4-15 CD CC 0.97 0.24** 0.26** female-biased
Hybrid E4-36 CC CC 0.90 0.61 0.55 unbiased
Hybrid E4-37 DD null 0.70 0.37 0.57 unbiased
Hybrid E4-61 CD null 0.85 0.00*** 0.15*** female-biased
SC SC-4 — — 0.82 0.49 0.64* unbiased
SC SC-21 — — 0.62 0.27** 0.55 female-biased
SC SC-70 — — 0.86 0.65* 0.55 male-biased
SD SD-6 — — 0.91 0.74** 0.67* male-biased
SD SD-55 — — 0.91 0.49 0.53 unbiased
SD SD-58 — — 0.76 0.08*** 0.29* female-biased
Chi-square test for deviation from an equal sex ratio: *p-value< 0:05, **p-value< 0:01, ***p-value< 0:001
yGenotyping was performed only on hybrid parents; alleles that failed to amplify are designated “null”, families that were not genotyped are indicated by “—”;
CC: SC allele homozygote; CD: heterozygote; DD: no SC allele present
yyMortality correction for family sex ratio assigns all missing animals to the rarer sex
parental genotype—because the indel is not fixed and because the smaller fragment amplifies preferentially,
offspring were grouped by whether their mothers were typed as homozygous SC (CC; no small allele), or
homozygous or heterozygous SD (CD or DD; one or possibly two copies of the small allele). Thus, bulk
extractions were created by grouping offspring of mothers with 0-1, 1, or 1-2 SD (small) alleles. Hybrid
families for individual sequencing were also chosen based on this assay, with two families (mother, father,
daughter, son) from each group selected randomly (Table 1). For parental populations, families were ran-
domly chosen for individual sequencing from high–survivorship (> 0:50) families. Offspring included in
individual sequencing and bulk extractions for all three populations were allocated within families using a
random number generator.
Librarypreparation
DNA extractions were sheared to an average size of 500bp using a Covaris S2 according to manu-
facturer guidelines. Both pooled and individual libraries were prepared using the NEBNext Ultra kit and
NEBNext Oligos for Multiplexing, according to the manufacturer’s instructions. The pooled libraries were
size-selected with Ampure XP beads and amplified with 6 rounds of PCR. Due to the low input from indi-
vidual samples (generally 20–50ng per individual), individual libraries were not size selected. All but two
individual libraries were amplified with 6 rounds of PCR (both individually-sequenced E4-61 offspring were
amplified with 10 rounds of PCR). All libraries were quantified by Qubit assays (Life Technologies/Thermo
110
Fisher Scientific) and a random subset were run on a Bioanalyzer high sensitivity chip (Agilent Technolo-
gies, Inc.) to estimate fragment size for equimolar pooling. The 12 pool-seq libraries were each sequenced
on 2 lanes of an Illumina HiSeq 2500 with 110bp paired-end reads. The individual libraries were pooled
23-24 per lane and run on two lanes of an Illumina HiSeq 2500 with 110bp paired-end reads.
Sequenceanalysis
Pooledsamples
Poolseq samples were processed for Phi-X and adapter contamination with BBDuk and mapped back
to the Tigriopus californicus draft genome (reference population SD, v3) without quality trimming using
BBMap 35.43 (B. Bushnell, sourceforge.net/projects/bbmap/). For each library, lanes were de-contaminated
and mapped separately, to assess technical variation. Output sam files were compressed, sorted, indexed, and
filtered for quality and duplicates with Samtools and Picard-Tools (Li 2011; http://broadinstitute.github.io/picard).
Realignment around indels was performed in GATK according to the Best Practices guide v3.5 (McKenna
et al. 2010, DePristo et al. 2011, Van der Auwera et al. 2013), and bams were converted into pileup and sync
files using Samtools, and PoPoolation2 v1.201 (Kofler et al. 2011). SC reads map poorly to the SD refer-
ence genome due to high divergence, so we restricted our analyses to genes annotated in the T. californicus
SD reference genome using the create-genewise-sync function in PoPoolation2. Cochran-Mantel-Hanzel
(CMH) tests were performed in PoPoolation2 to search for differences in SNP allele frequencies between
males and females across sequencing lanes for each set of pools, with a minimum coverage of 4 reads per
sample per lane and a minimum count of 2 reads to consider a minimum allele. FDR cutoffs were determined
using the R package fdrtool (Klaus and Strimmer 2015).
Individualsamples
Individual libraries were processed as described above. Joint genotyping was performed on all individ-
uals from all 3 populations together using GATK HaplotypeCaller, followed by hard filtering according to
the GATK Best Practices recommendations. PCA for the filtered set of SNPs and indels was conducted for
all samples and for the samples from each population alone using R/SNPRelate.
In species with degenerated sex chromosomes, coverage will differ between males and females on the
sex chromosomes but not the autosomes (Vicoso et al. 2013, Vicoso and Bachtrog 2015). For each individual
sample, we calculated normalized log2 coverage per scaffold (normalized to scaffold length) using bamtools
111
(Barnett et al. 2011). Thus, expected normalized coverage on autosomes is 0 and on sex chromosomes is -1.
Figure 1: Mean SD allele frequency per chromosome per month in males (orange triangles) and females
(purple circles) in the hybrid population over its first 21 months of admixture. Error bars indicate standard
error. Asterisks identify chromosomes with1 SNP where allele frequencies differ between males and
females (Fishers exact test, 10% FDR correction). Data from (Pritchard and Edmands 2013).
Results
Population–diagnosticgenotyping
SDallelesonchromosome10havebeenmaintainedinexcessinfemalesthroughoutthehybridpopu-
lation’sexistence
Population–diagnostic genotyping of the hybrid population was previously conducted on samples taken
from 3 to 21 months of admixture using a set of 45 SNPs (1–6 per chromosome; Pritchard and Edmands
2013). We re-analyzed this data to examine how allele frequencies changed in males and females over time.
112
Table 2: Number of loci with SD allele frequency changes (Fisher’s exact test) between 3 and 21 months of
admixture in the hybrid population, for each sex and chromosome.
Females Males
Chromosome Decrease No change Decrease No change
1 1 5 0 6
2 4 1 0 5
3 3 0 1 2
4 2 0 0 2
5 6 0 0 6
6 2 0 0 2
7 5 0 2 3
8 2 3 0 5
9 2 0 0 2
10 0 6 2 4
11 2 0 0 2
12 1 0 1 0
For every chromosome, the mean SD allele frequency was initially higher in females than in males (Figure
1). After FDR correction, the frequency of SD alleles was significantly different between males and females
on at least one SNP on chromosomes 1, 3, and 5 at month 3, on chromosomes 6 and 10 at month 6, and
only on chromosome 10 at months 9 and 21. In all cases, the allele frequency differences were due to higher
levels of SD ancestry in females than in males.
In addition to comparing allele frequencies between males and females, we compared allele frequencies
within each sex between the initial (3 month) and final (21 month) time points. In females, SD allele fre-
quencies decreased significantly at 1 locus per chromosome on all chromosomes except chromosome 10
(Fisher’s exact test, 6:22x10
5
p 1 for chromosomes 1-9, 11-12 and 0:804p 1 for chromosome
10; Table 2). In males, where initial SD allele frequencies were lower, significant decreases occurred for
loci on chromosomes 3, 7, 10, and 12 (p = 0:024). SD alleles did not increase in frequency at any locus in
either sex in the hybrid population.
SDancestryonchromosome10isstronglyassociatedwithsexafter4yearsofadmixture
Population–diagnostic genotyping of the hybrid swarm confirmed that the SC mitochondrial genome has
fixed in this population (as noted in Pritchard and Edmands (2013)) and that SC alleles are at or near fixation
113
Table 3: Mean SD allele frequency per chromosome (standard deviation), population–diagnostic SNPs
48 months 51 months Combined
Chromosome SNPs Females(41) Males(42) Females(48) Males(48) Females(89) Males(90)
1 18 0.09 (0:05) 0.08 (0:04) 0.11 (0:04) 0.08 (0:04) 0.10 (0:04) 0.08 (0:03)
2 27 0.04 (0:03) 0.07 (0:04) 0.04 (0:03) 0.03 (0:03) 0.04 (0:03) 0.05 (0:03)
3 21 0.05 (0:05) 0.03 (0:05) 0.04 (0:03) 0.03 (0:04) 0.04 (0:04) 0.03 (0:04)
4 8 0.10 (0:11) 0.15 (0:13) 0.12 (0:10) 0.16 (0:15) 0.11 (0:10) 0.16 (0:14)
5 20 0.00 (0:01) 0.01 (0:01) 0.03 (0:02) 0.02 (0:02) 0.02 (0:01) 0.01 (0:01)
6 9 0.02 (0:03) 0.02 (0:03) 0.02 (0:05) 0.01 (0:04) 0.02 (0:04) 0.01 (0:03)
7 17 0.01 (0:03) 0.02 (0:04) 0.02 (0:04) 0.02 (0:03) 0.02 (0:03) 0.02 (0:04)
8 18 0.06 (0:04) 0.06 (0:03) 0.07 (0:03) 0.06 (0:03) 0.07 (0:03) 0.06 (0:03)
9 16 0.02 (0:05) 0.02 (0:05) 0.03 (0:06) 0.03 (0:06) 0.03 (0:05) 0.03 (0:05)
10 16 0.28 (0:06) 0.04 (0:06) 0.31 (0:05) 0.04 (0:08) 0.30 (0:05) 0.04 (0:04)
11 10 0.02 (0:07) 0.02 (0:06) 0.03 (0:07) 0.03 (0:08) 0.03 (0:07) 0.03 (0:07)
12 8 0.00 (0:00) 0.00 (0:01) 0.00 (0:00) 0.00 (0:00) 0.00 (0:00) 0.00 (0:00)
Mt 2 0.00 (0:00) 0.00 (0:00) 0.00 (0:00) 0.00 (0:00) 0.00 (0:00) 0.00 (0:00)
across most of the nuclear genome (Table 3). Strikingly, in the face of near-complete genetic swamping of
the SC allele across the genome, SD alleles on chromosome 10 were present and largely unrecombined
in many females (Figure 2). Association tests and Fisher’s exact tests for SNPs statistically associated
with sex identified all loci on chromosome 10—and only loci on chromosome 10—as highly significant
after permutation testing in 48-month and the 51-month samples, and when the timepoints were combined
(Figure 2, only combined sample is shown).
Under principal components analysis, PC2, which explained 8.14% of the observed variance, separated
males and females, and was mostly underlain by SNPs on chromosome 10 (Figure 3). PC1 explained more
variance (9.92%) and was primarily underlain by SNPs on chromosome 2, but only separated a few rare
individuals (both males and females) who had retained large linkage blocks of SD ancestry on chromosome
2. Identity-by-state analysis, a measure of pairwise individual similarity, split the sample into two groups,
who clustered by sex and chromosome 10 ancestry.
Survivorshipandfamilysexratios
Clutch size differed between populations (ANOV A,p < 0:001), with SD having significantly smaller
clutch sizes than both the hybrid population and SC (Tukey’s HSD,p = 0:002 andp< 0:001, respectively).
114
Table 4: Survivorship and family sex ratios of hybrid swarm and parental populations
Clutch size Survivorship Family sex ratio
Population n (Std:Dev:) n (Std:Dev:) n (Std:Dev:)
Hybrid 48 42 (19) 48 0.59 (0:28) 47 0.43 (0:25)
SC 37 46 (21) 37 0.49 (0:32) 30 0.34 (0:20)
SD 37 28 (14) 37 0.44 (0:37) 27 0.41 (0:20)
Table 5: Genotype frequencies of parents from hybrid swarm family sex ratio assays at indel scaf-76 on
chromosome 10.
Females Males
Genotype n Frequency n Frequency
SC/SC 5 0.16 26 0.64
SC/SD 11 0.34 0 0
SD/SD 13 0.41 1 0.03
Null 3 0.09 11 0.33
Survivorship and family sex ratio, however, were not significantly different between populations (ANOV A,
p = 0:08 andp = 0:24, respectively). We had predicted that in the presence of a sex chromosome, family
sex ratios in the hybrid swarm would conform to a binomial distribution. We tested this hypothesis by
performing random sampling of clutches from a binomial distribution and plotting these against the actual
family sex ratios for the hybrid and parental populations. However, the distribution of family sex ratios in
the hybrid swarm did not differ from those of the parental populations, and none fit a binomial distribution
(Figure 4).
Fragmentanalysis
Fragment analysis was used to genotype sex ratio assay parents from the hybrid swarm for an indel
on chromosome 10. Allele frequencies were highly skewed in both males and females, though in opposite
directions, and amplification at this locus failed more often in males than in females (Table 5). The SC
allele produced less robust amplifications than the SD allele (as a function of both gel band brightness
and of relative peak height within heterozygotes; data not shown), and many of the missing genotypes can
probably be attributed to the poor amplification of this allele.
115
A
B
Figure 2: Population–diagnostic genotyping of the experimental hybrid population after 4 years of admixture
revealed sex-specific heterozygosity across a single chromosome, which is the only region of the genome
that is strongly associated with sex. (A) SD allele frequency at 187 parental-population–diagnostic SNPs
(females: purple; males: orange). Many hybrid population females have retained SD ancestry across the
entire mapped length of chromosome 10 despite genetic swamping of SC alleles across the remainder of
the genome. Dashed lines indicate chromosome boundaries. (B). Association tests for sex find a strong
effect of chromosome 10 and only of chromosome 10 after 10,000 permutations. Each SNP is colored by
chromosome. Black line indicates p = 0.05.
116
A B
C
Figure 3: Principal components and cluster analysis of population–diagnostic SNPs. A. Principal compo-
nents plot for the top two eigenvectors. PC 2 separates males from some females. Black: females; red:
males. B. Correlation coefficient for each SNP for the top two eigenvectors. Correlation coefficients for PC
2 are largest for chromosome 10. SNPs are colored by chromosome. C. Identity-by-state cluster analysis
groups females with SD ancestry on chromosome 10 separately from all but one male. Black: female; red:
male.
117
Figure 4: Quantile-quantile family sex ratio plot. Family sex ratios do not fit a binomial distribution–
as would be expected under chromosomal sex determination–in either the hybrid swarm or the parental
populations. Plot depicts actual versus expected family sex ratios under a binomial model with all theoretical
family sizes equal to the average clutch size for that population. Green: parental populations (SC and SD
combined); purple: hybrid population. Dashed gray line indicates the expectation when family sex ratios
are binomially distributed (as with chromosomal sex determination).
118
Table 6: Read number and mapping statistics, pools
Total Mapped as Mapped as Average Total Reads Reads Average Scaffolds with Reference
Phi-X Adapter contaminant mated incongruent insert size mapped mapped mapped coverage > 1 base bases covered
Population Group Sex Lane Total reads reads reads reads pairs pairs (std:dev:)* reads unambiguously ambiguously (std:dev:) covered > 1x
SC Pop Fem 1 25,100,726 280,814 161,293 1:18% 52:01% 11:86% 274 (319) 18,230,949 (73:50%) 63:46% 10:03% 10:8(186:74) 76:36% 93:60%
SC Pop Fem 2 26,332,572 307,148 162,550 1:23% 53:08% 12:28% 278 (330) 19,354,046 (74:41%) 64:30% 10:11% 11:46(189:56) 76:36% 93:84%
SC Pop Male 1 23,311,594 265,146 116,314 1:20% 47:19% 10:48% 287 (324) 16,196,665 (70:32%) 60:55% 9:77% 9:6(182:91) 76:03% 92:83%
SC Pop Male 2 24,414,730 284,864 117,458 1:22% 49:77% 11:25% 294 (328) 17,440,999 (72:32%) 62:38% 9:944% 10:34(185:96) 76:29% 93:25%
SD Pop Fem 1 17,157,184 294,294 64,492 1:75% 79:28% 7:73% 311 (229) 15,205,468 (90:20%) 84:21% 6:00% 9:01(140:17) 75:05% 95:80%
SD Pop Fem 2 18,090,958 317,506 64,913 1:79% 79:80% 7:87% 315 (232) 16,085,103 (90:53%) 84:54% 5:99% 9:53(144:84) 75:19% 95:94%
SD Pop Male 1 13,382,612 188,468 50,528 1:44% 74:74% 6:47% 325 (230) 11,528,220 (87:40%) 81:36% 6:04% 6:83(130:12) 75:01% 94:86%
SD Pop Male 2 14,178,108 204,970 50,062 1:48% 77:08% 6:78% 331 (236) 12,396,811 (88:75%) 82:70% 6:04% 7:35(134:99) 74:5% 95:24%
Hybrid Pop Fem 1 16,434,808 243,190 116,813 1:52% 43:36% 8:36% 276 (311) 10,324,863 (63:79%) 54:98% 8:82% 6:11(162:72) 76:36% 90:61%
Hybrid Pop Fem 2 17,304,706 263,828 119,148 1:56% 46:22% 9:02% 284 (325) 11,218,094 (65:85%) 56:87% 8:99% 6:64(165:27) 76:07% 91:32%
Hybrid Pop Male 1 10,902,090 146,996 94,753 1:43% 49:86% 11:42% 288 (326) 7,496,660 (69:76%) 60:10% 9:66% 4:44(132:34) 75:56% 87:91%
Hybrid Pop Male 2 11,442,700 157,444 95,840 1:46% 50:52% 11:72% 292 (338) 7,934,215 (70:36%) 60:67% 9:70% 4:7(135:51) 75:56% 88:72%
Hybrid CC Fem 1 14,265,042 226,552 97,181 1:65% 51:86% 10:65% 270 (315) 10,118,462 (72:12%) 62:26% 9:86% 5:99(158:72) 75:3% 90:63%
Hybrid CC Fem 2 14,957,294 246,942 98,882 1:70% 53:52% 11:14% 275 (324) 10,787,201 (73:37%) 63:41% 9:96% 6:39(161:47) 75:3% 91:33%
Hybrid CC Male 1 17,235,238 202,098 89,523 1:22% 50:17% 11:78% 286 (319) 12,428,432 (73:00%) 62:89% 10:11% 7:36(169:26) 75:92% 91:91%
Hybrid CC Male 2 18,067,264 214,998 90,804 1:23% 51:98% 12:39% 291 (329) 13,293,364 (74:50%) 64:25% 10:24% 7:88(171:71) 75:81% 92:61%
Hybrid CD Fem 1 16,785,654 211,894 119,106 1:31% 55:28% 10:82% 272 (313) 12,463,285 (75:24%) 65:18% 10:05% 7:38(168:10) 75:63% 92:14%
Hybrid CD Fem 2 17,555,436 227,578 119,716 1:34% 56:89% 11:24% 277 (322) 13,243,990 (76:47%) 66:33% 10:14% 7:84(170:26) 75:96% 92:69%
Hybrid CD Male 1 11,375,934 154,634 114,087 1:45% 53:17% 11:76% 278 (328) 8,365,209 (74:62%) 64:55% 10:07% 4:95(133:90) 75:23% 89:14%
Hybrid CD Male 2 11,951,934 163,826 115,477 1:46% 54:37% 12:21% 283 (330) 8,910,631 (75:66%) 65:49% 10:17% 5:27(137:58) 75:01% 89:94%
Hybrid DD Fem 1 15,502,328 214,080 97,934 1:44% 55:80% 11:38% 277 (323) 11,489,283 (75:20%) 65:20% 10:00% 6:8(163:17) 75:81% 91:87%
Hybrid DD Fem 2 16,268,962 228,356 98,542 1:46% 56:52% 11:67% 280 (316) 12,153,857 (75:81%) 65:77% 10:04% 7:2(165:79) 75:26% 92:22%
Hybrid DD Male 1 42,876,794 495,186 198,229 1:21% 50:88% 12:09% 296 (328) 31,240,038 (73:76%) 63:56% 10:20% 18:51(232:70) 76:50% 95:42%
Hybrid DD Male 2 45,040,764 535,900 199,190 1:25% 52:45% 12:67% 301 (336) 33,396,281 (75:08%) 64:78% 10:30% 19:79(236:38) 76:58% 95:70%
119
Sequencing
Pools
Twelve bulk-extracted prepared libraries were pooled and subjected to 110bp paired-end sequencing on
two lanes of an Illumina HiSeq 2500, generating 459,935,432 total reads (10,902,090 to 45,040,764 reads
per lane per library, or 22,344,790 to 87,917,558 reads per library; Table 6). Reads from each library and
lane were processed separately to assess biological and technical variation. Reads were filtered for Phi-X
and adapter contaminants and then mapped to the reference without quality trimming. Contaminant filtering
removed 155,728 to 561,282 reads (1.18 to 1.79 percent) per library per lane. The reference sequence
was built from the SD population, and, overall, reads from SD pools mapped more successfully to the
reference than reads from E4 or SC pools. Average coverage per lane ranged from 4.44x (132.34) to
19.79x (236.38) for E4 libraries, from 9.6x (182.91) to 11.46x (189.56) for SC libraries, and from
6.83x (130.12) to 9.53x (144.84) for SD libraries. More reads mapped as mated pairs and mapped
unambiguously in SD than in either the hybrid population or SC (ANOV A,p< 0:001 for both and Tukey’s
HSD tests, p < 0:001 for SC-SD and hybrid-SD, p = 0:99 and p = 0:69 for SC-SD). The number of
reference bases covered by at least 1 mapped read was higher in SD than in the hybrid population, but not
for other comparisons (ANOV A,p = 0:002 and Tukey’s HSD,p = 0:002 for SD-hybrid andp = 0:18 for
SC-hybrid andp = 0:26 for SC-SD), and the number of scaffolds with any coverage was higher in SD than
in either SC or the hybrid population (ANOV A, ,p < 0:001 and Tukey’s HSD,p = 0:006 for SD-hybrid
andp< 0:001 for SC-SD butp = 0:93 for SC-hybrid).
Reads from each pool mapped disproportionately to the assembled chromosomes, relative to the length
of each chromosome assembly (
2
tests for each pool set, all p < 0:001), with the largest deviations
on chromosome 3 in all populations with SC ancestry and on chromosome 12 in all populations (Figure
5). Despite the differential introgression of SD alleles across the genome that was previously observed
(Figure 2), reads did not map disproportionately to chromosomes with greater amounts of SD ancestry (e.g.,
chromosome 10). Rather, SC and hybrid reads mapped with slight excess to chromosome 3, which has
retained little SD ancestry (Figure 2), and proportionally more reads than expected mapped to chromosome
12 for all populations.
120
Individuals
Libraries from 47 individuals (Table 1) were sequenced at 23 or 24 libraries per lane with paired-end
110bp reads on an Illumina HiSeq2500, generating a total of 309,331,022 reads (between 1,549,812 and
27,376,138 reads per individual; Table 7). Phi-X and adapter contaminant filtering removed from 5,132
to 85,286 reads (0.09 to 2.33 percent). As in the pooled sequences, hybrid and SC libraries mapped to
the reference with less success than reads derived from SD individuals. More reads from SD individuals
mapped as mated pairs than did reads from hybrid or SC individuals (ANOV A, p < 0:001 and Tukey’s
HSD,p < 0:001 for SD-hybrid andp < 0:001 for SD-SC andp = 0:56 for SC-hybrid). The proportion
of reads that mapped unambiguously was also greater in SD than SC and the hybrid population (ANOV A,
p < 0:001 and Tukey’s HSD,p < 0:001 for SD-hybrid andp < 0:001 for SD-SC andp = 0:68 for SC-
hybrid). Though the percent of reference bases covered also differed between SD individuals and those from
the other populations (ANOV A,p< 0:001 and Tukey’s HSD,p< 0:001 for SD-hybrid andp< 0:001 for
SD-SC andp = 0:96 for SC-hybrid), the percent of covered scaffolds did not differ between populations
(ANOV A,p = 0:34), perhaps because average coverage per individual was low.
121
A B
C D
E F
Figure 5: Reads map disproportionately to chromosome 3 in hybrids and SC and to chromosome 12 in SD,
relative to assembled chromosome length. Each bar represents the percent difference between the proportion
of total reads unambiguously mapped to a chromosome per pool set (males and females summed across
sequencing lanes) and the proportion of the total assembled, mapped genome assigned to each chromosome.
Chi-squarep< 0:001 for all pool sets. A,B,C,E. Experimental hybrid population. D. Parental population
SC. F. Parental population SD. A, C, E. Pools composed of 50 siblings from families where mothers had
with 0-1 (A), 1 (C), or 2 (E) SD alleles at the scaf-76 indel on chromosome 10. B. Unrelated population
sample of hybrid population. Bars are colored by chromosome.
122
Table 7: Read number and mapping statistics, individuals.
Total Total Mapped as Mapped as Average Scaffolds Reference
Total Phi-X Adapter contam mapped mated incongruent Mapped Mapped insert size Average with> 1 bases
Pop. Fam. Gen.* Sex reads reads reads reads reads pairs pairs unambig ambig (std:dev:) coverage base covered covered 1x
Hybrid 1 P F 3783866 5914 12949 0:20% 2308742(61:14%) 46:37% 9:25% 53:18% 3:94% 352:72(383:34) 1:37(41:41) 71:18% 63:30%
Hybrid 1 P M 5727894 9188 43414 0:23% 3888365(68:04%) 52:33% 10:15% 58:99% 4:48% 318:65(372:78) 2:3(78:55) 72:46% 76:65%
Hybrid 1 O F 2202846 9226 14258 0:51% 1659863(75:73%) 57:50% 12:31% 64:94% 5:36% 311:24(351:5) 0:98(47:76) 69:68% 51:06%
Hybrid 1 O M 2718982 5650 3786 0:24% 2035758(75:05%) 56:76% 11:57% 65:20% 4:87% 366:45(395:24) 1:21(39:57) 70:23% 59:71%
Hybrid 11 P F 3359038 18376 22466 0:63% 2367753(70:94%) 55:17% 9:86% 62:19% 4:33% 331:59(370:19) 1:4(38:13) 70:92% 64:25%
Hybrid 11 P M 2451066 21122 16747 0:99% 1800557(74:20%) 56:56% 11:57% 63:94% 5:07% 317:52(370) 1:07(43:08) 70:08% 54:18%
Hybrid 11 O M 2456056 35042 11891 1:51% 1860659(76:92%) 59:19% 11:67% 66:62% 5:11% 313:46(361:96) 1:1(43:02) 70:30% 55:62%
Hybrid 11 O F 5062404 18570 35458 0:44% 3476054(68:97%) 53:80% 9:70% 60:24% 4:32% 307:65(365:91) 2:06(67:09) 72:27% 74:27%
Hybrid 15 P F 2148538 13482 13655 0:73% 1625257(76:20%) 59:36% 10:75% 66:73% 4:68% 342:17(387:15) 0:96(27:59) 70:01% 52:92%
Hybrid 15 P M 1976988 4596 7915 0:31% 1369596(69:49%) 52:73% 10:78% 60:25% 4:56% 353:36(374) 0:81(29:3) 69:03% 46:75%
Hybrid 15 O F 3178880 6820 23557 0:27% 2382355(75:15%) 57:93% 11:67% 64:70% 5:16% 304:41(325:51) 1:41(64:39) 71:36% 62:09%
Hybrid 15 O M 1549812 8826 9958 0:69% 1203402(78:19%) 60:50% 11:52% 67:99% 5:05% 321:34(370:4) 0:71(23:3) 68:44% 43:3%
Hybrid 36 P F 3667054 83890 13318 2:33% 2783997(77:73%) 60:02% 11:36% 67:78% 4:92% 332:14(372:02) 1:65(50:05) 71:8% 68:69%
Hybrid 36 P M 1675880 7186 7031 0:52% 1285839(77:13%) 58:60% 11:94% 66:80% 5:12% 359:97(384:03) 0:76(26:39) 69:61% 45:32%
Hybrid 36 O F 2711456 27758 10345 1:09% 2058183(76:74%) 58:05% 11:87% 66:59% 5:02% 358:07(391:74) 1:22(40:06) 70:45% 59:59%
Hybrid 36 O M 1582082 4588 6386 0:37% 1225942(77:78%) 59:16% 11:81% 67:60% 5:03% 357:4(382:05) 0:73(22:55) 68:99% 44:21%
Hybrid 37 P F 2980226 6658 21736 0:29% 2314875(77:90%) 60:11% 11:58% 67:63% 5:09% 324:35(368:43) 1:37(50:1) 71:29% 62:52%
Hybrid 37 P M 3178928 10050 17575 0:37% 2455172(77:52%) 59:43% 11:56% 67:41% 5:01% 338:87(377:73) 1:46(46:44) 71:54% 64:83%
Hybrid 37 O F 27376138 23646 117130 0:12% 21385030(78:21%) 59:52% 11:97% 67:86% 5:12% 345:07(383:98) 12:68(202:78) 75:81% 93:66%
Hybrid 37 O M 3212678 21012 16597 0:68% 2507518(78:59%) 61:64% 10:81% 68:85% 4:82% 317:06(370:22) 1:49(38:07) 71:69% 65:81%
Hybrid 61 P F 2854182 47186 11896 1:73% 1904115(67:89%) 52:18% 10:11% 59:12% 4:34% 349:48(368:46) 1:13(40:52) 70:52% 57:06%
Hybrid 61 P M 3823082 8646 22936 0:31% 2754958(72:28%) 55:28% 10:71% 62:87% 4:67% 341:05(367:8) 1:63(52:01) 72:38% 68:14%
Hybrid 61 O F 14700568 11612 23367 0:14% 11243024(76:58%) 58:37% 11:56% 66:77% 4:85% 378:23(390:75) 6:67(154:89) 74:43% 91:45%
Hybrid 61 O F 12355712 16518 27293 0:21% 9462191(76:74%) 58:04% 11:91% 66:67% 4:99% 371:64(390:18) 5:61(145:41) 74:86% 90:24%
SC 4 P F 2355786 15158 14875 0:74% 1374675(58:79%) 45:01% 8:80% 51:06% 3:82% 329:34(386:72) 0:81(27:22) 71:03% 46:89%
SC 4 P M 9528588 52700 46307 0:58% 6010545(63:45%) 48:17% 9:81% 54:89% 4:24% 338:4(382:57) 3:56(108:28) 80:85% 85:15%
SC 4 O F 3142294 8594 17290 0:31% 2443458(78:01%) 59:61% 11:66% 67:80% 5:06% 325:33(369:09) 1:45(41:92) 70:96% 64:66%
SC 4 O M 2746034 4948 12431 0:23% 2116103(77:24%) 58:73% 12:15% 66:64% 5:26% 323:2(353:01) 1:25(50:25) 70:74% 59:49%
SC 21 P F 2826632 10292 14656 0:43% 2082983(74:01%) 56:15% 11:47% 64:05% 4:94% 336:39(344:04) 1:23(44:13) 71:51% 59:57%
SC 21 P M 3234006 44344 4212 1:40% 2136120(66:99%) 50:81% 10:16% 58:27% 4:31% 348:7(370:11) 1:27(39:01) 75:26% 61:34%
SC 21 O M 4130084 14158 24462 0:42% 3194817(77:68%) 59:52% 11:60% 67:42% 5:08% 319:82(371:34) 1:89(62:54) 72:16% 71:85%
SC 70 P F 2094320 9904 2830 0:50% 1558031(74:77%) 56:45% 11:44% 65:13% 4:78% 378:5(406:37) 0:92(26:31) 69:65% 52:19%
SC 70 P M 2729664 35928 11479 1:39% 1968959(73:14%) 55:38% 11:30% 63:44% 4:81% 348:63(375:72) 1:17(37:56) 70:96% 58:13%
123
Table 7. Read number and mapping statistics, individualscontinued.
Total Total Mapped as Mapped as Average Scaffolds Reference
Total Phi-X Adapter contam mapped mated incongruent Mapped Mapped insert size Average with> 1 bases
Pop. Fam. Gen.* Sex reads reads reads reads reads pairs pairs unambig ambig (std:dev:) coverage base covered covered 1x
SC 70 O F 2920622 13060 8935 0:49% 2259279(77:74%) 58:83% 11:84% 67:67% 4:98% 366:37(379:89) 1:34(39:69) 70:56% 63:09%
SC 70 O M 2532810 15574 10264 0:68% 1945752(77:35%) 58:72% 11:79% 67:16% 5:04% 347:73(381:59) 1:15(35:36) 70:74% 57:99%
SD 6 P F 2496008 5748 23088 0:35% 1740653(69:98%) 62:38% 6:22% 64:33% 2:79% 287:93(240:86) 1:03(58:45) 69:90% 53:56%
SD 6 P M 5568020 7186 20798 0:18% 4138023(74:45%) 68:30% 4:51% 69:74% 2:34% 354:49(283:92) 2:45(56:69) 76:03% 84:01%
SD 6 O F 15359354 14858 137159 0:14% 13626101(88:84%) 81:73% 5:30% 83:15% 2:82% 311:22(254:7) 8:07(130:74) 72:67% 95:35%
SD 6 O M 13524496 15584 108792 0:15% 11951392(88:50%) 79:18% 7:57% 81:57% 3:43% 304:59(254:51) 7:08(186:33) 72:16% 94:92%
SD 55 P F 26069540 19772 146613 0:11% 20875714(80:16%) 73:90% 4:44% 75:31% 2:40% 336:97(272:03) 12:37(141:55) 73:99% 95:86%
SD 55 P M 18397834 10666 101769 0:09% 14235671(77:45%) 71:71% 4:10% 72:79% 2:31% 323:26(263:58) 8:44(110:09) 72:97% 95:40%
SD 55 O F 7782420 14122 30055 0:21% 6892420(88:75%) 81:89% 4:90% 83:42% 2:65% 354:48(279:32) 4:09(67:74) 70:96% 92:34%
SD 55 O M 15193052 13428 81157 0:12% 13714326(90:37%) 83:45% 4:97% 84:95% 2:68% 348:44(275:8) 8:13(109:81) 71:91% 95:39%
SD 58 P F 15943102 35074 29129 0:24% 12954446(81:45%) 74:86% 4:80% 76:42% 2:49% 344:93(264:36) 7:68(114:54) 73:04% 95:23%
SD 58 P M 8918512 8516 56319 0:13% 6867210(77:10%) 71:48% 3:86% 72:69% 2:18% 346:83(283:08) 4:07(54:83) 71:58% 92:30%
SD 58 O F 1969274 3662 13008 0:26% 1737964(88:48%) 81:01% 5:52% 82:70% 2:87% 333:21(276:16) 1:03(26:76) 67:82% 56:74%
SD 58 O M 23134214 21570 179117 0:12% 20422960(88:39%) 81:42% 5:06% 82:84% 2:75% 311:34(258:99) 12:1(154:49) 73:33% 95:78%
*Generation. P: parent; O: offspring
124
Figure 6: Degenerated sex chromosomes are absent in both the parental populations and the hybrid pop-
ulation. Log2 normalized read coverage per scaffold per individual is plotted for all scaffolds mapped
to chromosomes. Lines represent GAM-smoothed averages across all samples. Purple: females; orange:
males.
DegeneratedsexchromosomesarenotpresentinT.californicus
If a degenerated sex chromosome is present anywhere in the T. californicus genome, fewer reads will
map to that chromosome in the hemizygous sex. Log2 normalized read counts per scaffold, then, would
be expected to equal 0 on autosomes and on the sex chromosome in the homogametic sex but 1 on the sex
chromosome in the heterogametic sex. No differences in read depth are present between males and females
for any chromosome in the parental populations or in the hybrid population, indicating that degenerated sex
chromosomes are absent inT.californicus (Figure 6).
125
Figure 7: Male-female allele frequency differences for each sequenced pool set. A, B. Pools composed of
50 siblings from families from parental populationA. SC andB. SD.C,D,E. Pools composed of 50 siblings
from families where mothers had 0-1 (A), 1 (C), or 1-2 (E) SD alleles at the scaf-76 indel on chromosome
10. F. Random sample of hybrid population. Black line indicates FDR cutoff.
Allelefrequencydifferencesbetweenmalesandfemalesareconcentratedonchromosome10
Cochran-Mantel-Haenszel tests were conducted for each set of pools to search for allelic differences be-
tween male and female pools that were controlled across sequencing lanes. Many SNPs on all chromosomes
exceeded the FDR-corrected significance threshold (Figure 7), but significant SNPs were concentrated on
scaffolds mapped to chromosome 10 in the hybrid population but not in the parental populations (Figure 8).
Additionally, though the total number of SNPs detected on chromosome 10 was slightly higher in pool
sets with SD ancestry on chromosome 10 (Hybrid 1, Hybrid 1/2, and Hybrid Pop), the proportion of SNPs
with significantly different allele frequencies in males and females on chromosome 10 was disproportion-
ately higher in the hybrid pool sets with SD ancestry (Figure 9). In both the hybrid and parental populations,
the number of male–female differentiated SNPs detected per chromosome was strongly correlated to the
126
Table 8: Linear regression results for the total number of SNPs per chromosome versus chromosome length
and number of SNPs with male-female allele frequency differences per chromosome versus chromosome
length
Total SNPs Significant SNPs
Pool R
2
p R
2
p
Hybrid 0/1 0.43 0.02 0.10 0.31
Hybrid 1 0.40 0.03 0.00 0.95
Hybrid 1/2 0.46 0.02 0.00 0.96
Hybrid Pop 0.73 < 0:001 0.00 0.90
SC 0.93 < 0:001 0.90 < 0:001
SD 0.82 < 0:001 0.50 0.01
total number of SNPs detected on that chromosome, but this relationship explained much more of the total
variance in the parental populations. Additionally, in the parental populations, both the total number of
detected SNPs per chromosome and the number of differentiated SNPs per chromosome were significantly
correlated (p = 0.016 and 0.019, respectively), and each explains a similar proportion of the total variance
(R
2
= 0.24 versus 0.23). However, while a similar, though weaker, relationship is observed for the total
number of detected SNPs per chromosome in the hybrid population (R
2
= 0.09, p = 0.035), this relation-
ship does not exist for the set of SNPs that are significantly differentiated between males and females in
hybrids (R
2
= 0.00, p = 0.929). Similar results are obtained when each pool set is analyzed individually: the
total number of detected SNPs is correlated to assembled chromosome length in all populations and to the
number of significantly differentiated SNPs in SC and SD, but no significant relationships exist between the
number of significantly differentiated SNPs and chromosome length for any hybrid pool (Table 8).
Sex-associated variants explain inter-individual genomic variation in hybrids but not in the
parentalpopulation
Most of the variance in the full set of SNPs found in individual animals is attributable to SC–SD diver-
gence (PC1 and PC2, total variance explained = 81.3%, Figure 10A), while just over 1% of the variance
separates hybrid females from other animals when all 3 populations are considered together (PC4, Figure
10A). Within the hybrid population, females with SD ancestry on chromosome 10 are separated by PC 3,
which again accounts for a very small fraction of the total variance across the genome (Figure 10B). How-
127
Figure 8: Boxplots of p-values (Cochran-Mantel-Haenszel test) for male-female allele frequency differences
on each mapped scaffold for each sequenced pool set. A, B. Pools composed of 50 siblings from families
from parental population A. SC and B. SD. C, D, E. Pools composed of 50 siblings from families where
mothers had 0-1(A), 1(C), or 1-2(E) SD alleles at the scaf-76 indel on chromosome 10. F. Random sample
of hybrid population.
ever, in both parental populations, no components separating males and females were detected (Figure 10C,
D). Thus, males and females are likely not highly genetically differentiated within natural populations of
T. californicus and the genetic basis of sex determination is likely to be highly polygenic. The relatively
high divergence between males and females observed in the hybrid population, therefore, is an unusual
phenomenon within this species.
128
Figure 9: More SNPs with statistically significant allele frequency differences between males and females
are present on chromosome 10 in the experimental hybrid population but not in the parental populations,
and chromosome length is correlated with both the number of detected SNPs and the number of SNPs with
different allele frequencies in males and females in the hybrid population but not in the parental populations.
A. Proportion of all detected, mapped SNPs per chromosome in each pool set. B. Proportion of all mapped
SNPs with different allele frequencies in males and females per chromosome per pool set (CMH test, 10%
FDR correction). Each color represents a chromosome. C. Linear regressions of total number of detected
SNPs per chromosome versus chromosome length. D. Linear regressions of the number of SNPs with allele
frequencies that differ between males and females versus length of assembled chromosomes. E., Linear
regressions of the number of male-female divergent SNPs versus the total number of SNPs per chromosome.
C,D,E. Points are colored by chromosome as in(A,B). Purple line/text: hybrid population; green line/text:
parental populations.
Discussion
Non-chromosomalsexdeterminationinnaturalT.californicuspopulations
Polygenic sex determination in Tigriopus was first proposed in 1958 (Ar-Rushdi 1958, Belser 1959),
but it has required more than 50 years of additional investigation to demonstrate definitively. Polygenic sex
determination can be inferred when heteromorphic sex chromosomes are absent and family sex ratios (i) do
129
A B
C D
Figure 10: Principal components analysis of whole–genome re-sequencing for individual animals. A. PCA
of individuals from all three sequenced populations. Triangles: Parental population SC; Crosses: Parental
population SD; Circles: Hybrid population. Black: female; red: male. B. PCA of variants for sequenced
individuals for the hybrid population. Squares: Individuals from Hybrid 0/1 families; circles: individuals
from families Hybrid 1 families; triangles: individuals from Hybrid 1/2 families. C. PCA of variants for
sequenced individuals for the SC population. D. PCA of variants for sequenced individuals for the SD
population. Colors as inA.
130
not fit a binomial distribution, (ii) are heritable, and (iii) respond to selection, all of which have now been
experimentally demonstrated in T. californicus (Ar-Rushdi 1958, V oordouw and Anholt 2002, V oordouw
et al. 2005, Foley et al. 2013, Alexander et al. 2014, Chapter 3). And, though achiasmatic meiosis is more
commonly observed in the heterogametic sex (Haldane–Huxley rule), karyotypic studies have demonstrated
that heteromorphic sex chromosomes are absent (Ar-Rushdi 1963, Lazzaretto and Libertini 1985). How-
ever, QTL mapping for sex determination has thus far only been conducted in interpopulation hybrids, and,
because larvae are not sexually dimorphic, can only be conducted using adults (Alexander et al. 2015, Foley
et al. 2013, Chapter 3). Thus, sex determination loci and sex–specific viability loci may be confounded, and
loci that have a strong effect on sex in hybrids may not necessarily have those same effects in natural pop-
ulations. This study provides the first genome–wide examination of how the genome varies between males
and females in T. californicus. Consistent with previous approaches, we found no evidence for a degen-
erated sex chromosome anywhere in the assembled genome (Figure 6). We also found that sex-associated
variation comprises a very small portion of the total inter-individual variation (Figure 10). These results are
both exciting and sobering, as they suggest that a complete dissection of the genetic basis of polygenic sex
determination may be a formidable task.
The family sex ratio assays confirmed that the parental populations do not fit a binomial distribution,
as is expected if sex determination is polygenic. However, this is also true of the hybrid population, where
chromosome 10 does appear to be behaving as a sex chromosome. This may be because the neo-sex chro-
mosome is still segregating within the hybrid population, rather than fixed. Though we assayed a large
number of hybrid families, the subset that was genotyped does not comprise a large enough dataset to test
this idea only in families with SD ancestry on chromosome 10.
Chromosome10isassociatedwithsexinT.californicushybrids
Population–diagnostic genotyping of a large panel of individuals from an experimental hybrid popu-
lation revealed that many females had retained SD ancestry across an entire chromosome in spite of the
genetic swamping of SC alleles across the rest of the genome (Figure 2; Table 3), and that SD ancestry on
chromosome 10 had consistently been maintained in females but not in males (Figure 1). Outside of chro-
mosome 10, SD ancestry, when present, was generally seen only across small map distances. Chromosome
10 is the most differentiated region of the genome between males and females within the hybrid population
but not in the parental populations (Figures 7, 8, 9; Table 8), whereas sex-associated alleles account for
131
nearly 2% of the genetic variance within the hybrid population (Figure 10). In contrast, no detected set of
loci separated males from females by PCA in the parental populations.
Several other studies have also documented a strong effect of chromosome 10 on sex in crosses involving
the SD population. In the other replicates of this hybrid population, allele frequencies were also different
between males and females, with males tending to have more SC alleles and females more SD alleles. In
non–recombinant backcrosses with SC, the RP and AB populations (Los Angeles), and the FHL population
(San Juan Islands, Washington), males consistently exhibit deficits of SD alleles (Harrison and Edmands
2006, Pritchard et al. 2011, Chapter 3). F
2
males were consistently missing SC alleles in two different SC–
SD crosses (Pritchard et al. 2011, Foley et al. 2013), while females (genotyped only in Foley et al. 2013) had
deficits of SC homozygotes. Chromosome 10 was the only chromosome in which 2 QTL associated with
sex were detected in a mapping cross between SD and a Canadian population (Alexander et al. 2015). Taken
together, these results suggest that SD alleles on chromosome 10 have a feminizing effect inT.californicus,
at least within a hybrid context.
Chromosome10actsaneo-sexchromosomewithinthehybridpopulation
Sex chromosomes are generally thought to arise when the linkage of a sex-determining locus and a
sexual conflict locus causes selection for the suppression of recombination (Charlesworth 1991). However,
if recombination suppression is already present in one sex, gaining a sex chromosome requires only the
emergence of a locus with a strong effect on sex determination (Wright et al. 2016). In that sense, T.
californicus may be “pre–adapted” to female heterogamety.The widely documented feminizing effect of SD
alleles in a hybrid context, therefore, would seem a good candidate for a sex chromosome. And that the SD
chromosome 10 has remained intact over approximately 90 generations of hybridization is not surprising
if it also determines that its possessors will be female, as meiosis is achiasmatic in Tigriopus females (Ar-
Rushdi 1963, Burton et al. 1981, Harrison and Edmands 2006, Chapter 3), so it would not be subject to
recombination.
An additional frequently observed feature of sex chromosomes is that (usually due to the accumulation
of substitutions on the Y or W after recombination has ceased), the X and Y (or Z and W) chromosomes
exhibit high genetic divergence. Thus, due to the pre–existing high genetic divergence between the SC
and SD populations (d
S
= 0.0559, Barreto et al. in prep.), chromosomes with sex–specific introgression in
interpopulation hybrids may be pre-disposed to heterogamety.
132
Finally, inbreeding could speed the spread of a novel sex chromosome within a population with poly-
genic sex determination because alleles at sex-determining loci might quickly be fixed. The effective popula-
tion size of the hybrid population in its initial months was estimated at 4.3 (Pritchard and Edmands 2013), so
the combination of hybridization and inbreeding may have acted as a perfect storm in the early generations
of the hybrid population’s existence to tip this population’s sex determination system into disarray.
One major question that remains to be answered is what the feminizing or sex-determining locus on
chromosome 10 might be. Dmrt genes play major roles in the sex determination pathways of most metazoans
(Beukeboom and Perrin 2014, Wexler et al. 2014). Two genes in the T. californicus draft genome assembly
are currently annotated as dmrt genes (Barreto et al. in prep.), and both occur on scaffolds mapped to
chromosome 8. Additional BLAST searches revealed that numerous genes on all assembled chromosomes
display homology to sex determination pathway genes, including many on chromosome 10. Future work
to narrow down the candidate gene set may require mapping studies and investigations of sex–specific
expression.
Conclusions
In this study, we have identified the emergence of a novel sex chromosome in an experimental hybrid
population. We have demonstrated that this neo-sex chromosome is not highly differentiated within the
parental populations, where sex determination is polygenic. Though the neo-W chromosome is still segre-
gating, it is already highly differentiated from the neo-Z because the two chromosomes come from highly
divergent parental populations. Identification of the causal loci will require mapping experiments. To our
knowledge, this is the first experimental evolution of a novel sex chromosome when sex determination was
not already chromosomal.
Acknowledgements
Thanks to Wai Leong for help with the initial genotyping and to Laura Moore, who assisted with hybrid
population maintenance during its initial 21 months.
133
References
Alexander, H., J. Richardson, and B. Anholt. 2014. Multigenerational response to artificial selection for
biased clutch sex ratios inTigriopuscalifornicus populations. Journal of Evolutionary Biology27:1921–
1929.
Alexander, H., J. Richardson, S. Edmands, and B. Anholt. 2015. Sex without sex chromosomes: genetic
architecture of multiple loci independently segregating to determine sex ratios in the copepod Tigriopus
californicus. Journal of Evolutionary Biology28:2196–2207.
Ar-Rushdi, A. H. 1958. The polygenic basis of sex-ratio in Tigriopus. Paper presented at the Xth Interna-
tional Congress of Genetics, McGill University, Montreal, Canada.
Ar-Rushdi, A. H. 1963. The cytology of achiasmatic meiosis in the female Tigriopus (Copepoda). Chro-
mosoma13:1754–1763.
Barnett, D. W., E. K. Garrison, A. R. Quinlan, M. P. Stromberg, and G. T. Marth. 2011. BamTools: a C++
API and toolkit for analyzing and managing BAM files. Bioinformatics27:1691–1692.
Belser, W. L. 1959. The research frontier: Where is science taking us? The Saturday Review48:58.
Beukeboom, L. W., and N. Perrin. 2014. The Evolution of Sex Determination. Oxford University Press.
Bull, J. 1983. Evolution of Sex Determining Mechanisms. Benjamin/Cummings, Menlo Park, CA.
Bull, J., and E. Charnov. 1977. Changes in the heterogametic mechanism of sex determination. Heredity
39:1–14.
Bulmer, M., and J. Bull. 1982. Models of polygenic sex determination and sex ratio control. Evolution
36:13–26.
Burton, R., M. Feldman, and S. Swisher. 1981. Linkage relationships among 5 enzyme-coding gene loci in
the copepodTigriopuscalifornicus: A genetic confirmation of achiasmatic meiosis. Biochemical Genetics
19:1237–1245.
Charlesworth, B. 1991. The evolution of sex chromosomes. Science251:1030–1033.
DePristo, M. A., E. Banks, R. Poplin, K. Garimella, J. Maguire, C. Hartl, A. Philippakis, G. del Angel,
M. A. Rivas, M. Hanna, A. McKenna, T. Fennell, A. Kernytsky, A. Sivachenko, K. Cibulskis, S. Gabriel,
D. Altshuler, and M. Daly. 2011. A framework for variation discovery and genotyping using next-
generation DNA sequencing data. Nature Genetics43:491–498.
Fisher, R. 1930. TheGeneticalTheoryofNaturalSelection. Clarendon Press, Oxford, UK.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, and S. Edmands. 2013. Postzygotic isolation involves
strong mitochondrial and sex–specific effects in Tigriopus californicus, a species lacking heteromorphic
sex chromosomes. Heredity111:391–401.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, G. W. Moy, R. S. Burton, and S. Edmands. 2011.
A gene-based SNP resource and linkage map for the copepod Tigriopus californicus. BMC Genomics
12:568.
Franco, M., P. Rubini, and M. Vecchi. 1982. Sex–determinants and their distribution in various populations
ofMuscadomestica L. of Western Europe. Genetical Research40:279–293.
134
Ghiselli, F., L. Milani, P. L. Chang, D. Hedgecock, J. P. Davis, S. V . Nuzhdin, and M. Passamonti. 2012.
De novo assembly of the Manila clamRuditapesphilippinarum transcriptome provides new insights into
expression bias, mitochondrial doubly uniparental inheritance and sex determination. Molecular Biology
and Evolution29:771–786.
Green, D. 1988. Cytogenetics of the endemic New Zealand frog, Leiopelma hochstetteri: extraordinary
supernumerary chromosome variation and a unique sex-chromosome system. Chromosoma97:55–70.
Green, D., C. Zeyl, and T. Sharbel. 1993. The evolution of hypervariable sex and supernumerary (B)
chromosomes in the relict New Zealand frog, Leiopelma hochstetteri. Journal of Evolutionary Biology
6:417–441.
Harrison, J. S., and S. Edmands. 2006. Chromosomal basis of viability differences inTigriopuscalifornicus
interpopulation hybrids. Journal of Evolutionary Biology19:2040–2051.
Kitano, J., J. A. Ross, S. Mori, M. Kume, F. C. Jones, Y . F. Chan, D. M. Absher, J. Grimwood, J. Schmutz,
R. M. Myers, D. M. Kingsley, and C. L. Peichel. 2009. A role for a neo-sex chromosome in stickleback
speciation. Nature461:1079–1083.
Klaus, B., and K. Strimmer, 2015. fdrtool: Estimation of (Local) False Discovery Rates and Higher Criti-
cism, version 1.2.15 edition.
Kofler, R., R. V . Pandey, and C. Schl¨ otterer. 2011. PoPoolation2: identifying differentiation between
populations using sequencing of pooled DNA samples (Pool-Seq). Bioinformatics27:3435–6.
Lazzaretto, I., and A. Libertini. 1985. Karyological comparisons among three species of the genusTigriopus
(Copepods, Harpacticoida). Atti. Acc. Lincei Rend. fis.126:377–381.
Li, H. 2011. A statistical framework for SNP calling, mutation discovery, association mapping and popula-
tion genetical parameter estimation from sequencing data. Bioinformatics27:2987–2993.
Liew, W., R. Bartfai, Z. Lim, R. Sreenivasan, K. Siegfried, and L. Orban. 2012. Polygenic sex determination
system in zebrafish. PLoS ONE7:e34397.
McKenna, A., M. Hanna, E. Banks, A. Sivachenko, K. Cibulskis, A. Kernytsky, K. Garimella, D. Altshuler,
S. Gabriel, M. Daly, and M. A. DePristo. 2010. The Genome Analysis Toolkit: a MapReduce framework
for analyzing next-generation DNA sequencing data. Genome Research20:1297–1303.
Parnell, N., and J. Streelman. 2013. Genetic interactions controlling sex and color establish the potential for
sexual conflict in Lake Malawi cichlid fishes. Heredity110:239–246.
Peterson, D., K. Kubow, M. Connolly, L. Kaplan, M. Wetkowski, W. Leong, B. Phillips, and S. Edmands.
2013. Reproductive and phylogenetic divergence of tidepool copepod populations across a narrow geo-
graphical boundary in Baja California. Journal of Biogeography40:1664–1675.
Pritchard, V ., L. Dimond, J. Harrison, C. Velasquez, J. Zieba, R. Burton, and S. Edmands. 2011. Interpopu-
lation hybridization results in widespread viability selection across the genome inTigriopuscalifornicus.
BMC Genetics12:54.
Pritchard, V ., and S. Edmands. 2013. The genomic trajectory of hybrid swarms: Outcomes of repeated
crosses between populations ofTigriopuscalifornicus. Evolution67:774–791.
135
Pritchard, V ., V . Knutson, M. Lee, J. Zeiba, and S. Edmands. 2013. Fitness and morphological outcomes of
many generations of hybridization in the copepodTigriopuscalifornicus. Journal of Evolutionary Biology
26:416–433.
Purcell, S., B. Neale, K. Todd-Brown, L. Thomas, M. A. Ferreira, D. Bender, J. Maller, P. Sklar, P. I.
De Bakker, M. J. Daly, et al. 2007. PLINK: a tool set for whole-genome association and population-
based linkage analyses. The American Journal of Human Genetics81:559–575.
Ramachandra, N., and H. Ranganath. 1986. The chromosomes of two Drosophila races: D. nasuta nasuta
andD.n.albomicana: IV . Hybridization and karyotype repatterning. Chromosoma93:243–248.
Rice, W. 1986. On the instability of polygenic sex determination: the effect of sex-specific selection.
Evolution40:633–639.
Rice, W. 1992. Sexually antagonistic genes: experimental evidence. Science256:1436–1439.
Rice, W. 1994. Degeneration of a nonrecombining chromosome. Science263:230–232.
Roberts, R. B., J. R. Ser, and T. D. Kocher. 2009. Sexual Conflict Resolved by Invasion of a Novel Sex
Determiner in Lake Malawi Cichlid Fishes. Science326:998–1001.
Rodrigues, N., C. Betto-Colliard, H. Jourdan-Pineau, and N. Perrin. 2013. Within-population polymorphism
of sex-determination systems in the common frog (Rana temporaria). Journal of Evolutionary Biology
26:1569–1577.
Rodrigues, N., Y . Vuille, and N. Perrin. 2015. Sex-chromosome differentiation and ‘sex races’ in the com-
mon frog (Ranatemporaria). Proceedings of the Royal Society B: Biological Sciences282:20142726.
Schuelke, M. 2000. An economic method for the fluorescent labeling of PCR fragments. Nature Biotech-
nology18:233–234.
Ser, J., R. Roberts, and T. Kocher. 2010. Multiple interacting loci control sex determination in Lake Malawi
cichlid fish. Evolution64:486–501.
Sharbel, T., and D. Green. 1998. B-chromosome origin in the endemic New Zealand frog Leiopelma
hochstetteri through sex chromosome devolution. Genome41:14–22.
Tanuja, M., N. Ramachandra, and H. Ranganath. 1999. Evolution of a recent neo-Y sex chromosome in a
laboratory population ofDrosophila. Journal of Genetics78:81–85.
Team, R. C., 2015. R: a language and environment for statistical computing.
The Tree of Sex Consortium. 2014. Sex determination: Why so many ways of doing it? PLoS Biology
12:e1001899.
Van der Auwera, G. A., M. O. Carneiro, C. Hartl, R. Poplin, G. del Angel, A. Levy-Moonshine, T. Jordan,
K. Shakir, D. Roazen, J. Thibault, E. Banks, K. V . Garimella, D. Altshuler, S. Gabriel, and M. A. DePristo.
2013. From FastQ data to high-confidence variant calls: The Genome Analysis Toolkit best practices
pipeline. Current Protocols in Bioinformatics43:11.10.1–11.10.33.
van Doorn, G., and M. Kirkpatrick. 2007. Turnover of sex chromosomes induced by sexual conflict. Nature
449:909–912.
136
van Doorn, G., and M. Kirkpatrick. 2010. Transitions between male and female heterogamety caused by
sex–antagonistic selection. Genetics186:629–645.
Vandeputte, M., M. Dupont-Nivet, H. Chavanne, and B. Chatain. 2007. A polygenic hypothesis for sex
determination in the European sea bassDicentrarchuslabrax. Genetics176:1049–1057.
Vicoso, B., and D. Bachtrog. 2015. Numerous transitions of sex chromosomes in Diptera. PLoS Biology
13:e1002078.
Vicoso, B., J. J. Emerson, Y . Zektser, S. Mahajan, and D. Bachtrog. 2013. Comparative sex chromosome
genomics in snakes: Differentiation, evolutionary strata, and lack of global dosage compensation. PLoS
Biology11:e1001643.
V oordouw, M., and B. Anholt. 2002. Heritability of sex tendency in a harpacticoid copepod, Tigriopus
californicus. Evolution56:1754–1763.
V oordouw, M., H. Robinson, and B. Anholt. 2005. Paternal inheritance of the primary sex ratio in a copepod.
Journal of Evolutionary Biology18:1304–1314.
V oordouw, M. J., G. Stebbins, H. E. Robinson, M.-J. Perrot-Minnot, T. Rigaud, and B. R. Anholt. 2008.
Genetic variation in the primary sex ratio in populations of the intertidal copepod,Tigriopuscalifornicus,
is widespread on Vancouver Island. Evolutionary Ecology Research10:1007–1023.
Wexler, J. R., D. C. Plachetzki, and A. Kopp. 2014. Pan-metazoan phylogeny of the DMRT gene family: a
framework for functional studies. Development Genes and Evolution224:175–181.
Wright, A. E., R. Dean, F. Zimmer, and J. E. Mank. 2016. How to make a sex chromosome. Nature
Communications7:12087.
Yu, Y ., F. Lin, and H. Chang. 1999. Stepwise chromosome evolution in Drosophila albomicans. Heredity
83:39–45.
Yusa, Y . 2007. Causes of variation in sex ratio and modes of sex determination in the Mollusca-an overview.
American Malacological Bulletin23:89–98.
Yusa, Y ., S. Breton, and W. Hoeh. 2013. Population genetics of sex determination in Mytilus mussels:
Reanalyses and a model. Journal of Heredity104:380–385.
Zheng, X., D. Levine, J. Shen, S. Gogarten, C. Laurie, and B. Weir. 2012. A high-performance computing
toolset for relatedness and principal component analysis of SNP data. Bioinformatics28:3326–3328.
137
Conclusions
The work in this dissertation illustrates how the speciation process in taxa that do not have sex chromo-
somes may differ from the canonical rules and highlights the necessity of examining taxa beyond conven-
tional model organisms in the pursuit of explanations regarding how biodiversity arises and is maintained.
It examines the speciation process from the general to the specific, and finds commonalities across taxa and
approaches. Key findings include that speciation may occur at different rates in the presence and absence
of sex chromosomes, that hybridization may have sex–specific effects even without sex chromosomes, and
that hybridization may be able to catalyze the emergence of a sex chromosome in taxa that lack them.
Slower speciation in the absence of sex chromosomes
In diverging lineages, sex chromosomes often have an outsized role in postzygotic isolation and may
accumulate incompatibility loci more quickly than other parts of the genome. In their absence, therefore,
speciation might occur more slowly. We find support for this hypothesis in reptiles in Chapter 1, where
net diversification intervals are longer in groups that mostly lack heteromorphic sex chromosomes (turtles
and crocodilians). However, we also found that birds, which mostly have highly heteromorphic ZW chro-
mosomes, also have long net diversification intervals. Recent work has illuminated the extent to which W
chromosome degeneration differs within birds (and also snakes), which will allow for a more nuanced analy-
sis in future (Vicoso et al. 2013, Demuth 2014, Zhou et al. 2014). A different approach to the same question
drew on hybridization records from numerous taxa and found that at similar levels of genetic divergence,
taxa lacking sex chromosomes exhibited less reproductive isolation than did taxa with homomorphic (inter-
mediate) or heteromorphic (highest) sex chromosomes, suggesting that slower speciation in the absence of
degenerate sex chromosomes may be a common pattern (Lima 2014).
In Chapter 2, we documented the extreme divergence present within T. californicus. We found vari-
ation within T. californicus at 153 sites in our mt12S alignment (out of 260 total variable sites across all
4 Tigriopus species assessed). Chapters 3 and 4, however, demonstrate the resilience of T. californicus to
hybridization, even across large genetic distances. Though all but the closest crosses exhibited some degree
of postzygotic isolation, reproductive isolation was incomplete, even over large genetic distances (Chapter
3), and hybrid populations of T. californicus were able to persist more or less indefinitely (Chapter 4). By
contrast, an analysis of 7 species within the Drosophila nasuta subgroup found only 3 variable sites in the
138
mt12S gene (Nagaraja et al. 2004). Members of the D. nasuta subgroup experience such extreme genomic
dysfunction upon hybridization that hybrid lines have gained a novel sex chromosome (Ramachandra and
Ranganath 1986, Tanuja et al. 1999).
In Chapter 3, we observed the strongest reproductive isolation documented between populations of
T. californicus (outside of the Baja populations demonstrated to be a different species in Chapter 2) in
one direction of our high–divergence cross—near complete mortality of larvae at or shortly after hatching.
However, we found far less extreme reproductive isolation in the other direction of the cross—indeed, it
was the cross with the greatest increase in survivorship relative to the additive expectation (though not
statistically significant), and we found fewer chromosomes with segregation distortion and fewer epistatic
incompatibilities in this cross than in crosses between less genetically divergent populations. Attempts
to correlate postzygotic incompatibility with genetic divergence have previously met with mixed results,
suggesting that postzygotic isolation may often emerge non–linearly with time (Edmands 2002).
The species status of the Tigriopus populations at the southern edge of the species range has been un-
certain for over two decades now (Ganz and Burton 1995). Chapter 2 clarifies this issue, finding that these
geographically contiguous but reproductively isolated populations are a novel species. The geological his-
tory of the Baja peninsula is complex and still contentious, but phylogeographic breaks can be found for
many flora and fauna in the region, including other intertidal species (Riddle et al. 2000, Bernardi et al.
2003, Leach´ e et al. 2007, Hurtado et al. 2010).
Sex–specific hybrid fitness effects even in the absence of sex chromosomes
Haldane’s rule—the observation that deleterious effects of hybridization, when sex–specific, tend to
impact the heterogametic sex more severely—has been confirmed in numerous taxa, under both male and
female heterogamety, in animals and in plants, with homomorphic and heteromorphic sex chromosomes, and
even in haplodiploids (Presgraves and Orr 1998, Naisbit et al. 2002, Presgraves 2002, Coyne and Orr 2004,
Koevoets and Beukeboom 2008, Brothers and Delph 2010, Schilthuizen et al. 2011, Dufresnes et al. 2016).
In each of these cases, however, the heterogametic sex possesses a chromosome absent in the homogametic
sex—a Y or W in heteromorphic and homomorphic XY or ZW taxa, and an entire set of chromosomes in
the case of haplodiploids.
In Chapter 3, we demonstrate that sex–specific deleterious effects of hybridity can occur even in the
absence of a differential chromosomal complement between males and females. In backcrosses spanning a
139
wide range of geographic and genetic distances withinT.californicus, family sex ratios are consistently more
female–biased than the additive expectation, and males exhibit more segregation distortion (a manifestation
of genetic inviability) than females. Additionally, although average heterozygosity did not differ genome–
wide between males and females in both the crosses where animals of both sexes were genotyped, they
did differ in heterozygosity at from 1/4 to 1/2 of their chromosomes, with males nearly always being more
heterozygous. The only chromosome where females were more heterozygous than males was chromosome
10, on which SD alleles are frequently associated with femaleness (homozygous animals in both crosses
where males and females were genotyped lack SD alleles at that locus; Chapter 3, Chapter 4, Pritchard and
Edmands 2012, Foley et al. 2013). These results contrast with a previous non–recombinant backcross study,
where sex ratios were male–biased in 2 of 3 crosses, and females were more heterozygous than males at
most loci with differences in heterozygosity between sexes (Harrison and Edmands 2006).
The faster–male evolution hypothesis, frequently invoked as a partial explanation for Haldane’s rule, can
also apply to taxa where males are not heterogametic (Wu and Davis 1993, Wu et al. 1996, Willett 2008). In
most taxa, hybrid sterility evolves before hybrid inviability (Coyne and Orr 2004), but hybrid sterility (for
either sex) is rarely observed in T. californicus (Willett 2008, Foley et al. 2013). Foley et al. (2013) found
that male–specific estimates of survival rates were consistently lower than female–specific rates, regardless
of mitochondrial background, in an F
2
cross, consistent with the missing adult males in the family sex ratio
assays in Chapter 3. However, faster–male evolution is more often invoked to explain sterility than viability
because it is often assumed that loci affecting male viability will also affect female viability (Wu et al. 1996).
An alternative possible explanation for the sex–specific viability observed in Chapter 3 is the mother’s
curse hypothesis, which states that alleles with male–deleterious effects may accumulate in the mitochondria
because the uniparental inheritance of mitochondria shields them from selection in males (Gemmell et al.
2004). Mitochondria play vital roles in the functioning of eukaryotic organisms, but their genomes are
reduced to only a few genes, and they rely on the importation of approximately 1000 nuclear–encoded
genes to function (Rand et al. 2004). Much previous work in has focused on cytonuclear dysfunction in
T. californicus hybrids, including in some of the crosses examined in Chapter 3 (Ellison and Burton 2006,
2008, 2010), and all of the crosses utilized in Chapter 3 were backcrossed to the paternal population, such
that only heterozygotes would have maternal population alleles that matched the mitochondrial population–
of–origin. If males are more sensitive to mito–nuclear dysfunction than females, this could potentially
explain both why they were more heterozygous than females at many loci and why they apparently had
140
higher mortality. In Harrison and Edmands (2006), female F
1
s were always backcrossed to the maternal,
rather than the paternal population, so the higher heterozygosity observed in females would again fit a pattern
of stronger sensitivity of males to mito–nuclear dysfunction, as only heterozygotes would potentially carry
the mismatched alleles in that case.
Emergence of a sex chromosome as a consequence of hybridization
Polygenic sex determination has received increasing attention in recent years, as observations in more
taxa demonstrate that it is more common (and potentially more stable) than previously believed. In Chapter
3, QTL mapping of sex-determining loci was consistent with previous results that identified multiple chro-
mosomes, including a large effect of chromosome 10, on sex in interpopulation hybrids (Alexander et al.
2015). However, because sex–specific viability and sex determination are entangled in studies conducted
with hybrids, it is important to confirm these results in natural populations. Chapter 4 demonstrates that
males and females from two natural populations ofT.californicus are not differentiated in any region of the
genome in a manner consistent with a sex chromosome and provides the first genome–wide, intrapopulation
examination of male–female divergence inT.californicus, finding it to be extremely low.
Chapter 4 also tracks the differentiation between males and females in an experimental hybrid popu-
lation, where SD alleles have been maintained along the full length of chromosome 10 in females but not
in males for 90 or more generations. The association of SD alleles with femaleness has been observed in
multiple interpopulation crosses, and two independent sex QTL were mapped to chromosome in a cross be-
tween SD and a Canadian population (Chapter 3, Harrison and Edmands 2006, Foley et al. 2013, Alexander
et al. 2015).
When recombination is absent in one sex, the evolution of a sex chromosome requires only a single
step: the emergence of a strong sex–determiner within a population (Wright et al. 2016). The absence of
meiotic recombination in T. californicus is well–documented (Chapter 3, Ar-Rushdi 1963, Burton et al.
1981, Harrison and Edmands 2006), and the allopatric nature of Tigriopus populations, coupled with the
high genetic divergence often observed between populations (Chapter 2, Edmands 2001, Peterson et al.
2013), could lead to highly divergent sets of sex determination alleles in T. californicus populations. The
introduction of novel allelic combinations in hybrids, therefore, could provide an opportunity for takeover
by a strong sex determination allele, leading to the rapid evolution of a novel sex chromosome.
In the hybrid population examined in Chapter 4, SD alleles on chromosome 10 are found almost exclu-
141
sively in females. Females are always heterozygous for this chromosome because they cannot inherit an SD
chromosome 10 from their fathers, and females who have SD alleles on chromosome 10 are highly differen-
tiated from males because the two parental populations are highly differentiated, and rather than decaying,
the linkage around the sex–determining region on chromosome 10 is being maintained because meiosis in
females is achiasmatic. Therefore, the total evidence suggests that chromosome 10 is functioning as a novel
sex chromosome in the hybrid population but not in the parental populations.
Although the emergence of a sex chromosome from polygenic sex determination has been predicted
for more than 30 years (Rice 1986), this is to our knowledge the first time it has occurred in a laboratory.
However, as chromosome 10 appears to exert a strong effect on sex, at least in crosses involving SD, this
phenomenon may well be replicable.
Final remarks
Emerging evidence suggests that the speciation process may occur fundamentally differently in taxa that
lack sex chromosomes. Validating this hypothesis will require examining genetic differentiation and repro-
ductive isolation in numerous taxa outside traditional model species. Work on the genetics of postzygotic
isolation has traditionally concentrated onDrosophila andMus in large part because of the wide availability
of genetic, genomic, and experimental resources for these organisms. However, in the genomics era, this
focus is beginning to expand, allowing for a more comparative framework wherein it will be possible to
draw inferences from numerous taxa, each of whom has both unique and shared characteristics with other
taxa. Currently, nearly 50% of published animal genomes currently come from vertebrates, while 26% are
from insects (Dunn and Ryan 2015). Expanding the resources for taxa outside these groups will be vital to
an understanding of the speciation process, which underlies biodiversity. Results to date suggest that taxa
outside the boundaries of traditional work can offer surprising and important perspectives on the speciation
process.
142
References
Alexander, H., J. Richardson, S. Edmands, and B. Anholt. 2015. Sex without sex chromosomes: genetic
architecture of multiple loci independently segregating to determine sex ratios in the copepod Tigriopus
californicus. Journal of Evolutionary Biology 28:2196–2207.
Ar-Rushdi, A. H. 1963. The cytology of achiasmatic meiosis in the female Tigriopus (Copepoda). Chro-
mosoma 13:1754–1763.
Bernardi, G., L. Findley, and A. Rocha-Olivares. 2003. Vicariance and dispersal across Baja California in
disjunct marine fish populations. Evolution 57:1599–1609.
Brothers, A. N., and L. F. Delph. 2010. Haldane’s rule is extended to plants with sex chromosomes.
Evolution 64:3643–3648.
Burton, R., M. Feldman, and S. Swisher. 1981. Linkage relationships among 5 enzyme-coding gene loci in
the copepodTigriopuscalifornicus: A genetic confirmation of achiasmatic meiosis. Biochemical Genetics
19:1237–1245.
Coyne, J. A., and H. A. Orr. 2004. Speciation. Sinauer, Sunderland, MA.
Demuth, J. P. 2014. Do sex chromosomes affect speciation rate? BioEssays 36:632.
Dufresnes, C., T. Majtyka, S. J. E. Baird, J. F. Gerchen, A. Borzee, R. Savary, M. Ogielska, N. Perrin, and
M. Stock. 2016. Empirical evidence for large X-effects in animals with undifferentiated sex chromo-
somes. Scientific Reports 6:21029.
Dunn, C. W., and J. F. Ryan. 2015. The evolution of animal genomes. Current Opinions in Genetics and
Development 35:25–32.
Edmands, S. 2001. Phylogeography of the intertidal copepod Tigriopus californicus reveals substantially
reduced population differentiation at northern latitudes. Molecular Ecology 10:1743–1750.
Edmands, S. 2002. Does parental divergence predict reproductive compatibility? Trends in Ecology and
Evolution 17:520–527.
Ellison, C. K., and R. S. Burton. 2006. Disruption of mitochondrial function in interpopulation hybrids of
Tigriopuscalifornicus. Evolution 60:1382–1391.
Ellison, C. K., and R. S. Burton. 2008. Interpopulation hybrid breakdown maps to the mitochondrial
genome. Evolution 62:631–638.
Ellison, C. K., and R. S. Burton. 2010. Cytonuclear conflict in interpopulation hybrids: The role of RNA
polymerase in mtDNA transcription and replication. Journal of Evolutionary Biology 23:528–538.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, and S. Edmands. 2013. Postzygotic isolation involves
strong mitochondrial and sex–specific effects in Tigriopus californicus, a species lacking heteromorphic
sex chromosomes. Heredity 111:391–401.
Ganz, H., and R. Burton. 1995. Genetic differentiation and reproductive incompatibility among Baja
California populations of the copepodTigriopuscalifornicus. Marine Biology 123:821–827.
Gemmell, N. J., V . J. Metcalf, and F. W. Allendorf. 2004. Mother’s curse: the effect of mtDNA on individual
fitness and population viability. Trends in Ecology and Evolution 19:238–244.
143
Harrison, J. S., and S. Edmands. 2006. Chromosomal basis of viability differences inTigriopuscalifornicus
interpopulation hybrids. Journal of Evolutionary Biology 19:2040–2051.
Hurtado, L. A., M. Mateos, and C. A. Santamaria. 2010. Phylogeography of supralittoral rocky intertidal
Ligia isopods in the Pacific region from Central California to Central Mexico. PLoS One 5:e11633.
Koevoets, T., and L. W. Beukeboom. 2008. Genetics of postzygotic isolation and Haldane’s rule in hap-
lodiploids. Heredity 102:16–23.
Leach´ e, A. D., S. C. Crews, and M. J. Hickerson. 2007. Two waves of diversification in mammals and
reptiles of Baja California revealed by hierarchical Bayesian analysis. Biology Letters 3:646–650.
Lima, T. G. 2014. Higher levels of sex chromosome heteromorphism are associated with markedly stronger
reproductive isolation. Nature Communications 5:4743.
Nagaraja, J. Nagaraju, and H. A. Ranganath. 2004. Molecular phylogeny of the nasuta subgroup of
Drosophila based on 12S rRNA, 16S rRNA, and CoI mitochondrial genes, RAPD and ISSR polymor-
phisms. Genes & Genetic Systems 79:293–299.
Naisbit, R. E., C. D. Jiggins, M. Linares, C. Salazar, and J. Mallet. 2002. Hybrid sterility, Haldane’s rule
and speciation inHeliconiuscydno andH.melpomene. Genetics 161:1517–1526.
Peterson, D., K. Kubow, M. Connolly, L. Kaplan, M. Wetkowski, W. Leong, B. Phillips, and S. Edmands.
2013. Reproductive and phylogenetic divergence of tidepool copepod populations across a narrow geo-
graphical boundary in Baja California. Journal of Biogeography 40:1664–1675.
Presgraves, D. C. 2002. Patterns of postzygotic isolation in Lepidoptera. Evolution 56:1168–83.
Presgraves, D. C., and H. A. Orr. 1998. Haldane’s rule in taxa lacking a hemizygous X. Science 282:952–
954.
Pritchard, V ., and S. Edmands. 2012. The genomic trajectory of hybrid swarms: Outcomes of repeated
crosses between populations ofTigriopuscalifornicus. Evolution 67:774–791.
Ramachandra, N., and H. Ranganath. 1986. The chromosomes of two Drosophila races: D. nasuta nasuta
andD.n.albomicana: IV . Hybridization and karyotype repatterning. Chromosoma 93:243–248.
Rand, D., R. Haney, and A. Fry. 2004. Cytonuclear coevolution: the genomics of cooperation. Trends in
Ecology and Evolution 19:645–653.
Rice, W. 1986. On the instability of polygenic sex determination: The effect of sex-specific selection.
Evolution 40:633–639.
Riddle, B. R., D. J. Hafner, L. F. Alexander, and J. R. Jaeger. 2000. Cryptic vicariance in the historical
assembly of a Baja California Peninsular Desert biota. Proceedings of the National Academy of Sciences
97:14438–14443.
Schilthuizen, M., M. C. W. G. Giesbers, and L. W. Beukeboom. 2011. Haldane’s rule in the 21st century.
Heredity 107:95–102.
Tanuja, M., N. Ramachandra, and H. Ranganath. 1999. Evolution of a recent neo-Y sex chromosome in a
laboratory population ofDrosophila. Journal of Genetics 78:81–85.
144
Vicoso, B., J. J. Emerson, Y . Zektser, S. Mahajan, and D. Bachtrog. 2013. Comparative sex chromosome
genomics in snakes: Differentiation, evolutionary strata, and lack of global dosage compensation. PLoS
Biology 11:e1001643.
Willett, C. S. 2008. No evidence for faster male hybrid sterility in population crosses of an intertidal copepod
(Tigriopuscalifornicus). Genetica 133:129–136.
Wright, A. E., R. Dean, F. Zimmer, and J. E. Mank. 2016. How to make a sex chromosome. Nature
Communications 7:12087.
Wu, C., and A. Davis. 1993. Evolution of postmating reproductive isolation: the composite nature of
Haldane’s rule and its genetic bases. American Naturalist 142:187–212.
Wu, C.-I., N. A. Johnson, and M. F. Palopoli. 1996. Haldane’s rule and its legacy: Why are there so many
sterile males? Trends in Ecology and Evolution 11:281–284.
Zhou, Q., J. Zhang, D. Bachtrog, N. An, Q. Huang, E. D. Jarvis, M. T. P. Gilbert, and G. Zhang. 2014.
Complex evolutionary trajectories of sex chromosomes across bird taxa. Science 346:1332.
145
Bibliography
Alexander, H., J. Richardson, and B. Anholt. 2014. Multigenerational response to artificial selection for
biased clutch sex ratios inTigriopuscalifornicus populations. Journal of Evolutionary Biology 27:1921–
1929.
Alexander, H., J. Richardson, S. Edmands, and B. Anholt. 2015. Sex without sex chromosomes: genetic
architecture of multiple loci independently segregating to determine sex ratios in the copepod Tigriopus
californicus. Journal of Evolutionary Biology 28:2196–2207.
Allendorf, F., R. Leary, P. Spruell, and J. Wenburg. 2001. The problems with hybrids: setting conservation
guidelines. Trends in Ecology and Evolution 16:613–622.
Anderson, J. L., A. R. Mari, I. Braasch, A. A. P. Hohenlohe, P. Batzel, and J. H. Postlethwait. 2012.
Multiple sex–associated regions and a putative sex chromosome in zebrafish revealed by RAD mapping
and population genomics. PLoS One 7:e40701.
Ar-Rushdi, A. H. 1958. The polygenic basis of sex-ratio in Tigriopus. Paper presented at the Xth Interna-
tional Congress of Genetics, McGill University, Montreal, Canada.
Ar-Rushdi, A. H. 1963. The cytology of achiasmatic meiosis in the female Tigriopus (Copepoda). Chro-
mosoma 13:1754–1763.
Baker, C. F., 1912. First Annual Report of the Laguna Marine Laboratory at Laguna Beach, Orange County,
California, Chapter: Notes on crustacea of laguna beach . Claremont College.
Barnett, D. W., E. K. Garrison, A. R. Quinlan, M. P. Stromberg, and G. T. Marth. 2011. BamTools: a C++
API and toolkit for analyzing and managing BAM files. Bioinformatics 27:1691–1692.
Barreto, F. S., G. W. Moy, and R. S. Burton. 2011. Interpopulation patterns of divergence and selection
across the transcriptome of the copepodTigriopuscalifornicus. Molecular Ecology 20:560–72.
Belser, W. L. 1959. The research frontier: Where is science taking us? The Saturday Review 48:58.
Benjamini, Y ., and Y . Hochberg. 1995. Controlling the false discovery rate: A practical and powerful
approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological) 57:289–
300.
Bernardi, G., L. Findley, and A. Rocha-Olivares. 2003. Vicariance and dispersal across Baja California in
disjunct marine fish populations. Evolution 57:1599–1609.
Beukeboom, L. W., and N. Perrin. 2014. The Evolution of Sex Determination. Oxford University Press.
Broman, K. W., H. Wu, S. Sen, and G. A. Churchill. 2003. R/qtl: QTL mapping in experimental crosses.
Bioinformatics 19:889–890.
146
Brothers, A. N., and L. F. Delph. 2010. Haldane’s rule is extended to plants with sex chromosomes.
Evolution 64:3643–3648.
Brown, A. F. 1991. Outbreeding depression as a cost of dispersal in the harpacticoid copepod, Tigriopus
californicus. Biological Bulletin 181:123–126.
Bull, J. 1983. Evolution of Sex Determining Mechanisms. Benjamin/Cummings, Menlo Park, CA.
Bull, J., and E. Charnov. 1977. Changes in the heterogametic mechanism of sex determination. Heredity
39:1–14.
Bulmer, M., and J. Bull. 1982. Models of polygenic sex determination and sex ratio control. Evolution
36:13–26.
Burton, R. 1987. Differentiation and integration of the genome in populations of the marine copepod
Tigriopuscalifornicus. Evolution 41:504–513.
Burton, R. 1990a. Hybrid breakdown in developmental time in the copepodTigriopuscalifornicus. Evolu-
tion 44:1814–1822.
Burton, R. 1990b. Hybrid breakdown in physiological response: A mechanistic approach. Evolution
44:1806–1813.
Burton, R. 1997. Genetic evidence for long term persistence of marine invertebrate populations in an
ephemeral environment. Evolution 51:993–998.
Burton, R., and F. Barreto. 2012. A disproportionate role for mtDNA in Dobzhansky-Muller incompatibili-
ties? Molecular Ecology 20:4942–4957.
Burton, R., M. Feldman, and J. Curtsinger. 1979. Population genetics ofTigriopuscalifornicus (Copepoda,
Harpacticoida) 1. Population structure along the central California coast. Mar Ecol Prog Ser 1:29–39.
Burton, R., M. Feldman, and S. Swisher. 1981. Linkage relationships among 5 enzyme-coding gene loci in
the copepodTigriopuscalifornicus: A genetic confirmation of achiasmatic meiosis. Biochemical Genetics
19:1237–1245.
Burton, R., E. Metz, J. Flowers, and C. Willet. 2005. Unusual structure of ribosomal DNA in the copepod
Tigriopuscalifornicus: intergenic spacer sequences lack internal subrepeats. Gene 344:105–113.
Burton, R., P. Rawson, and S. Edmands. 1999. Genetic architecture of physiological phenotypes: Empirical
evidence for coadapted gene complexes. American Zoologist 39:451–462.
Burton, R. S. 1985. Mating system of the intertidal copepod Tigriopus californicus. Marine Biology
86:247–252.
Burton, R. S., R. J. Byrne, and P. D. Rawson. 2007. Three divergent mitochondrial genomes from California
populations of the copepodTigriopuscalifornicus. Gene 403:53–59.
Burton, R. S., C. K. Ellison, and J. S. Harrison. 2006. The sorry state of F2 hybrids: Consequences of rapid
mitochondrial DNA evolution in allopatric populations. American Naturalist 168:S14–S24.
Burton, R. S., and M. W. Feldman. 1981. Population genetics of Tigriopus californicus: II. Differentiation
among neighboring populations. Evolution 35:1192–1205.
147
Burton, R. S., and S. G. Swisher. 1984. Population structure of the intertidal copepodTigriopuscalifornicus
as revealed by field manipulation of allele frequencies. Oecologia 65:108–111.
Camus, M. F., D. J. Clancy, and D. K. Dowling. 2012. Mitochondria, maternal inheritance, and male aging.
Current Biology 22:1717–1721.
Ceballos, G., P. R. Ehrlich, A. D. Barnosky, A. Garcia, R. M. Pringle, and T. M. Palmer. 2015. Accel-
erated modern human–induced species losses: Entering the sixth mass extinction. Science Advances
1:e1400253.
Chalker-Scott, L. 1995. Survival and sex-ratios of the intertidal copepod, Tigriopus californicus, following
ultraviolet-B (290-320 nm) radiation exposure. Marine Biology 123:799–804.
Charlesworth, B. 1991. The evolution of sex chromosomes. Science 251:1030–1033.
Charlesworth, D., and J. E. Mank. 2010. The birds and the bees and the flowers and the trees: Lessons from
genetic mapping of sex determination in plants and animals. Genetics 186:9–31.
Chou, J.-Y ., and J.-Y . Leu. 2010. Speciation through cytonuclear incompatibility: Insights from yeast and
implications for higher eukaryotes. BioEssays 32:401–411.
Coyne, J., and H. Orr. 1997. “Patterns of speciation inDrosophila” revisited. Evolution 51:295–303.
Coyne, J. A., and H. A. Orr. 1989a. Patterns of speciation inDrosophila. Evolution 43:362–381.
Coyne, J. A., and H. A. Orr, 1989b. Speciation and Its Consequences, Chapter: Two rules of speciation,
pages 180–207 . Sinauer, Sunderland, MA.
Coyne, J. A., and H. A. Orr. 2004. Speciation. Sinauer, Sunderland, MA.
Darriba, D., G. Taboada, R. Doallo, and D. Posada. 2012. jModelTest2: More models, new heuristics and
parallel computing. Nature Methods 9:772.
Demuth, J. P. 2014. Do sex chromosomes affect speciation rate? BioEssays 36:632.
Dethier, M. N. 1980. Tidepools as refuges: predation and the limits of the harpacticoid copepod Tigriopus
californicus (Baker). Journal of Experimental Marine Biology and Ecology 42:99–111.
Dobzhansky, T. 1937. Genetics and the Origin of Species. Columbia University Press, New York.
Dufresnes, C., T. Majtyka, S. J. E. Baird, J. F. Gerchen, A. Borzee, R. Savary, M. Ogielska, N. Perrin, and
M. Stock. 2016. Empirical evidence for large X-effects in animals with undifferentiated sex chromo-
somes. Scientific Reports 6:21029.
Dunn, C. W., and J. F. Ryan. 2015. The evolution of animal genomes. Current Opinions in Genetics and
Development 35:25–32.
Edgar, R. C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput.
Nucleic Acids Research 32:1792–1797.
Edmands, S. 1999. Heterosis and outbreeding depression in interpopulation crosses spanning a wide range
of divergence. Evolution 53:1757–1768.
Edmands, S. 2001. Phylogeography of the intertidal copepod Tigriopus californicus reveals substantially
reduced population differentiation at northern latitudes. Molecular Ecology 10:1743–1750.
148
Edmands, S. 2002. Does parental divergence predict reproductive compatibility? Trends in Ecology and
Evolution 17:520–527.
Edmands, S., S. Northrup, and A. Hwang. 2009. Maladapted gene complexes within populations of the
intertidal copepodTigriopuscalifornicus? Evolution 63:2184–2192.
Ellegren, H. 2009. Genomic evidence for a large-Z effect. Proceedings of the Royal Society B: Biological
Sciences 276:361.
Ellison, C. K., and R. S. Burton. 2006. Disruption of mitochondrial function in interpopulation hybrids of
Tigriopuscalifornicus. Evolution 60:1382–1391.
Ellison, C. K., and R. S. Burton. 2008a. Genotype-dependent variation of mitochondrial transcriptional
profiles in interpopulation hybrids. Proceedings of the National Academy of Sciences of the United
States of America 105:15831–15836.
Ellison, C. K., and R. S. Burton. 2008b. Interpopulation hybrid breakdown maps to the mitochondrial
genome. Evolution 62:631–638.
Ellison, C. K., and R. S. Burton. 2010. Cytonuclear conflict in interpopulation hybrids: The role of RNA
polymerase in mtDNA transcription and replication. Journal of Evolutionary Biology 23:528–538.
Ellison, C. K., O. Niehuis, and J. Gadau. 2008. Hybrid breakdown and mitochondrial dysfunction in hybrids
ofNasonia parasitoid wasps. Journal of Evolutionary Biology 21:1844–1851.
Eo, S. H., and J. A. DeWoody. 2010. Evolutionary rates of mitochondrial genomes correspond to diversifi-
cation rates and to contemporary species richness in birds and reptiles. Proceedings of the Royal Society
B: Biological Sciences 277:3587–92.
Fisher, R. 1930. TheGeneticalTheoryofNaturalSelection. Clarendon Press, Oxford, UK.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, and S. Edmands. 2013. Postzygotic isolation involves
strong mitochondrial and sex–specific effects in Tigriopus californicus, a species lacking heteromorphic
sex chromosomes. Heredity 111:391–401.
Foley, B. R., C. G. Rose, D. E. Rundle, W. Leong, G. W. Moy, R. S. Burton, and S. Edmands. 2011.
A gene-based SNP resource and linkage map for the copepod Tigriopus californicus. BMC Genomics
12:568.
Fox, J., and S. Weisberg. 2011. An R Companion to Applied Regression. 2
nd
edition edition. Sage.
Franco, M., P. Rubini, and M. Vecchi. 1982. Sex–determinants and their distribution in various populations
ofMuscadomestica L. of Western Europe. Genetical Research 40:279–293.
Ganz, H., and R. Burton. 1995. Genetic differentiation and reproductive incompatibility among Baja
California populations of the copepodTigriopuscalifornicus. Marine Biology 123:821–827.
Gemmell, N. J., V . J. Metcalf, and F. W. Allendorf. 2004. Mother’s curse: the effect of mtDNA on individual
fitness and population viability. Trends in Ecology and Evolution 19:238–244.
Ghiselli, F., L. Milani, P. L. Chang, D. Hedgecock, J. P. Davis, S. V . Nuzhdin, and M. Passamonti. 2012.
De novo assembly of the Manila clamRuditapesphilippinarum transcriptome provides new insights into
expression bias, mitochondrial doubly uniparental inheritance and sex determination. Molecular Biology
and Evolution 29:771–786.
149
Graves, J. A. M. in press. Did sex chromosome turnover promote divergence of the major mammal groups?
BioEssays 38.
Green, D. 1988. Cytogenetics of the endemic New Zealand frog, Leiopelma hochstetteri: extraordinary
supernumerary chromosome variation and a unique sex-chromosome system. Chromosoma 97:55–70.
Green, D., C. Zeyl, and T. Sharbel. 1993. The evolution of hypervariable sex and supernumerary (B)
chromosomes in the relict New Zealand frog, Leiopelma hochstetteri. Journal of Evolutionary Biology
6:417–441.
Guindon, S., and O. Gascuel. 2003. A simple, fast and accurate method to estimate large phylogenies by
maximum-likelihood. Systematic Biology 52:696–704.
Haldane, J. 1922. Sex ratio and unisexual sterility in hybrid animals. Journal of Genetics 12:101–109.
Handschumacher, L., M. B. Steinarsdottir, S. Edmands, and A. Ingolfsson. 2010. Phylogeography of the
rock-pool copepodTigriopusbrevicornis (Harpacticoida) in the northern North Atlantic, and its relation-
ship to other species of the genus. Mar Biol 157:1357–1366.
Harrison, J., and R. Burton. 2006. Tracing hybrid incompatibilities to single amino acid substitutions.
Molecular Biology and Evolution 23:559–564.
Harrison, J. S., and S. Edmands. 2006. Chromosomal basis of viability differences inTigriopuscalifornicus
interpopulation hybrids. Journal of Evolutionary Biology 19:2040–2051.
Hausdorf, B. 2011. Progress toward a general species concept. Evolution 65:923–931.
Hendry, A. P., D. I. Bolnick, D. Berner, and C. L. Peichel. 2009. Along the speciation continuum in
sticklebacks. Journal of Fish Biology 75:2000–2036.
Hulsenbeck, J. P., F. Ronquist, R. Nielsen, and J. P. Bollback. 2001. Bayesian inference of phylogeny and
its impact on evolutionary biology. Science 294:2310–2314.
Hurtado, L. A., M. Mateos, and C. A. Santamaria. 2010. Phylogeography of supralittoral rocky intertidal
Ligia isopods in the Pacific region from Central California to Central Mexico. PLoS One 5:e11633.
Huys, R., J. Mackenzie-Dodds, and J. Llewellyn-Hughes. 2009. Cancrincolidae (Copepoda, Harpacticoida)
associated with land crabs: A semiterrestrial leaf of the ameirid tree. Molecular Phylogenetics and Evo-
lution 51:143–156.
Hwang, A., S. Northrup, J. Alexander, K. V o, and S. Edmands. 2011. Long-term experimental hybrid
swarms between moderately incompatible Tigriopus californicus populations: hybrid inferiority in early
generations yields to hybrid superiority in later generations. Conservation Genetics 12:895–909.
Hwang, A., S. Northrup, D. Peterson, Y . Kim, and S. Edmands. 2012. Long-term experimental hybrid
swarms between nearly incompatible Tigriopus californicus populations: persistent fitness problems and
assimilation by the superior population. Conservation Genetics 13:567–579.
Irwin, D. E. 2002. Phylogeographic breaks without geographic barriers to gene flow. Evolution 56:2283–
2394.
Ki, J.-S., K.-W. Lee, H. G. Park, S. Chullasorn, H.-U. Dahms, and J.-S. Lee. 2009. Phylogeography of the
copepodTigriopusjaponicus along the Northwest Pacific rim. J Plankton Res 31:209–221.
150
Kitano, J., and C. L. Peichel. 2012. Turnover of sex chromosomes and speciation in fishes. Environmental
Biology of Fishes 94:549–558.
Kitano, J., J. A. Ross, S. Mori, M. Kume, F. C. Jones, Y . F. Chan, D. M. Absher, J. Grimwood, J. Schmutz,
R. M. Myers, D. M. Kingsley, and C. L. Peichel. 2009. A role for a neo-sex chromosome in stickleback
speciation. Nature 461:1079–1083.
Klaus, B., and K. Strimmer, 2015. fdrtool: Estimation of (Local) False Discovery Rates and Higher Criti-
cism, version 1.2.15 edition.
Koevoets, T., and L. W. Beukeboom. 2008. Genetics of postzygotic isolation and Haldane’s rule in hap-
lodiploids. Heredity 102:16–23.
Kofler, R., R. V . Pandey, and C. Schl¨ otterer. 2011. PoPoolation2: identifying differentiation between
populations using sequencing of pooled DNA samples (Pool-Seq). Bioinformatics 27:3435–6.
Larkin, M., G. Blackshields, N. Brown, R. Chenna, P. McGettigan, H. McWilliam, F. Valentin, I. Wallace,
A. Wilm, R. Lopez, J. Thompson, T. Gibson, and D. Higgins. 2007. Clustal W and Clustal X version 2.0.
Bioinformatics 23:2947–2948.
Lazzaretto, I., and A. Libertini. 1985. Karyological comparisons among three species of the genusTigriopus
(Copepods, Harpacticoida). Atti. Acc. Lincei Rend. fis. 126:377–381.
Leach´ e, A. D., S. C. Crews, and M. J. Hickerson. 2007. Two waves of diversification in mammals and
reptiles of Baja California revealed by hierarchical Bayesian analysis. Biology Letters 3:646–650.
Lee, K., H. Dahms, and J. Lee. 2012. Multigenerational life history traits of the copepod Tigriopus from
the Northwest Pacific rim in crossbreeding experiments. Journal of Experimental Marine Biology and
Ecology 436–437:56–62.
Lenormand, T., and J. Dutheil. 2005. Recombination difference between sexes: a role for haploid selection.
PLoS Biology 3:e63.
Li, C., Z. Wang, and J. Zhang. 2013. Toward genome-wide identification of Bateson-Dobzhansky-Muller
incompatibilities in yeast: A simulation study. Genome Biology and Evolution 5:1261–1272.
Li, H. 2011. A statistical framework for SNP calling, mutation discovery, association mapping and popula-
tion genetical parameter estimation from sequencing data. Bioinformatics 27:2987–2993.
Li, H., and R. Durbin. 2009. Fast and accurate short read alignment with Burrows–Wheeler transform.
Bioinformatics 25:1754–1760.
Liew, W., R. Bartfai, Z. Lim, R. Sreenivasan, K. Siegfried, and L. Orban. 2012. Polygenic sex determination
system in zebrafish. PLoS One 7:e34397.
Lijtmaer, D., B. Mahler, and P. Tubaro. 2003. Hybridization and postzygotic isolation patterns in pigeons
and doves. Evolution 57:1411–1418.
Lima, T. G. 2014. Higher levels of sex chromosome heteromorphism are associated with markedly stronger
reproductive isolation. Nature Communications 5:4743.
Locey, K. J., and J. T. Lennon. 2016. Scaling laws predict global microbial diversity. Proceedings of the
National Academy of Sciences 113:5970–5975.
151
Mank, J., and J. Avise. 2009. Evolutionary Diversity and Turn-Over of Sex Determination in Teleost Fishes.
Sexual Development 3:60–67.
Masly, J. P., and D. C. Presgraves. 2007. High-resolution genome-wide dissection of the two rules of
speciation inDrosophila. PLoS Biology 5:1890–1898.
Matsubara, K., H. Tarui, M. Toriba, K. Yamada, C. Nishida-Umehara, K. Agata, and Y . Matsuda. 2006.
Evidence for different origin of sex chromosomes in snakes, birds, and mammals and step-wise differen-
tiation of snake sex chromosomes. Proceedings of the National Academy of Sciences 103:18190–18195.
Mayr, E. 1942. Systematics and the Origin of Species. Columbia University Press, New York.
McDermott, S., and M. Noor. 2010. The role of meiotic drive in hybrid male sterility. Philosophical
Transactions of the Royal Society B: Biological Sciences 365:1265–1272.
Mora, C., D. P. Tittensor, S. Adl, A. G. B. Simpson, and B. Worm. 2011. How many species are there on
Earth and in the ocean? PLoS Biology 9:e1001127.
Moyle, L., and E. Graham. 2005. Genetics of hybrid incompatibility betweenLycopersiconesculentum and
L.hirsutum. Genetics 169:355–373.
Moyle, L. C., and T. Nakazato. 2010. Hybrid incompatibility “snowballs” betweenSolanum species. Science
329:1521–3.
Muller, H. J., 1940. The New Systematics, Chapter: Bearing of the Drosophila work on systematics, pages
185–268 . Clarendon Press.
Nagaraja, J. Nagaraju, and H. A. Ranganath. 2004. Molecular phylogeny of the nasuta subgroup of
Drosophila based on 12S rRNA, 16S rRNA, and CoI mitochondrial genes, RAPD and ISSR polymor-
phisms. Genes & Genetic Systems 79:293–299.
Naisbit, R. E., C. D. Jiggins, M. Linares, C. Salazar, and J. Mallet. 2002. Hybrid sterility, Haldane’s rule
and speciation inHeliconiuscydno andH.melpomene. Genetics 161:1517–1526.
Ooijen, J. W. V ., 2006. Join Map 4, Software for the calculation of genetic linkage maps in experimental
populations. Kyazma B.V ., Wageningen, Netherlands.
Palmer, C., and S. Edmands. 2000. Mate choice in the face of both inbreeding and outbreeding depression
in the intertidal copepodTigriopuscalifornicus. Marine Biology 136:693–698.
Parnell, N., and J. Streelman. 2013. Genetic interactions controlling sex and color establish the potential for
sexual conflict in Lake Malawi cichlid fishes. Heredity 110:239–246.
Pereira, R. J., F. S. Barreto, and R. S. Burton. 2014. Ecological novelty by hybridization: Experimental
evidence for increased thermal tolerance by transgressive segregation inTigriopuscalifornicus. Evolution
68:204–215.
Pereira, R. J., F. S. Barreto, N. T. Pierce, M. Carneiro, and R. S. Burton. 2016. Transcriptome-wide patterns
of divergence during allopatric evolution. Molecular Ecology 25:1478–1493.
Peterson, D., K. Kubow, M. Connolly, L. Kaplan, M. Wetkowski, W. Leong, B. Phillips, and S. Edmands.
2013. Reproductive and phylogenetic divergence of tidepool copepod populations across a narrow geo-
graphical boundary in Baja California. Journal of Biogeography 40:1664–1675.
152
Phillips, B. C., and S. Edmands. 2012. Does the speciation clock tick more slowly in the absence of
heteromorphic sex chromosomes? BioEssays 34:166–169.
Pokorn´ a, M., and L. Kratochv´ ıl. 2009. Phylogeny of sex–determining mechanisms in squamate reptiles: are
sex chromosomes an evolutionary trap? Zoological Journal of the Linnean Society 156:168–183.
Powlik, J., 1996. Ecology of Tigriopus californicus (Copepoda, Harpacticoida) in Barkley Sound, British
Columbia. Ph.D. thesis, University of British Columbia.
Presgraves, D. C. 2002. Patterns of postzygotic isolation in Lepidoptera. Evolution 56:1168–83.
Presgraves, D. C. 2008. Sex chromosomes and speciation inDrosophila. Trends in Genetics 24:336–343.
Presgraves, D. C. 2010. The molecular evolutionary basis of species formation. Nature Reviews Genetics
11:175–180.
Presgraves, D. C., and H. A. Orr. 1998. Haldane’s rule in taxa lacking a hemizygous X. Science 282:952–
954.
Price, T., and M. Bouvier. 2002. The evolution of F
1
postzygotic incompatibilities in birds. Evolution
56:2083–2089.
Pritchard, V ., L. Dimond, J. Harrison, C. Velasquez, J. Zieba, R. Burton, and S. Edmands. 2011. Interpopu-
lation hybridization results in widespread viability selection across the genome inTigriopuscalifornicus.
BMC Genetics 12:54.
Pritchard, V ., and S. Edmands. 2012. The genomic trajectory of hybrid swarms: Outcomes of repeated
crosses between populations ofTigriopuscalifornicus. Evolution 67:774–791.
Pritchard, V ., V . Knutson, M. Lee, J. Zeiba, and S. Edmands. 2013. Fitness and morphological outcomes of
many generations of hybridization in the copepodTigriopuscalifornicus. Journal of Evolutionary Biology
26:416–433.
Purcell, S., B. Neale, K. Todd-Brown, L. Thomas, M. A. Ferreira, D. Bender, J. Maller, P. Sklar, P. I.
De Bakker, M. J. Daly, et al. 2007. PLINK: a tool set for whole-genome association and population-
based linkage analyses. The American Journal of Human Genetics 81:559–575.
Ramachandra, N., and H. Ranganath. 1986. The chromosomes of two Drosophila races: D. nasuta nasuta
andD.n.albomicana: IV . Hybridization and karyotype repatterning. Chromosoma 93:243–248.
Rand, D., R. Haney, and A. Fry. 2004. Cytonuclear coevolution: the genomics of cooperation. Trends in
Ecology and Evolution 19:645–653.
Rice, W. 1986. On the instability of polygenic sex determination: The effect of sex-specific selection.
Evolution 40:633–639.
Rice, W. 1992. Sexually antagonistic genes: experimental evidence. Science 256:1436–1439.
Rice, W. 1994. Degeneration of a nonrecombining chromosome. Science 263:230–232.
Riddle, B. R., D. J. Hafner, L. F. Alexander, and J. R. Jaeger. 2000. Cryptic vicariance in the historical
assembly of a Baja California Peninsular Desert biota. Proceedings of the National Academy of Sciences
97:14438–14443.
153
Rieseberg, L. 2001. Chromosomal rearrangements and speciation. Trends in Ecology and Evolution
16:351–358.
Rieseberg, L. H., M. A. Archer, and R. K. Wayne. 1999. Transgressive segregation, adaptation and specia-
tion. Heredity 83:363–372.
Rieseberg, L. H., and B. K. Blackman. 2010. Speciation genes in plants. Annals of Botany 106:439–55.
Roberts, R. B., J. R. Ser, and T. D. Kocher. 2009. Sexual conflict resolved by invasion of a novel sex
determiner in Lake Malawi cichlid fishes. Science 326:998–1001.
Robinson, J. T., H. Thorvaldsd´ ottir, W. Winckler, M. Guttman, E. S. Lander, G. Getz, and J. P. Mesirov.
2011. Integrative genomics viewer. Nature Biotechnology 29:24–26.
Rodrigues, N., C. Betto-Colliard, H. Jourdan-Pineau, and N. Perrin. 2013. Within-population polymorphism
of sex-determination systems in the common frog (Rana temporaria). Journal of Evolutionary Biology
26:1569–1577.
Rodrigues, N., Y . Vuille, A. Brelsford, J. Merila, and N. Perrin. 2016. The genetic contribution to sex deter-
mination and number of sex chromosomes vary among populations of common frogs (Ranatemporaria).
Heredity 117:25–32.
Rodrigues, N., Y . Vuille, and N. Perrin. 2015. Sex-chromosome differentiation and ‘sex races’ in the com-
mon frog (Ranatemporaria). Proceedings of the Royal Society B: Biological Sciences 282:20142726.
Rothfels, K. 1989. Speciation in black flies. Genome 32:500–509.
Sasa, M., P. Chippindale, and N. A. Johnson. 1998. Patterns of postzygotic isolation in frogs. Evolution
52:1811–1820.
Schilthuizen, M., M. C. W. G. Giesbers, and L. W. Beukeboom. 2011. Haldane’s rule in the 21st century.
Heredity 107:95–102.
Schneider, C. A., W. S. Rasband, and K. W. Eliceiri. 2012. NIH Image to ImageJ: 25 years of image
analysis. Nature Methods 9:671–675.
Schuelke, M. 2000. An economic method for the fluorescent labeling of PCR fragments. Nature Biotech-
nology 18:233–234.
Ser, J., R. Roberts, and T. Kocher. 2010. Multiple interacting loci control sex determination in Lake Malawi
cichlid fish. Evolution 64:486–501.
Sharbel, T., and D. Green. 1998. B-chromosome origin in the endemic New Zealand frog Leiopelma
hochstetteri through sex chromosome devolution. Genome 41:14–22.
Shaw, K. L., and S. P. Mullen. 2014. Speciation continuum. Journal of Heredity 105:741–742.
Shields, G. F. 2013. The significance of sex–linked chromosomal inversions in the speciation process of
the Simulium arcticum complex of black flies (Diptera: Simuliidae). Monographs of the Western North
American Naturalist 6:64–86.
Smith, S., C. Turbill, and F. Suchentrunk. 2010. Introducing mother’s curse: low male fertility associated
with an imported mtDNA haplotype in a captive colony of brown hares. Molecular Ecology 19:36–43.
154
Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and S. Kumar. 2011. MEGA5: Molecular evo-
lutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony
methods. Molecular Biology and Evolution 28:2731–2739.
Tanuja, M., N. Ramachandra, and H. Ranganath. 1999. Evolution of a recent neo-Y sex chromosome in a
laboratory population ofDrosophila. Journal of Genetics 78:81–85.
Tao, Y ., and D. Hartl. 2003. Genetic dissection of hybrid incompatibilities between Drosophila simulans
and D. mauritiana: III. Heterogeneous accumulation of hybrid incompatibilities, degree of dominance,
and implications for Haldane’s rule. Evolution 57:2580–2598.
Team, R. C., 2015. R: a language and environment for statistical computing.
The Tree of Sex Consortium. 2014. Sex determination: Why so many ways of doing it? PLoS Biology
12:e1001899.
Tittensor, D. P., C. Mora, W. Jetz, H. K. Lotze, D. Ricard, E. V . Berghe, and B. Worm. 2010. Global patterns
and predictors of marine biodiversity across taxa. Nature 466:1098–1101.
Tubaro, P., and D. Lijtmaer. 2002. Hybridization patterns and the evolution of reproductive isolation in
ducks. Biological Journal of the Linnean Society 77:193–200.
Turelli, M., and D. Begun. 1997. Haldane’s rule and X-chromosome size in Drosophila. Genetics 147:1799–
1815.
Turelli, M., and L. C. Moyle. 2007. Asymmetric postmating isolation: Darwin’s corollary to Haldane’s rule.
Genetics 176:1059–1088.
Turelli, M., and H. Orr. 1995. The dominance theory of Haldane’s rule. Genetics 140:389–402.
Uricaru, R., G. Rizk, V . Lacroix, E. Quillery, O. Plantard, R. Chikhi, C. Lemaitre, and P. Peterlongo. 2015.
Reference-free detection of isolated SNPs. Nucleic acids research 43:e11–e11.
Vacquier, V . 1962. Hydrostatic pressure has a selective effect on the copepod Tigriopus. Science 135:724–
725.
Vacquier, V ., and W. Belser. 1965. Sex conversion induced by hydrostatic pressure in the marine copepod
Tigriopuscalifornicus. Science 150:1619–1621.
Van der Auwera, G. A., M. O. Carneiro, C. Hartl, R. Poplin, G. del Angel, A. Levy-Moonshine, T. Jordan,
K. Shakir, D. Roazen, J. Thibault, E. Banks, K. V . Garimella, D. Altshuler, S. Gabriel, and M. A. DePristo.
2013. From FastQ Data to High-Confidence Variant Calls: The Genome Analysis Toolkit Best Practices
Pipeline. Current Protocols in Bioinformatics .
van Doorn, G., and M. Kirkpatrick. 2007. Turnover of sex chromosomes induced by sexual conflict. Nature
449:909–912.
van Doorn, G., and M. Kirkpatrick. 2010. Transitions between male and female heterogamety caused by
sex–antagonistic selection. Genetics 186:629–645.
Vandeputte, M., M. Dupont-Nivet, H. Chavanne, and B. Chatain. 2007. A polygenic hypothesis for sex
determination in the European sea bassDicentrarchuslabrax. Genetics 176:1049–1057.
155
Vavrek, M. J. Palaeontologia Electronica. fossil: Palaeoecological and palaeogeographical analysis tools.
2011 14:1.
Vicoso, B., and D. Bachtrog. 2015. Numerous transitions of sex chromosomes in Diptera. PLoS Biology
13:e1002078.
Vicoso, B., J. J. Emerson, Y . Zektser, S. Mahajan, and D. Bachtrog. 2013. Comparative sex chromosome
genomics in snakes: Differentiation, evolutionary strata, and lack of global dosage compensation. PLoS
Biology 11:e1001643.
Vitt, L. J., and J. P. Caldwell. 2009. Herpetology.
Vittor, B. A., 1971. Effects of the environment on fitness-related life history characters inTigriopuscalifor-
nicus. Ph.D. thesis, University of Oregon.
V oordouw, M., and B. Anholt. 2002a. Environmental sex determination in a splash pool copepod. Biological
Journal of the Linnean Society 76:511–520.
V oordouw, M., and B. Anholt. 2002b. Heritability of sex tendency in a harpacticoid copepod, Tigriopus
californicus. Evolution 56:1754–1763.
V oordouw, M., H. Robinson, and B. Anholt. 2005a. Paternal inheritance of the primary sex ratio in a
copepod. Journal of Evolutionary Biology 18:1304–1314.
V oordouw, M., H. Robinson, G. Stebbins, A. Albert, and B. Anholt. 2005b. Larval density and the Charnov-
Bull model of adaptive environmental sex determination in a copepod. Canadian Journal of Zoology
83:943–954.
V oordouw, M. J., G. Stebbins, H. E. Robinson, M.-J. Perrot-Minnot, T. Rigaud, and B. R. Anholt. 2008.
Genetic variation in the primary sex ratio in populations of the intertidal copepod,Tigriopuscalifornicus,
is widespread on Vancouver Island. Evolutionary Ecology Research 10:1007–1023.
Warnes, G. R., B. Bolker, T. Lumley, and R. C. Johnson, 2015. gmodels: Various R Programming Tools for
Model Fitting, r package version 2.16.2 edition.
Wexler, J. R., D. C. Plachetzki, and A. Kopp. 2014. Pan-metazoan phylogeny of the DMRT gene family: a
framework for functional studies. Development Genes and Evolution 224:175–181.
Willett, C., and R. Burton. 2001. Viability of cytochrome C genotypes depends on cytoplasmic backgrounds
inTigriopuscalifornicus. Evolution 55:1592–1599.
Willett, C., and R. Burton. 2003. Environmental influences on epistatic interactions: Viabilities of cy-
tochrome c genotypes in interpopulation crosses. Evolution 57:2286–2292.
Willett, C. S. 2006. Deleterious epistatic interactions between electron transport system protein-coding loci
in the copepodTigriopuscalifornicus. Genetics 173:1465–1477.
Willett, C. S. 2008a. No evidence for faster male hybrid sterility in population crosses of an intertidal
copepod (Tigriopuscalifornicus). Genetica 133:129–136.
Willett, C. S. 2008b. Significant Variation for Fitness Impacts of ETS Loci in Hybrids between Populations
ofTigriopuscalifornicus. Journal of Heredity 99:56–65.
156
Willett, C. S. 2011. Complex deleterious interactions associated with malic enzyme may contribute to
reproductive isolation in the copepodTigriopuscalifornicus. PLoS One 6:e21177.
Willett, C. S., and J. N. Berkowitz. 2007. Viability effects and not meoitic drive cause dramatic departures
from Mendelian inheritance for malic enzyme in hybrids of Tigriopus californicus populations. Journal
of Evolutionary Biology 20:1196–1205.
Willett, C. S., T. G. Lima, I. Kovaleva, and L. Hatfield. in press. Chromosome-wide impacts on the
expression of incompatibilities in hybrids ofTigriopuscalifornicus. G3: Genes, Genomes, Genetics .
Wilson, C. A., S. K. High, B. M. McCluskey, A. Amores, Y . lin Yan, T. A. Titus, J. L. Anderson, P. Batzel,
M. J. C. III, M. Schartl, and J. H. Postlethwait. 2014. Wild sex in zebrafish: Loss of the natural sex
determinant in domesticated strains. Genetics 198:1291–1308.
Wolff, J., E. Ladoukakis, J. Enriquez, and D. Dowling. 2014. Mitonuclear interactions: evolutionary con-
sequences over multiple biological scales. Philosophical Transactions of the Royal Society B: Biological
Sciences 369:20130443.
Wright, A. E., R. Dean, F. Zimmer, and J. E. Mank. 2016. How to make a sex chromosome. Nature
Communications 7:12087.
Wu, C., and A. Davis. 1993. Evolution of postmating reproductive isolation: the composite nature of
Haldane’s rule and its genetic bases. American Naturalist 142:187–212.
Wu, C.-I. 1992. A note on Haldane’s rule: Hybrid inviability versus hybrid sterility. Evolution 46:1584–
1587.
Wu, C.-I., N. A. Johnson, and M. F. Palopoli. 1996. Haldane’s rule and its legacy: Why are there so many
sterile males? Trends in Ecology and Evolution 11:281–284.
Yu, Y ., F. Lin, and H. Chang. 1999. Stepwise chromosome evolution inDrosophila. Heredity 83:39–45.
Yusa, Y . 2007. Causes of variation in sex ratio and modes of sex determination in the Mollusca-an overview.
American Malacological Bulletin 23:89–98.
Yusa, Y ., S. Breton, and W. Hoeh. 2013. Population genetics of sex determination in Mytilus mussels:
Reanalyses and a model. Journal of Heredity 104:380–385.
Zheng, X., D. Levine, J. Shen, S. Gogarten, C. Laurie, and B. Weir. 2012. A high-performance computing
toolset for relatedness and principal component analysis of SNP data. Bioinformatics 28:3326–3328.
Zhou, Q., J. Zhang, D. Bachtrog, N. An, Q. Huang, E. D. Jarvis, M. T. P. Gilbert, and G. Zhang. 2014.
Complex evolutionary trajectories of sex chromosomes across bird taxa. Science 346:1332.
157

Abstract (if available)

Abstract The processes and patterns underlying speciation are a major focus of evolutionary biology. These processes often result in the evolution of reproductive isolation, and sex chromosomes often play an important role in postzygotic reproductive isolation. However, sex chromosomes—and, more generally, sex determination mechanisms—are themselves highly evolutionarily labile, and many lineages lack sex chromosomes but still manage to speciate. However, the vast majority of work on the evolution of postzygotic reproductive isolation has been conducted in taxa that do possess sex chromosomes, leaving many open questions regarding how postzygotic isolation may evolve in taxa without sex chromosomes. This dissertation examines how the speciation process may differ in taxa that do not have sex chromosomes. ❧ Chapter 1 asks whether the presence of sex chromosomes in a clade accelerates speciation. Sex chromosomes disproportionately accumulate postzygotic incompatibilities (large X–effect) and contribute to sex–specific hybrid inviability and infertility (Haldane’s rule). Therefore, in the absence of sex chromosomes, postzygotic reproductive isolation may take longer to accumulate to the same level, which could ultimately result in slower speciation rates. We addressed this question in reptiles, where we found that speciation intervals are shorter in snakes and lizards (where sex chromosomes are common) than in turtles and crocodiles (where sex chromosomes are rare or absent). ❧ Chapters 2–4 examine phylogeography, hybrid incompatibility, and sex determination in the intertidal copepod Tigriopus californicus, a species with polygenic (genetic but non-chromosomal) sex determination. Chapter 2 presents a phylogeographic study of T. californicus across the species range. T. californicus populations have been documented from southern Alaska, USA to central Baja California, Mexico, and previous phylogeographic studies have employed only a single mitochondrial locus (mtCOI). We used a different mitochondrial locus (mt12S), as well as 7 nuclear loci (18S, ITS1, 5.8S, and 4 transcriptome–based loci) to resolve relationships between major clades and to clarify the taxonomic status of populations at the southern edge of the range, which are genetically and reproductively isolated from other T. californicus populations. We found definitive evidence that these southern populations constitute a novel species that is geographically continuous with T. californicus, identify the presence of major regional clades, and document a general isolation–by–distance pattern within regions. ❧ In Chapter 3, non-recombinant backcrosses between multiple pairs of T. californicus populations were used to map postzygotic reproductive isolation, morphometric features, and sex. We identified segregation distortion in some, but not all crosses and detected few epistatic incompatibilities. We mapped associations with sex to several chromosomes, consistent with polygenic sex determination, and we found evidence of transgressive segregation for morphometric traits in some crosses. In these crosses, we found that males consistently fare worse than females, despite not being heterogametic (the traditional explanations for sex-specific deleterious effects invoke heterogamety). Thus, explanations for sex-specific deleterious effects of hybridization may require the generation of alternative hypotheses in taxa without sex chromosomes. ❧ Finally, Chapter 4 investigates polygenic sex determination and the emergence of a novel sex chromosome in a hybrid population of T. californicus. Polygenic sex determination, where sex determination is genetic but non-chromosomal, is predicted to be unstable and prone to takeover by strong sex-determiners, which can then become sex chromosomes. We found that alleles from one parental population (SD, from San Diego, California, USA) had introgressed sex–specifically on one chromosome (chromosome 10), despite swamping by the other population’s alleles (SC, from Santa Cruz, California, USA) across the rest of the genome. SD alleles on chromosome 10 appear to have a strong feminizing effect, and—in the absence of meiotic recombination in females—to have been transmitted almost exclusively to females. Whole-genome re-sequencing of the parental progenitor populations did not reveal allele frequency differences between males and females to be concentrated anywhere in the genome, whereas in the hybrid population these differences were elevated on chromosome 10. The total evidence suggests that chromosome 10 is acting as a novel sex chromosome within the hybrid population but not within the parental populations, making it perhaps the first sex chromosome to emerge from a polygenic sex determination system in a laboratory. ❧ The speciation process is currently poorly understood outside of a few model taxa, where sex chromosomes are involved in most aspects of postzygotic reproductive isolation. Increasing the taxonomic breadth of this research is an imperative in the Anthropocene, where human activities are directly affecting the speciation process through anthropogenically–mediated hybridization and extinction. Speciation in the absence of sex chromosomes may proceed fundamentally differently, as depicted in Chapters 1, 2, and 3. Alternatively, as seen in Chapter 4, hybridization could in some cases lead to the emergence of a novel sex chromosome, which could then modify the speciation process.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The evolution of pollution tolerance in the marine copepod Tigriopus

PDF

Genetic architecture underlying variation in different traits in the Pacific oyster Crassostrea gigas

PDF

Sex differences in aging and the effects of mitochondria

PDF

Multiple generations of hybridization between populations of the intertidal copepod Tigriopus californicus

PDF

Molecular and behavioral mechanisms of circatidal biological rhythms in intertidal mollusks

PDF

Transgenerational inheritance of thermal tolerance in two coral species in the Florida Keys

PDF

Biological interactions on the behavioral, genomic, and ecological scale: investigating patterns in Drosophila melanogaster of the southeast United States and Caribbean islands

PDF

Genome-wide analysis of genetic load and larval mortality in a highly fecund marine invertebrate, the Pacific Oyster Crassostrea gigas

PDF

Computational and experimental approaches for the identification of genes and gene networks in the Drosophila sex-determination hierarchy

PDF

Innovative sequencing techniques elucidate gene regulatory evolution in Drosophila

PDF

Dynamics of protein metabolism in larvae of marine invertebrates

PDF

Application of evolutionary theory and methods to aquatic ecotoxicology

PDF

Applications of next generation sequencing in sessile marine invertabrates

PDF

Male-female conflict after mating: function and dynamics of the copulatory plug in mice (Mus domesticus)

PDF

The evolution of gene regulatory networks

PDF

Evolutionary genomic analysis in heterogeneous populations of non-model and model organisms

PDF

Thermal acclimation and adaptation of key phytoplankton groups and interactions with other global change variables

PDF

Robustness and stochasticity in Drosophila development

PDF

Unraveling the molecular mechanisms of heterochromatic double strand break repair in Drosophila cells

PDF

The structural and functional configurations of glutamate, GABA, and catecholamine pre-synaptic terminals in the parvicellular neuroendocrine part of the paraventricular nucleus of the hypothalamus

Asset Metadata

Creator Phillips, Barret C. (author)

Core Title Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the copepod Tigriopus californicus

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Marine Biology and Biological Oceanography

Publication Date 09/22/2018

Defense Date 06/06/2016

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

Tag marine Invertebrates,OAI-PMH Harvest,sex determination,speciation

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Edmands, Suzanne (committee chair), Conti, David V. (committee member), Dean, Matthew D. (committee member), Hedgecock, Dennis (committee member)

Creator Email barretphillips@gmail.com,bcphilli@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-303880

Unique identifier UC11281135

Identifier etd-PhillipsBa-4792.pdf (filename),usctheses-c40-303880 (legacy record id)

Legacy Identifier etd-PhillipsBa-4792.pdf

Dmrecord 303880

Document Type Dissertation

Format application/pdf (imt)

Rights Phillips, Barret C.

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