Understanding the development of sexually selected traits

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Understanding the Evolution and Development of Sexually Selected Traits
By
Caleb Roberto Ghione
A Dissertation presented to the
FACULTY OF USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MOLECULAR BIOLOGY)
DECEMBER 2024
Copyright 2024 Caleb Roberto Ghione

ii
Acknowledgements
First, I would like to begin by thanking so many people that have been with me through the thick
and thin of grad school. I have had the privilege of being able to get my PhD at the University of
Southern California under the tutelage of Dr. Matthew Dean, PhD. Matt, thank you so much for
your mentorship, your guidance, and your friendship. Thank you for allowing me to join your
lab. I’m grateful that we were able to work together and study incredible things! I would also like
to thank my entire committee, Sergey Nuzhdin, PhD., Berenice Benayoun, PhD., and Andrew
McMahon, PhD. I’m grateful for all of your help, comments, and guidance on my project
throughout my time in my PhD career. Your help was invaluable throughout my career.
Second, I would like to thank my cohort. We joined the MCB program at USC in 2017, and I
couldn’t have made it through this PhD as easily without you. Thank you, Meghan Petrie
Mendez, PhD., Joshua Park, Joseph Hale, PhD., Nicole Stuhr, PhD., Mezmur Belew, PhD., Dan
Ma, PhD., and Yiwei He, PhD. The seven of you were a fantastic group of friends to have during
my PhD career. We made it a point to have dinner and hangouts as much as possible and for that
I’m grateful. In particular thank you to my lunch crew, Meghan and Josh, you two were always
there for me, and I love you all very much. I would like to thank a couple of other friends that I
made throughout my PhD career. There are many of you to name, but I’m severely grateful. One
friend in particular, Anne Nguyen, PhD. Anne, thank you so much for your friendship, comfort,
and mentorship as a PhD student. You will always be my Mariah Carey! Dr. Maddelyn Harden,
thank you for helping me maintain my sanity throughout my PhD career. Our coffee dates and
chats will forever be in my heart. I would also like to thank the members of the Dean lab, past
and present, that welcomed me into their lab and offered me lifelong friendship. Nick Schultz,
PhD., Sara Keeble, PhD., and Michael Lough-Stevens, PhD., Thank you very much! To the
current students of the Dean lab, Charles Toney III, Maeve Secor, and Quinn Fagersten
(incoming student), you three are incredible people. Thank you for making the last two years of
PhD career in the Dean lab amazing and fun. I look forward to seeing the fantastic work you do
in the future.
Lastly, I would like to thank my friends and family that have been with me throughout my entire
life. Mama, Papa, Abuela, Josué, Alonso, Gita, Tia Marcella, Kris, Marissa, y familia, muchas
gracias por apoyarme durante mi tiempo en mi doctorado. Los amo bastante y mi carrera no
hubiera sido posible in ustedes. To the rest of my family, I love you all so much! Alyssa
Necesito, Camly Nguyen, Leilani Simmons, Wendy Ramirez, Andrea Muriel, PA., and others,
your friendship has been invaluable. You have been with me throughout most of my life. I would
not have been able to make it through my PhD career without your support. I love you all very
much.

iii
Table of Contents
Acknowledgements ....................................................................................................................... ii
List of Tables................................................................................................................................ vi
List of Figures.............................................................................................................................. vii
List of Supplementary Figures ................................................................................................. viii
Abstract......................................................................................................................................... ix
Chapter 1: Introduction ............................................................................................................... 1
1.1 Sexually Selected Traits........................................................................................................ 1
1.2 Why do sexually selected traits exist? .................................................................................. 2
1.3 What are the genetic pathways and modes of regulation underlying sexually selected
traits?........................................................................................................................................... 5
Conclusion .................................................................................................................................. 8
Chapter 2: Manuscript Status and Contribution.............................................................................. 9
Chapter 2: An inferred role of androgen signaling in the
evolution of sexual size dimorphism across mammals ............................................................ 10
Abstract..................................................................................................................................... 10
Introduction............................................................................................................................... 11
Materials & Methods ................................................................................................................ 12
2.1 Body Mass....................................................................................................................... 13
2.2 ARE and ERE counts across whole genomes ................................................................. 13
2.3 ARE and ERE counts near protein coding regions.......................................................... 14
2.4 Phylogenetically controlled linear modeling................................................................... 15
2.5 Genome quality................................................................................................................ 16
Results....................................................................................................................................... 16
2.6 Across whole genomes and including all species,
SSD correlated with body size, but not ARE or ERE counts................................................ 16
2.7 Near genes and in small-bodied species,
SSD correlated with ARE counts but not overall body size.................................................. 17
2.8 Gene-centric analyses revealed persistent, positive influence of ARE on SSD.............. 18
2.9 Genome quality................................................................................................................ 20
Discussion................................................................................................................................. 20
Figure legends.......................................................................................................................... 25
Figure 1.1.............................................................................................................................. 25
Figure 1.2.............................................................................................................................. 25

iv
Figure 1.3.............................................................................................................................. 25
Supplemental File 1.1 .......................................................................................................... 25
Supplemental File 1.2 .......................................................................................................... 25
Supplemental File 1.3 .......................................................................................................... 25
Supplemental File 1.4 .......................................................................................................... 25
Supplemental File 1.5 .......................................................................................................... 25
Supplemental File 1.6 .......................................................................................................... 25
Chapter 3: Manuscript Status and Contribution............................................................................ 48
Chapter 3: Androgen receptor (AR) Co-immuno precipitation
and sequencing (ChIP-seq) and single-cell RNA-sequencing identify
major differences in cells that give rise to the baculum (penis bone)
compared to cells that give rise to forelimb and hindlimb skeletal system. .......................... 49
Abstract..................................................................................................................................... 49
Introduction............................................................................................................................... 50
Materials and Methods.............................................................................................................. 53
3.1 Animals............................................................................................................................ 53
3.2 Determining sampling timepoint for sampling baculum precursors............................... 53
3.3 Determining sampling timepoint for sampling forelimb and hindlimb precursors......... 55
3.4 Androgen Receptor Chromatin Immunoprecipitation Sequencing (AR ChIP-seq)........ 56
3. 5 Single-cell dissociation and Library construction .......................................................... 57
3.5.1 Baculum Dissociation................................................................................................... 57
3.5.2 Forelimb and Hindlimb Dissociation ........................................................................... 58
3.5.3 Read Mapping............................................................................................................... 58
3.5.4 Seurat v5.1.0 Quality Control Metrics and Normalization........................................... 58
3.5.5 Comparing Chondrocytes Between Developing P6 Baculum
and Developing E15 Hindlimb and Forelimb Bone Systems................................................ 59
Results....................................................................................................................................... 60
3.6 Inferred ARE’s cluster near genes involved in bone and cartilage formation,
but only in developing mouse and rat penises....................................................................... 60
3.7 Analysis of single-cell RNA-sequencing of developing P6 mouse penis
and E15 mouse hindlimb and forelimb via Integration ......................................................... 63
Discussion................................................................................................................................. 65
Limitations of the study ............................................................................................................ 68
Acknowledgements................................................................................................................... 69

v
Figure Legends........................................................................................................................ 70
Figure 2.1.............................................................................................................................. 70
Figure 2.2.............................................................................................................................. 70
Supplementary Figure 2.1................................................................................................... 70
Table 2.1 ............................................................................................................................... 71
Table 2.2 ............................................................................................................................... 75
Chapter 4: Conclusion................................................................................................................ 79
Introduction............................................................................................................................... 79
Materials and Methods.............................................................................................................. 82
4.1 Genetic Disruption of the Baculum’s Development using
the Cre-Lox System to knock out the Androgen Receptor in
Runx2 Expressing Cells......................................................................................................... 82
4.2 Histology Analysis of FArRunx2-cre Mutant Baculum ....................................................... 86
4.3 MicroCT Scanning and Bone Morphometric Analysis of FArRunx2-cre
Mutant Baculum .................................................................................................................... 86
4.4 Assessing Male Mouse Mating Behavior........................................................................ 86
Results....................................................................................................................................... 87
4.5 Androgen Receptor Knockout via Cre-Lox in Runx2 Expressing
Cells Disrupts the Baculum’s Development in Mice ............................................................ 87
4.6 Mating Behavior of the FArRunx2-cre Mouse with the Underdeveloped Baculum ............ 88
Discussion................................................................................................................................. 88
Figure Legends........................................................................................................................ 92
Figure 3.1.............................................................................................................................. 92
References..................................................................................................................................... 94
Appendix.................................................................................................................................... 110
Appendix A. Male-derived copulatory plugs enhance implantation
success in female Mus musculus............................................................................................. 110
Appendix B. Complex genetics cause and constrain fungal persistence
in different parts of the mammalian body............................................................................... 122
Appendix C. The lifespan of corpora lutea in non-pregnant females
is positively correlated with gestation length.......................................................................... 139

vi
List of Tables
Table 1.1....................................................................................................................................... 29
Table 2.1....................................................................................................................................... 71
Table 2.2....................................................................................................................................... 75

vii
List of Figures
Figure 1.1 ..................................................................................................................................... 26
Figure 1.2 ..................................................................................................................................... 27
Figure 1.3 ..................................................................................................................................... 28
Figure 2.1 ..................................................................................................................................... 72
Figure 2.2 ..................................................................................................................................... 74
Figure 3.1 ..................................................................................................................................... 93

viii
List of Supplementary Figures
Supplementary Figure 1.1.......................................................................................................... 30
Supplementary Figure 1.2.......................................................................................................... 31
Supplementary Figure 1.3.......................................................................................................... 37
Supplementary Figure 1.5.......................................................................................................... 38
Supplementary Figure 1.6.......................................................................................................... 47
Supplementary Figure 2.1.......................................................................................................... 78

ix
Abstract
My dissertation centers around understanding the development of two important sexually
selected traits, sexual size dimorphism (SSD) and the baculum (penis bone). In my first chapter, I
am interested in answering the questions, 1) how does SSD develop and 2) how do two
individuals of opposite sex deal with the conflict of body size differences when they share a
similar genome? Via a meta-analysis, I test how androgen signaling might regulate SSD via their
genomic architecture hormone-DNA binding sites (ARE). In the study, I directly tested whether
there was a correlation between SSD and the number of AREs present across 268 mammalian
genomes. I found that SSD is positively correlated with the number of AREs present in the
genome but only when near genes and significantly in small-bodied mammals (rodents and bats).
More importantly, SSD is significantly positively correlated with the number of AREs present
near very specific genes called vomeronasal genes. This study expands on the knowledge of how
the conflict of SSD might be regulated via androgen signaling. In my second chapter, I am
interested in understanding the development of the baculum which is a unique bone that is
morphologically hyper diverse, has a unique evolutionary history of independent development
and loss across mammal lineages, and is androgen sensitive as it develops. More specifically, I’m
interested in answering the questions, 1) how does the baculum develop, 2) Are there any
androgen binding sites near bone genes during the baculum’s development, 3) are the genetic and
regulatory pathways between the baculum and other bone systems that do not have a unique
evolutionary history like fore- and hindlimbs the same or different, and 4) can we disrupt the
development of the baculum to better understand its function. To address these question, I use
co-immuno precipitation and sequencing, single-cell RNA sequencing, and the cre-lox system to
understand the development, regulation and developmental disruption of the baculum. My study

x
shows that there are some distinctions between the chondrocytes of the developing baculum and
the developing limb systems. In particular, the chondrocytes in the developing baculum heavily
express Ar. The aforementioned results describe a model for the baculum’s development in
which androgen signaling must be present to drive the transcription of bone development genes
so that the baculum can be made. I conclude my thesis by presenting a genetically engineered
mouse via the cre-lox system that has a drastically developmentally retarded baculum as a novel
model for studying the function of the baculum which has been heavily disputed among
scientists. The genetically mutated male mouse, known as FArRunx2-cre, has the androgen receptor
gene knocked down only in cells expressing RUNX2 which are predominantly bone and
cartilage cells. The FArRunx2-cre mice are morphologically and behaviorally normal to their
wildtype counterparts except for the disrupted baculum and their inability to intromit
successfully. At the current time of this dissertation, only one FArRunx2-cre male mouse has been
able to mate successfully. This final study opens up a plethora of questions about the function
and development of the baculum for future studies.

1
Chapter 1: Introduction
1.1 Sexually Selected Traits
In 1858/9 Charles Darwin coined the term ‘natural selection’ as a process by which a population
adapts and evolves to survive. While many variable character traits (e.g. beak size of Darwin’s
finches) are explained by natural selection, unorthodox character traits like the vibrant feathers of
a peacock, which could attract predators, did not make sense with the survival concept alone and
left the theory of natural selection incomplete. Because most animals reproduce sexually, Darwin
coined the term ‘sexual selection’, described as competition for mates where an individual has a
competitive advantage over other individuals, presenting this natural process in combination with
natural selection as an avenue for speciation. An animal’s advantage can manifest in
outcompeting a member of the same sex (intrasexual selection) or impressing a member of the
opposite sex via an attractive display or physical characteristic resulting in mating [1-3]. The
“advantage” that the competing individual has is reflected by some sort of weapon/ornament (in
some cases overly exaggerated), behavior, or molecular component (proteins, immune cells,
sperm, fluids, etc.) that allows them to outcompete members of the same sex or be chosen by
members of the opposite sex for mating. Advantageous character traits become sexually selected
and are passed down to their progeny, albeit dependent on the degree of advantage given to
reproducing individuals. The degree of advantage is a precarious system that requires a balance
between reproduction and survivability – essentially a trade-off of investment. Some character
traits produce loud and vibrant colors or large character traits which aid the animal in
reproducing, but they are traits that make the animal easily spotted by predators. [4, 5]. Such
types of character traits can be observed in the colorful plumage in birds [6, 7], horns, antlers in

2
both mammals and insects and canine teeth in mammals [8-12], chromatic displays in
fish/marine organisms and reptiles amongst many others [13-16], and elaborate mating
displays/courtship dances (behaviors) that attract opposite sex conspecifics (individuals of the
same species) [17]. While animals exhibit ornaments, weaponry, and other sexually selected
traits, many of these traits, size, color, shape, and appendages (horns, feathers, fins, and teeth),
tend to be exhibited in a sexually dimorphic manner, meaning that males and females from a
particular species exhibit differing physical characteristics or behaviors [3, 6, 18-20]. The large
amount of sexually dimorphic traits raises important questions of why they exist/possible
functions and how genetics influence their development.
1.2 Why do sexually selected traits exist?
Sexually selected traits are often influenced by intrasexual selection (male/male competition),
intersexual selection (male/female mate choice), and ecological factors. In any case, sexually
selected traits influence an individual’s reproductive success – the production of offspring
throughout a breeding season or lifetime.
Intrasexual selection can be easily observed in animals where fighting between males allows the
winner to mate with their female counterparts allowing the character trait of the winner to be
passed down to the next generation. In many cases, intrasexual selection is seen in animals
exhibiting polygynous mating systems. In these systems, males will mate with multiple females
usually after a male fights against other males and wins. In some cases, intrasexual selection
involves actual physical fighting [21], while others are observed internally via sperm/seminal
fluid related competition [22-25]. The latter are examples of selection occurring after mating but
before fertilization, post-copulatory, pre-zygotic selection, while the former is selection

3
occurring before mating, pre-copulatory selection. A notable example of intrasexual selection can
be seen between differently sized males of elephant seals normally resulting in the largest
elephant seal winning and mating with the multiple females. That then leads to the female
elephant seals choosing to reproduce with the largest and strongest male elephant seal [21, 26].
Subsequently, it has been proposed that animals with exaggerated size or structures/weapons,
such as horns, are used as aggressive signals. Typically, in males, these signals are meant to deter
other males from approaching their territory or harem allowing them to protect their females.
Animals like rams, deer, bovids, beetles and many others of the male sex use their horns to fight
off other males in mating displays for the females to choose the winning male to reproduce [12,
27, 28]. Interestingly, in deer, there was no correlation between large antlers, body size, and
winning/losing fights [29]. However, sperm competition acts as a secondary place where males
compete with each other to sire progeny. Sperm competition occurs inside the female
reproductive tract where sperm from two or more males compete using a variety of molecules,
shapes, and numbers to outcompete one another [30-34]. Sperm competition in mammals is
known to cause an increase in sperm number and sperm size to better compete with other males
[35]. Intrasexual selection seems to exaggerate sexually selected traits that are important for
reproduction.
On the female side of the equation, intersexual selection, otherwise known as mate choice, can
also influence sexually selected traits [36]. In this case, males typically compete with other males
using courtship displays or other tactics in hopes of being chosen by females. Courtship displays
typically include species specific mannerisms or structures which lead to species-species
recognition resulting in possible speciation [17, 37, 38]. Clear examples can be seen in peacock
spiders when the males “dance” to attract their female counterparts [39], or in mating displays

4
produced via visual signals by many species of birds of paradise [40]. Examples of speciesspecific structures can be observed when male bowerbirds build complicated bowers and even
decorate them to attract a female mate [41]. One of the most fascinating instances of mate choice
leading to speciation is observed in bearded manakins. Bearded manakins have a unique
acrobatic courtship display that has allowed them to evolve one of the fastest limb muscles
recorded. Miles et al. [42] observed that bearded manakins were able to specialize their limb
muscles extensively enough to cause divergence in courtship display speed making room for
sympatric speciation of the bearded manakins from other manakins in their genus Manacus. The
importance of speed lies in the female manakins ability to discriminate between millisecondscale differences in mating displays performed by their male counterparts [43]. While colorful
plumage and extravagant courtship displays can increase an animal’s chance of mating, the same
traits and others that are under sexual selection can actually showcase the robustness of an
individual’s genes, in other words, honest signals.
Honest signals are signals where the individuals displaying them cannot lie about the quality of
their genes. Many of the honest signals depend on good genes, proper care, and access to
resources. In cases where animals have horns, apart from using their horn structures as weapons,
healthy and robust horns can function as measurements for overall reproductive health and
access to resources [44]. Studies in fish, birds, and reptiles have found similar results in the color
of plumage and scales. Brighter more pronounced colors suggest that male individual is healthy
and has “good genes” or honest signals [45]. Studies have also shown that larger animals tend to
have more dominance than their subordinate counterparts allowing them to have access to more
resources. A male individual animal with access to more resources gives females more of a
reason to mate with a healthier, dominant male counterpart [44, 46, 47]. These examples are just

5
a few of many sexually selected traits that give us an idea of why they might exist. These
examples suggest that sexually selected traits are required for reproductive success and
speciation throughout the animal kingdom.
1.3 What are the genetic pathways and modes of regulation underlying sexually selected traits?
Intrasexual selective interactions lead to an organism evolving exaggerated weaponry or gametes
while intersexual selective interactions lead to an organism evolving greater variation, behaviors,
and elaborate ornaments [48]. A question important to our understanding of the development of
sexually selected traits is, what are the genetic pathways that make and maintain said structures?
Understanding the genotype to phenotype relationship behind sexually selected traits is
imperative to our understanding of variation, speciation, and overall biology.
1.3.1 Genetics Behind Intrasexual Selection
Intrasexual selection has led to many exaggerated character traits, but what are the genetic
mechanisms behind these traits? Many of the genetic underpinnings of exaggerated character
traits are species specific. For example, studies in Swordtail fish (swords on the tail of male fish
are attractive to females), genus Xiphophorus, have found that the “sword” on the tail fin of
males is a sexually selected trait driven via hormones acting on msx and fgfr1 genes just to name
a few [49]. Studies carried out in rhinoceros beetles have discovered novel genetic mechanisms
underlying the exaggeration of the horn as a sexually selected male weapon [9, 50]. Genetic
mutations in the domains of hh, wg, and dpp signals alters the location of the horn’s outgrowth.
Emlen et al. [27] dug deeper and found that many of the genes responsible for exaggeration in
horn growth in beetles were insulin sensitive [27].

6
1.3.2 Genetics Behind Intersexual Selection
Coloration in scales and feathers of fish, reptiles, and birds are prime examples of impressive
displays of color, signaling, and courtship displays for mate attraction, intersexual selection.
Genes such as coatomer protein complex and subunit zeta-1 (copz-1) have been identified as
important genes for coloration in cichlid fishes [51]. In bird species, the gene Eorix has been
identified as a gene responsible for carotenoid-signaling allowing for a proper mating display
[52]. Studies in Drosophila courtship behavior found that male and female reproductive
behaviors rely on a single regulatory gene called fruitless (fru) [53, 54]. The interesting aspect of
the fru gene is alternative mRNA splicing and has caused certain isoforms (P1 transcript) to only
form in males allowing for their male reproductive behavior to occur. If a female Drosophila
expresses the male isoform of fru, she will begin to exhibit male-specific behaviors [55, 56].
Functional studies have even found specific loci in Drosophila that contribute to the genetic
basis of male courtship behaviors [57]. In rodents, making and maintaining sex-specific
hormones is important for proper sex-specific mating behavior. Reviewed by Hull and
Dominguez [58], studies have shown that interfering with/knocking down testosterone and
dihydrotestosterone can heavily affect mating behavior of male rodents, however in some species
can be rescued by the administration of estradiol. The aforementioned studies lead us to question
the important role of genetic regulation for sexually selected traits.
1.3.3 Gene Regulation
While the above references are just some examples of the genes necessary for the development
of sexually selected traits, how these genes are regulated is an important part of the puzzle.
Without proper regulation of genes, sexually selected traits cannot flourish. From the examples

7
listed above, the insulin-like growth factor gene cascade/pathway has been shown to be one of
the fundamental ways that genes pertaining to sexual traits are regulated. In mice, studies have
shown that by artificially reducing insulin-like growth factor (IGF) — insulin like proteins that
are part of the mechanism by which an organism interacts with their physical environment [59]
— content, the individual’s body size was reduced by 30% relative to their wildtype littermates
[60].
Scientists have also found that sexually selected traits can be driven via sex-specific
developmental mechanisms [61]. In that case, sex-determination becomes an important avenue
of study which then leads to sex specific gene expression. Out of the well-known genes, tens of
thousands have dimorphic expression. In Drosophila, 57% of known genes showed some kind of
sex-bias [62]. In somatic tissue of mice over 10,000 genes showed sex-bias [63]. Studies have
shown that male-biased gene expression in certain mammalian tissues such as kidneys is
androgen dependent. Xiong et al. (2023) showed that removing AR activity was crucial for
establishing the male phenotype in kidney proximal tubule (PT) cells. They also observed an
enrichment in androgen binding near male-biased genes in their chromatin immunoprecipitation
and sequencing (ChIP-seq) analysis showcasing a mechanism for regulating the sex-biased genes
that establish the male phenotype in kidney PT cells [64]. Other studies have observed similar
androgen dependency for sex-biased gene expression across the animal kingdom. Kitano et al.
(2020) observed major androgen dependency for sex-biased genes in juvenile three-spined
stickleback fish. They hypothesized that androgen dependency was important during specific
developmental stages where a lot of gonadal steroid hormones are being expressed and growth is
occurring [65]. Seeing as how androgen molecules are more prominent in males of any species,
it makes sense that they might regulate their sex-biased genes using androgen signaling during a

8
time when development and growth are crucial. In large, sex-biased gene expression and
hormone regulation serve as mechanisms for regulating the development of sexually selected
traits.
Conclusion
The combination of mate choice, good genes, regulatory networks, and function make sexually
selected traits very interesting to study. For my dissertation, I focused on studying two sexually
selected traits – sexual size dimorphism (SSD) and the baculum (penis bone) found in many
male mammals. Both character traits are heavily studied by evolutionary biologists, but the
underlying genetics, mechanics, and functions of each are still currently unknown. In my
dissertation I aim to expand on the current knowledge of how SSD and the baculum might be
regulated.

9
Chapter 2: Manuscript Status and Contribution
The current manuscript entitled “An inferred role of androgen signaling in the evolution of
sexual size dimorphism across mammals” is under revision at the Journal of Evolution.
My contribution to this manuscript involved the entirety of the project. I contributed to the data
collection, data analysis, and manuscript writing.

10
Chapter 2: An inferred role of androgen signaling in the evolution of sexual size
dimorphism across mammals
Abstract
Sexual size dimorphism (SSD) is common throughout the animal kingdom, but its evolution is
puzzling because males and females largely share the same genome. Sex-specific regulation of
the genome via hormonal signaling offers one potential solution to this conundrum. Here we test
if/how SSD is correlated with the number of putative binding motifs for androgen and estrogen
receptors. Across 268 mammal species, we find almost no relationship between SSD and the
number of estrogen receptor binding motifs in their genome. In contrast, the magnitude of SSD is
positively correlated with the number of androgen receptor binding motifs, but only in relatively
small-bodied lineages (Chiroptera and Rodentia), and only when we focus our analyses on
regions near protein coding genes. In fact, ~90% of protein coding genes show a positive
relationship between the magnitude of SSD and the number of their androgen receptor binding
motifs. The most strongly correlated genes are enriched for chemical detection pathways,
including pheromone detection. We discuss these patterns in the context of different life history
features. We hypothesize that larger-bodied lineages have relatively longer periods of time to
accumulate SSD, making them less reliant on hormonal signaling and sex-specific regulation.
Key Words: androgen receptor binding motifs, sexual size dimorphism, androgen receptor,
hormonal signaling, genomes, sex-specific regulation

11
Introduction
Sexual size dimorphism, where males and females of the same species tend towards different
body size, is a ubiquitous feature across mammals [18, 66, 67]. Because body size correlates
with life history traits such as lifespan, metabolic rate, and developmental trajectories,
differences in size likely belie some kind of sexual conflict. In other words, an allele that
increases body size may be beneficial to the larger sex but deleterious to the smaller sex, and
vice versa [3, 67-70]. The long-term resolution to this sexual conflict over body size is a
fundamentally important question in evolutionary biology. On first principles, however, it is not
clear how such conflict can be resolved because males and females largely share the same
genome.
One way this conflict could be resolved is through sex-specific regulation of the genome
[71-74]. For example, growth-enhancing genes could be upregulated in the larger sex and
downregulated in the other. Sex hormones like testosterone and estrogen offer genetic
mechanisms for males and females to differentially regulate their genomes. Androgen and
estrogen circulate at different levels in males and females, and their main downstream targets are
Androgen receptor (Ar) and two Estrogen receptors (Erα and Erβ). Upon binding to their
hormonal ligands, these receptors undergo conformational changes, migrate into the cell nucleus,
bind to their respective DNA motifs, and orchestrate expression of nearby genes [75-80].
Therefore, androgen and estrogen receptors act as transcription factors that potentially regulate
thousands of genes in a sex-specific manner [65, 81-84], offering a potential molecular
mechanism to achieve sex-specific size [81, 82, 85-89].
If hormonal regulation represents a common genomic mechanism to resolve sexual
conflict over body size, then the variation in sexual size dimorphism across species should
correlate with the number of androgen / estrogen receptor elements (ARE’s and ERE’s) – the

12
DNA motifs that serve as binding sites – especially near protein coding regions. In a recent study
of 26 primate species, Anderson and Jones [90] found no support that sexual size dimorphism
correlated positively with the number of androgen receptor binding site. However, the relatively
recent accumulation of genomes and gene annotations from hundreds of mammalian species
means we can explore this question across a more heterogeneous collection of species and ask if
and how the relationship varies across a phylogeny.
Here we test whether sexual size dimorphism is correlated with the number of
ARE’s/ERE’s across 268 mammalian genomes. We uncovered a positive correlation between
sexual size dimorphism and the number of ARE’s – but only in relatively small-bodied lineages
(bats and rodents) and only in genomic regions near protein coding genes. In fact, ~90% of
protein coding genes show a positive effect between their number of ARE’s and sexual size
dimorphism. Smaller bodied species tend to reach sexual maturity in relatively short time
periods. One hypothesis to explain our findings is that hormonal signaling is more important for
short-lived species to achieve optimal body size in relatively shorter periods of time. Our study
uncovers previously unappreciated heterogeneity in the relationship between hormonal signaling
and sexual size dimorphism.
Materials & Methods
Our main goal was to test whether sexual size dimorphism covaried with the number of ARE’s
and ERE’s, as well as overall body size, using phylogenetically controlled linear modeling.
Therefore, we needed to collect body size data for sexually mature males and females, then
associate their sexual size dimorphism with the number of ARE’s and ERE’s across multiple
genomes. Because androgen receptor and estrogen receptors act as transcription factors, we also
focused protein coding genes.

13
2.1 Body Mass
We gathered body mass for sexually mature males and females from literature searching. We
only included body mass data taken outside of breeding season, primarily to avoid pregnant or
lactating females. Most body mass data were taken from Silva and Downing [91], supplemented
with additional literature sources (Supplementary File 1) [11, 91-98]. Sexual size dimorphism
(SSD) was calculated as the log(male/female) body size.
2.2 ARE and ERE counts across whole genomes
We counted ARE’s and ERE’s across all genome assemblies of the Zoonomia consortium, using
genome annotations from the Zoonomia Project [99-101]. Zoonomia sequenced and assembled
(or linked to already-assembled genomes stored in NCBI) 524 genomes from 464 unique species.
The canonical ARE is AGAACANNNTGTTCT, where the three N’s indicate a three
nucleotide spacer flanked by palindromic sequence [102-106]. 17 additional ARE’s have been
identified, all of which are highly similar in this basic sequence but vary in their specificity of
binding with AR. Following Anderson and Jones [90], we focused on these 18 ARE’s. A single
estrogen receptor element (ERE) has been identified as the sequence GGTCANNNTGACC [75,
78, 106-108]. We counted ARE’s and ERE’s using the VCOUNTPATTERN function from the
BIOSTRINGS package in R (v 2.68.1) [109]. For any non-palindromic ARE’s, we added up the
counts in the forward and reverse-complement directions of the genome.
For the 18 ARE’s, we also performed phylogenetically controlled Principal Components
Analysis using the PHYL.PCA function in the R package PHYTOOLS [110]. ARE counts across all
18 motifs were used as input, then we tested whether different Principal Components correlated
with sexual size dimorphism.

14
2.3 ARE and ERE counts near protein coding regions
Androgen receptor and the two estrogen receptors are transcription factors, so we might expect
ARE’s/ERE’s near protein coding regions to have a larger impact on gene regulation than those
far from genes.
We used the gene annotations of Kirilenko et al. [111], who developed A Tool to infer
Orthologs from Genome Alignments (TOGA) to essentially lift over annotations from a reference
genome (the house mouse genome annotation mm10) to all other genome assemblies. TOGA
makes six different gene calls: i) Intact genes, where the middle 80% of the mm10 reference gene
is identified without inactivating mutations (frameshifts or early termination) in the target
genome, ii) Partially intact genes, which are intact genes that show gaps in the genome
assembly, iii) Lost genes, where the middle 80% of the mm10 gene is present but contains at least
one inactivating mutation, iv) Missing genes, where less than 50% of the mm10 gene is present,
v) Uncertain lost genes, which are genes that would be classified as Lost but where evidence is
not strong, and vi) Paralogous projection genes, which are Intact genes that are not orthologous
to the reference mm10 gene, for example a lineage-specific duplication of a gene.
We recounted ARE’s and ERE’s that fell within 10 Kb, 100 Kb, or 1000 Kb of the
transcription start/stop of Intact and Paralogous Projection genes. We only counted unique
ARE’s and ERE’s, even if they fell near multiple protein coding genes. We included Paralogous
Projection genes because our hypothesis was focused on sex-specific expression of the genome,
regardless of whether a gene was duplicated, for example. The proximity over which ARE’s and
ERE’s can affect gene expression is a topic of some debate, but ARE’s have been shown to
influence genes more than 1000 Kb away [112]. We arrive at highly similar conclusions
regardless of which cutoff we use.

15
To better understand the contribution of individual genes to our study, we also tested
whether SSD correlated with ARE’s or ERE’s on a gene-by-gene basis. This analysis represents
an agnostic approach toward revealing genes or gene classes that explained variation in sexual
size dimorphism. As above, we counted ARE’s and ERE’s that fell within 10 Kb, 100 Kb, and
1000 Kb of the transcription start/stop of Intact and Paralogous Projection genes. In this genecentric approach only, ARE’s/ERE’s could be counted more than once if they fell near multiple
genes. Any genes that were annotated in fewer than 20 species were discarded, as were any genes
that showed no variance in counts.
2.4 Phylogenetically controlled linear modeling
We implemented phylogenetically controlled linear models using the PHYLOLM function in the R
package PHYLOLM [113] The basic model tested was SSD~ARE/ERE + body_size, where SSD
was sexual size dimorphism – log(male/female), ARE/ERE was the log-transformed count of
either androgen or estrogen response elements (whole genome, near protein-coding genes, or
per-gene), and body_size was the average male+female body size in a species. Average body size
was included as a covariate because of “Rensch’s Rule”, the observation that SSD in male-larger
species increases with body size [114]. The phylogeny we employed was that of Upham et al.
[115], trimmed to match our data (Supplemental File 2).
Following our gene-centric analyses, we tested whether genes that were correlated with
SSD (uncorrected p<0.05) were over-represented for Gene Ontology terms, using PantherDB
[116].

16
2.5 Genome quality
Some of our analyses could be confounded by variation in genome quality. For example,
Kirilenko et al. [111] annotated more genes from relatively high-quality genomes, but this is
probably a methodological artifact related to identifying more protein coding regions from wellassembled genomes. We tested whether ARE or ERE counts correlated with contig N50, a
common metric of genome quality, taken from Kirilenko et al. [111]. N50 is the contig length or
longer which includes half the bases of the assembly.
Results
From the 455 unique TOGA-annotated species from Zoonomia, 268 overlapped the
Upham phylogeny [115] and had body size data in the literature. Phylogenetically controlled
linear models from various subsets of our data are presented in Supplemental File 3, but these
268 species are the focus of the remainder of the manuscript. Among these 268 species, SSD
ranged from -0.99 (dugongs) to 1.55 (southern elephant seals). The raw number of ARE’s ranged
from 4,870 to 61,776 (mean=9,998) across whole genomes; 471- 29,294 (mean=3,283) within 10
Kb of genes; 501-43,387 (mean=5,450) within 100 Kb of genes; and 501- 44,382 (mean=7,511)
within 1000 Kb of genes. The raw number of ERE’s ranged from 160-1,524 (mean=344.1)
across whole genomes; 11-656 (mean=108.7) within 10 Kb of genes; 12-1,125 (mean=191.0)
within 100 Kb of genes; and 12-1,464 (mean=264.4) within 1000 Kb of genes.
2.6 Across whole genomes and including all species, SSD correlated with body size, but not
ARE or ERE counts
SSD was positively correlated with overall body size (p = 0.001) but not ARE counts across
whole genomes (p = 0.568) (Table 1). The full model, SSD = 0.044 * log(ARE_count) + 0.03 *

17
log(body_size), accounted for 4.4% of the variance in SSD (Table 1). The positive correlation
between SSD and body size (and lack of correlation between SSD and ARE) held if we analyzed
the 18 different ARE motifs separately, or their Principal Components (Supplemental File 3). In
the case of ERE’s, the full model was SSD = -0.06 * log(ERE_count) + 0.031 * log(body_size),
with body size explaining a significant amount of variance (p = 0.001) but not ERE counts (p =
0.552) (Table 1).
The relationship between SSD, ARE/ERE counts, and body size differed across the
phylogeny. Therefore, we repeated the analysis for orders that had at least 20 species in them.
For the relatively large-bodied orders Cetartiodactyla (83 species), Primates (42 species), and
Carnivora (48 species), we largely observe the same trend as above, where SSD correlated with
body size (p = 0.006, 0.047, and 0.065, respectively) but not total ARE count (p = 0.651, 0.636,
and 0.607, respectively) (Table 1). In contrast, in the relatively small-bodied orders Chiroptera
(21 species) and Rodentia (41 species), SSD was not correlated with body size (p = 0.904 and
0.772, respectively) (Table 1), and rodents trended towards a correlation between SSD and
number of ARE’s (p=0.055).
There were no correlations between SSD and ERE’s when these five orders were
analyzed separately (Table 1).
2.7 Near genes and in small-bodied species, SSD correlated with ARE counts but not overall
body size
The difference between small-bodied and large-bodied organisms became even more pronounced
after we confined our analyses to ARE’s and ERE’s near protein coding genes. The three
different cutoffs (10 Kb, 100 Kb, and 1000 Kb) yielded qualitatively identical results
(Supplemental File 4); we present the 1000 Kb for simplicity. For ARE’s within 1000 Kb of

18
protein coding genes, SSD correlated with both body size (p = 0.001) and ARE count (p =
0.039), with the full model, SSD = 0.07 * log(ARE_count) + 0.029 * log(body_size) accounting
for 5.8% of the variance (Table 1).
These relationships varied across orders (Figure 1). In the relatively large-bodied orders
Carnivora, Cetartiodactyla, and Primates, SSD was significantly positively correlated with
overall body size (p = 0.015, 0.005, and 0.049, respectively) but not ARE counts (p = 0.172,
0.832, and 0.802, respectively). In the relatively small-bodied orders Chiroptera and Rodentia, on
the other hand, SSD was significantly positively correlated with ARE counts (p = 0.006 and
0.017, respectively) but not overall body size (p = 0.72 and 0.874, respectively). These trends
held for the 100 Kb and 10 Kb cutoffs, although Rodentia slightly lost significance between SSD
and ARE counts (p = 0.063 and 0.059, respectively) and Primates slightly lost significance
between SSD and overall body size (p = 0.052 and 0.054, respectively) (Supplemental File 3).
All these results should be treated with caution as the number of species analyzed is small and
unequal once we break the analyses into orders. In addition, we did not correct for multiple
testing. Nevertheless, the general trends held even with the very stringent cutoff of 10 Kb.
For ERE’s within 1000 Kb of genes, SSD correlated with body_size (Table 1). There
were almost no significant correlations between SSD and ERE counts near genes, the lone
exception being in Chiroptera (Table 1).
2.8 Gene-centric analyses revealed persistent, positive influence of ARE on SSD
To tease apart the contributions of specific genes to the overall pattern just described, we tested
each gene separately for a correlation between its ARE’s/ERE’s and SSD, again using average
body size as a covariate. At the 1000 Kb cutoff, 20,884 genes were annotated from at least 20

19
species and had at least some variance in total ARE counts. Interestingly, 17,694 (84.7%) yielded
a positive effect of ARE count on SSD, regardless of statistical significance (Figure 2). Of the
2,135 genes that showed a significant correlation between SSD and ARE counts (uncorrected
p<0.05) this skew became even more apparent - 2,118 (99.2%) showed a positive effect of ARE
on SSD. These trends held for the 100 Kb and 10 Kb cutoffs.
Among the 2,118 significant and positive-effect genes, several Gene Ontology terms
related to detection of pheromones were significantly overrepresented (Supplemental File 5).
Interestingly, there were 60 vomeronasal receptor genes among these 2,118 genes. One
representative gene, Vmn1r209, is shown in Figure 3. Like Vmn1r209, the other 59 vomeronasal
receptors genes in this group largely share several features: 1) significant Vmnr’s are distributed
throughout the genome, 2) the correlations are not driven by a few outliers, 3) the correlations
are not driven by discretization of ARE counts (in other words, species show largely continuous
variation across the x-axis), and 4) the correlations are not driven by just a few taxonomic groups
(that is, orders are spread out along the x-axis). In short, the enrichment of protein coding genes
is likely to be a true positive.
ERE counts showed a much different result. At the 1000 Kb cutoff, 20,869 were
annotated from at least 20 species and had at least some variance in ERE counts. Of these,
11,271 (54%) showed a positive effect on SSD. Of the 1,287 genes with a significant correlation
between SSD and ERE counts (uncorrected p<0.05), 850 (66%) showed a positive effect on
SSD. Thus, although there is still skew towards positive influence on SSD, it is much less
dramatic than with ARE’s, and no Gene Ontology terms were enriched (Supplemental File 5).

20
2.9 Genome quality
ARE and ERE counts did not correlate with contig N50 (p=0.5, 0.6, respectively). Therefore,
differences in genome quality are unlikely to influence our overall conclusions.
Discussion
The widespread phenomenon of sexual size dimorphism implies that the sexes have different
fitness optima in many species, and that sexual conflict has been at least partially resolved. One
mechanism of conflict resolution is through sex-specific regulation of the genome. Here we
tested a prediction of this hypothesis, that SSD is evolutionarily correlated to the number ARE’s
and ERE’s in a genome. Three important patterns emerge from our study. First, ARE’s are more
strongly correlated with SSD than ERE’s, which show almost no correlation. Second, the largebodied and small-bodied lineages show strikingly different patterns (Figure 1). In large-bodied
lineages (Carnivora, Cetartiodactyla, and Primates), SSD correlated with body size but not ARE
counts, while in small-bodied lineages (Chiroptera and Rodentia), the opposite pattern was
observed. Third, the strongest correlations between SSD and ARE’s are seen near protein coding
genes (Table 1). On a gene-by-gene basis, ARE’s persistently and positively correlated with SSD
(Figure 2), and those that were significantly correlated with SSD were enriched for processes
related to pheromone detection. Taken together our study suggests that androgen signaling is an
important mechanism enabling males and females to approach different body size optima,
especially in small-bodied species, and especially near genes. These insights were only possible
by expanding previous work to a larger set of species and exploiting their genome annotations.
We lack fundamental data required to directly link androgen signaling to sexual size
dimorphism, and the proxy we use (the number of ARE’s) is likely to be a noisy measurement.

21
The presence of an ARE does not necessarily mean it serves as a binding site for AR. The 18
different ARE’s have variable specificity to binding androgen receptor and can be bound by other
transcription factors [79, 102, 105, 117-120]. Chromatin structure should also influence the
accessibility of ARE’s to be bound by AR. Even if we knew which ARE sites were bound by AR,
we would not know which genes are affected and how their expression is impacted in both
spatially and temporally. ARE’s can act cooperatively to induce gene expression in non-additive
ways [80], further obscuring the connection between ARE counts and transcription. Lastly, even
if we knew all the downstream target genes of AR+ARE’s, we would also need to know each
gene’s contribution to male versus female body size. This relationship can be complicated – for
example, in two closely related species of lizards, testosterone had opposite effects on male body
size [121, 122]. In short, there are many reasons to believe that we lack power to detect a
correlation between SSD and ARE counts. The correlations we find between SSD and ARE
counts, especially near genes in small-bodied lineages, could actually be much stronger than
reported here. In addition, it should be noted that Anderson and Jones [90] showed that, in
humans, ARE’s congregated near male-biased genes and androgen-responsive genes, supporting
the notion that this proxy is linked to meaningful genomic regulation.
What could explain the difference we observe in large- versus small-bodied lineages?
Two consistently strong correlates of body size are lifespan and metabolic rate, with smallbodied species generally living relatively short lives under relatively high metabolic rates [70,
123-131]. Sexual size dimorphism could arise in different ways according to lifespan. For largebodied lineages, whichever sex is larger can accumulate body size over multiple growing
seasons. In fact, the larger sex can simply take longer to develop than the smaller sex, a

22
phenomenon known as “bimaturism” [74, 132-136]. For long-lived species, bimaturism could be
important if it can extend over multiple growing seasons.
On the other hand, small-bodied lineages generally have less absolute time to accumulate
body size – for example, many rodent species live less than a year in nature. Perhaps short-lived
species are more reliant on hormonal signaling to differentially regulate growth genes over their
relatively short lives. This hypothesis is inconsistent with some data – for example, the bat
Myotis brandtii can live 41 years even though it weighs no more than 6-8 grams [137-140]. On
the other hand, lifespan per se is probably not as important as the age of sexual maturity and
peak reproductive performance, data which are currently lacking.
Another hypothesis to explain the difference between large and small lineages is related
to the magnitude and direction of sexual size dimorphism. The three large-bodied lineages all
have relatively large SSD, ranging from an average of 0.195 (Cetartiodactyla) to 0.375
(Carnivora). In contrast, the small-bodied lineages have fairly modest SSD of 0.073 (Rodentia)
and in fact shows an average female-larger SSD of -0.058 in Chiroptera, a pattern described in
other studies [141, 142]. It is possible that the effect of androgen signaling on SSD is only
observed when the magnitude of SSD is small. Said another way, many more factors than ARE
counts may be required to achieve the more dramatic forms of SSD often observed in largebodied species.
Unexpectedly, we find that most genes show a positive effect between their number of
ARE’s and SSD (Fig. 2). This pattern suggests that most genes contribute to body size in some
way. Alternatively, or additionally, many genes that are regulated by androgen may be indirectly
connected to size dimorphism. For example, a species with large magnitude SSD may require

23
sex-specific regulation of genes involved in metabolism or resource acquisition related to
supporting different body sizes.
Among genes whose ARE counts significantly and positively correlated with SSD, there
was an enrichment of genes involved in pheromone detection. We are unaware of any direct links
between vomeronasal receptor variation and body size. Perhaps the connection between ARE’s at
vomeronasal receptor genes and SSD is indirect. Large males in male-larger species are likely to
perform better in dominance contests [44, 47, 143]. In such species, perhaps there is correlated
selection for males to detect pheromones of reproductive females, or for females to detect signals
of dominant males. While completely speculative, such hypotheses could reconcile the puzzling
association between ARE’s, SSD, and detection of chemical signaling.
How organisms resolve sexual conflict over body size is an important question in
evolutionary and molecular biology. The existence of sexual size dimorphism implies that sexual
conflict over body size has been at least partially resolved, although it is unlikely to be fully
resolved [71, 144]. Full resolution requires flexibility to accumulate mutations with sex-specific
effects. While some genetic variants have been identified that differentially influence male vs.
female body size [145], others influence both sexes [146]. This latter pattern likely poses
constraints on what is possible, leading Pennell et al. [71] to coin the phrase “you can’t always
get what you want” with respect to sexual conflict over body size. Multiple empirical studies
have shown that sexually antagonistic alleles are common in most studied populations [147-150],
implying that the conflict remains unresolved. Nevertheless, our study implicates androgen
signaling as a mediator of sex-specific body size, at least in some lineages. Future mapping of
the molecular biology of androgen signaling and body size promises to increase our
understanding of this ubiquitous pattern.

24

25
Figure legends
Figure 1.1
SSD is significantly positively correlated with total ARE counts in Chiroptera and Rodentia but
not Carnivora, Cetartiodactyla, Primates and not when all species are analyzed. Each point on
the plot is a species based on regions within 1000 Kb of protein coding genes.
Figure 1.2
The -log10 x p-value and estimated effect size of ARE’s (red) and ERE’s (blue) per gene. For
ARE’s, a large majority of genes, regardless of statistical significance, has a positive influence on
SSD. This pattern contrasts with ERE’s.
Figure 1.3
Representative plot of 1 of 60 vomeronasal receptor genes that are significantly positively
correlated with SSD. Letters indicate orders: Primates, CetartiodactYla, Rodentia, or Other.
For supplementary files please contact the author: ghione@usc.edu
Supplemental File 1.1 All species, ARE’s, and ERE’s analyzed in the manuscript.
Supplemental File 1.2 The phylogenetic tree relating the species analyzed in the manuscript.
Supplemental File 1.3 All phylogenetically controlled linear models performed, across all
subsets of data.
Supplemental File 1.4 Repeat of Table 1, but with the different cutoffs.
Supplemental File 1.5 Results of Gene Ontology analyses.
Supplemental File 1.6 Vomeronasal receptor genes across the whole mouse genome (black
dashes), and the 60 whose ARE’s are significantly positively correlated with SSD (red dashes).

26
Figure 1.1

27
Figure 1.2

28
Figure 1.3

29
Table 1.1

30
Supplementary Figure 1.1
Double click on the Excel spreadsheet and scroll through table
Assembly_Names UPHAM_NOTATION Species MEAN_MALE_BODY_WEIGHT_KG MEAN_FEMALE_BODY_WEIGHT_KG hklk mvdp p21 c3
HLtacAcu2TACHYGLOSSUS_ACULEATUS_TACHYGLOSSIDAE_MONOTREMATA TACHYGLOSSUS_ACULEATUS 5.5 4.35 289 583 32 62
HLornAna4ORNITHORHYNCHUS_ANATINUS_ORNITHORHYNCHIDAE_MONOTREMATA ORNITHORHYNCHUS_ANATINUS 2.12 1.44 326 544 35 42
HLscaAqu1SCALOPUS_AQUATICUS_TALPIDAE_EULIPOTYPHLA SCALOPUS_AQUATICUS 0.1034 0.0792 346 787 71 32
conCri1 CONDYLURA_CRISTATA_TALPIDAE_EULIPOTYPHLA CONDYLURA_CRISTATA 0.042 0.041 294 752 42 30
eriEur2 ERINACEUS_EUROPAEUS_ERINACEIDAE_EULIPOTYPHLA ERINACEUS_EUROPAEUS 1.64 1.16 414 1045 33 31
HLsunEtr1 SUNCUS_ETRUSCUS_SORICIDAE_EULIPOTYPHLA SUNCUS_ETRUSCUS 0.00185 0.0021 761 893 38 31
sorAra2 SOREX_ARANEUS_SORICIDAE_EULIPOTYPHLA SOREX_ARANEUS0.0098 0.00996 679 773 97 57
HLchrBra1 CHRYSOCYON_BRACHYURUS_CANIDAE_CARNIVORA CHRYSOCYON_BRACHYURUS 23 20 389 831 57 44
HLlycPic3 LYCAON_PICTUS_CANIDAE_CARNIVORA LYCAON_PICTUS 25 25 371 826 55 46
canFam5 CANIS_LUPUS_CANIDAE_CARNIVORA CANIS_LUPUS_FAMILIARIS 45 41 381 853 56 47
HLvulVul2 VULPES_VULPES_CANIDAE_CARNIVORA VULPES_VULPES 6.7 5.4 407 857 68 42
HLvulLag1 VULPES_LAGOPUS_CANIDAE_CARNIVORA VULPES_LAGOPUS3.8 3.1 399 1253 75 44
HLlepWed2LEPTONYCHOTES_WEDDELLII_PHOCIDAE_CARNIVORA LEPTONYCHOTES_WEDDELLII 322 348 401 911 58 46
HLmirLeo1 MIROUNGA_LEONINA_PHOCIDAE_CARNIVORA MIROUNGA_LEONINA 4000 850 377 894 60 46
HLmirAng2 MIROUNGA_ANGUSTIROSTRIS_PHOCIDAE_CARNIVORA MIROUNGA_ANGUSTIROSTRIS 2300 900 388 883 59 46
HLphoVit1 PHOCA_VITULINA_PHOCIDAE_CARNIVORA PHOCA_VITULINA154 148 382 875 69 56
HLhalGry1 HALICHOERUS_GRYPUS_PHOCIDAE_CARNIVORA HALICHOERUS_GRYPUS 250.4 187.21 380 868 73 57
HLodoRos1ODOBENUS_ROSMARUS_ODOBENIDAE_CARNIVORA ODOBENUS_ROSMARUS 1270 850 377 827 51 51
HLzalCal1 ZALOPHUS_CALIFORNIANUS_OTARIIDAE_CARNIVORA ZALOPHUS_CALIFORNIANUS 390 110 383 848 53 52
HLeumJub1EUMETOPIAS_JUBATUS_OTARIIDAE_CARNIVORA EUMETOPIAS_JUBATUS 1000 273 383 848 56 57
HLarcGaz2 ARCTOCEPHALUS_GAZELLA_OTARIIDAE_CARNIVORA ARCTOCEPHALUS_GAZELLA 200 50 382 853 46 52
HLspiGra1 SPILOGALE_GRACILIS_MEPHITIDAE_CARNIVORA SPILOGALE_GRACILIS 0.734 0.3545 399 855 61 62
HLpotFla1 POTOS_FLAVUS_PROCYONIDAE_CARNIVORA POTOS_FLAVUS 3.53 3.27 396 824 93 67
HLproLot1 PROCYON_LOTOR_PROCYONIDAE_CARNIVORA PROCYON_LOTOR8.51 6.51 419 877 86 86
HLbasAst1 BASSARISCUS_ASTUTUS_PROCYONIDAE_CARNIVORA BASSARISCUS_ASTUTUS 1.32 0.94 422 915 92 78
HLgulGul1 GULO_GULO_MUSTELIDAE_CARNIVORA GULO_GULO 14.2 9.41 385 876 67 62
HLmarZib1 MARTES_ZIBELLINA_MUSTELIDAE_CARNIVORA MARTES_ZIBELLINA1.5 1.125 380 886 80 58
HLmusPut1MUSTELA_PUTORIUS_MUSTELIDAE_CARNIVORA MUSTELA_PUTORIUS1.11 0.69 397 908 70 53
HLpteBra2 PTERONURA_BRASILIENSIS_MUSTELIDAE_CARNIVORA PTERONURA_BRASILIENSIS 30 24 422 933 63 48
enhLutKen1ENHYDRA_LUTRIS_MUSTELIDAE_CARNIVORA ENHYDRA_LUTRIS_KENYONI 32.8 21.5 418 911 81 46
HLlutLut1 LUTRA_LUTRA_MUSTELIDAE_CARNIVORA LUTRA_LUTRA 9.6 6.75 406 877 67 50
HLaonCin1AONYX_CINEREA_MUSTELIDAE_CARNIVORA AONYX_CINEREUS3.47 3.2 424 886 71 55
HLmelCap1MELLIVORA_CAPENSIS_MUSTELIDAE_CARNIVORA MELLIVORA_CAPENSIS 9.7 6.2 401 951 81 56
HLailFul2 AILURUS_FULGENS_AILURIDAE_CARNIVORA AILURUS_FULGENS5.4 5.2 436 886 66 59
HLursAme2URSUS_AMERICANUS_URSIDAE_CARNIVORA URSUS_AMERICANUS86 54.05 417 979 78 71
ursMar1 URSUS_MARITIMUS_URSIDAE_CARNIVORA URSUS_MARITIMUS360 187 382 909 74 72
HLursArc1 URSUS_ARCTOS_URSIDAE_CARNIVORA URSUS_ARCTOS_HORRIBILIS 200 111.9 404 921 76 69
HLailMel2 AILUROPODA_MELANOLEUCA_URSIDAE_CARNIVORA AILUROPODA_MELANOLEUCA 117 97 409 919 83 60

31
Supplementary Figure 1.2
((TACHYGLOSSUS_ACULEATUS_TACHYGLOSSIDAE_MONOTREMATA:48.0786375,O
RNITHORHYNCHUS_ANATINUS_ORNITHORHYNCHIDAE_MONOTREMATA:48.07863
75):136.2110822,((((((SCALOPUS_AQUATICUS_TALPIDAE_EULIPOTYPHLA:27.2860280
4,CONDYLURA_CRISTATA_TALPIDAE_EULIPOTYPHLA:27.28602805):40.8665164,(ERI
NACEUS_EUROPAEUS_ERINACEIDAE_EULIPOTYPHLA:65.63081099,(SUNCUS_ETRU
SCUS_SORICIDAE_EULIPOTYPHLA:42.84229489,SOREX_ARANEUS_SORICIDAE_EUL
IPOTYPHLA:42.84229489):22.78851611):2.52173345):1.654445039,(((((((CHRYSOCYON_B
RACHYURUS_CANIDAE_CARNIVORA:3.672647349,LYCAON_PICTUS_CANIDAE_CAR
NIVORA:3.672647349):1.30617229,CANIS_LUPUS_CANIDAE_CARNIVORA:4.978819639
):4.219171201,(VULPES_VULPES_CANIDAE_CARNIVORA:2.984010697,VULPES_LAGO
PUS_CANIDAE_CARNIVORA:2.984010697):6.213980144):25.33656525,(((((LEPTONYCH
OTES_WEDDELLII_PHOCIDAE_CARNIVORA:5.211750031,(MIROUNGA_LEONINA_PH
OCIDAE_CARNIVORA:1.216689665,MIROUNGA_ANGUSTIROSTRIS_PHOCIDAE_CAR
NIVORA:1.216689665):3.995060366):3.675572144,(PHOCA_VITULINA_PHOCIDAE_CAR
NIVORA:1.643826777,HALICHOERUS_GRYPUS_PHOCIDAE_CARNIVORA:1.643826777)
:7.243495397):5.128284907,(ODOBENUS_ROSMARUS_ODOBENIDAE_CARNIVORA:9.31
4390541,((ZALOPHUS_CALIFORNIANUS_OTARIIDAE_CARNIVORA:2.780615032,EUM
ETOPIAS_JUBATUS_OTARIIDAE_CARNIVORA:2.780615032):0.4360668693,ARCTOCEP
HALUS_GAZELLA_OTARIIDAE_CARNIVORA:3.216681901):6.09770864):4.701216541):1
4.50333561,(SPILOGALE_GRACILIS_MEPHITIDAE_CARNIVORA:25.46638851,(((POTOS
_FLAVUS_PROCYONIDAE_CARNIVORA:18.29461097,(PROCYON_LOTOR_PROCYONI
DAE_CARNIVORA:7.404806405,BASSARISCUS_ASTUTUS_PROCYONIDAE_CARNIVO
RA:7.404806406):10.88980457):4.344342792,(((GULO_GULO_MUSTELIDAE_CARNIVOR
A:6.116994626,MARTES_ZIBELLINA_MUSTELIDAE_CARNIVORA:6.116994626):5.33991
5752,(MUSTELA_PUTORIUS_MUSTELIDAE_CARNIVORA:10.73441616,(PTERONURA_
BRASILIENSIS_MUSTELIDAE_CARNIVORA:8.586733338,(ENHYDRA_LUTRIS_MUSTE
LIDAE_CARNIVORA:6.264886763,(LUTRA_LUTRA_MUSTELIDAE_CARNIVORA:4.161
91016,AONYX_CINEREA_MUSTELIDAE_CARNIVORA:4.16191016):2.102976603):2.3218
46575):2.147682822):0.7224942177):0.7106599007,MELLIVORA_CAPENSIS_MUSTELIDA
E_CARNIVORA:12.16757028):10.47138349):1.966256427,AILURUS_FULGENS_AILURID
AE_CARNIVORA:24.60521019):0.8611783102):3.052554187):0.8297055414,((URSUS_AME
RICANUS_URSIDAE_CARNIVORA:2.98174859,(URSUS_MARITIMUS_URSIDAE_CARN
IVORA:1.015542626,URSUS_ARCTOS_URSIDAE_CARNIVORA:1.015542626):1.96620596
4):7.310993415,AILUROPODA_MELANOLEUCA_URSIDAE_CARNIVORA:10.29274201):
19.05590623):5.185907862):4.045921526,(((CROCUTA_CROCUTA_HYAENIDAE_CARNIV
ORA:3.790696793,HYAENA_HYAENA_HYAENIDAE_CARNIVORA:3.790696793):14.3406
1031,(CRYPTOPROCTA_FEROX_EUPLERIDAE_CARNIVORA:14.86781076,(SURICATA_
SURICATTA_HERPESTIDAE_CARNIVORA:7.161673142,(MUNGOS_MUNGO_HERPESTI
DAE_CARNIVORA:6.664924206,HELOGALE_PARVULA_HERPESTIDAE_CARNIVORA:
6.664924206):0.496748936):7.706137624):3.263496338):4.775393927,((PANTHERA_TIGRIS
_FELIDAE_CARNIVORA:3.476828061,(((PANTHERA_LEO_FELIDAE_CARNIVORA:1.62
8092416,PANTHERA_PARDUS_FELIDAE_CARNIVORA:1.628092416):0.4926073741,PAN
THERA_UNCIA_FELIDAE_CARNIVORA:2.12069979):0.3656546811,PANTHERA_ONCA_
FELIDAE_CARNIVORA:2.486354471):0.9904735902):3.073753729,((FELIS_CATUS_FELID

32
AE_CARNIVORA:4.94730293,PRIONAILURUS_BENGALENSIS_FELIDAE_CARNIVORA
:4.94730293):0.5084889709,((LYNX_PARDINUS_FELIDAE_CARNIVORA:1.343913275,LY
NX_CANADENSIS_FELIDAE_CARNIVORA:1.343913275):3.431094678,(ACINONYX_JUB
ATUS_FELIDAE_CARNIVORA:3.188407831,PUMA_CONCOLOR_FELIDAE_CARNIVOR
A:3.188407831):1.586600122):0.6807839488):1.094789889):16.35611924):15.67377659):26.11
909243,((EQUUS_ZEBRA_EQUIDAE_PERISSODACTYLA:37.7221973,(CERATOTHERIU
M_SIMUM_RHINOCEROTIDAE_PERISSODACTYLA:6.614749088,DICEROS_BICORNIS
_RHINOCEROTIDAE_PERISSODACTYLA:6.614749088):31.10744822):24.27769263,((VIC
UGNA_VICUGNA_CAMELIDAE_CETARTIODACTYLA:2.076781803,LAMA_GUANICOE
_CAMELIDAE_CETARTIODACTYLA:2.076781803):57.6563594,((CATAGONUS_WAGNE
RI_TAYASSUIDAE_CETARTIODACTYLA:26.24237762,BABYROUSA_BABYRUSSA_SUI
DAE_CETARTIODACTYLA:26.24237762):30.18571862,((HIPPOPOTAMUS_AMPHIBIUS_
HIPPOPOTAMIDAE_CETARTIODACTYLA:38.89291665,((KOGIA_BREVICEPS_PHYSET
ERIDAE_CETARTIODACTYLA:16.07299762,((INIA_GEOFFRENSIS_INIIDAE_CETARTIO
DACTYLA:11.80117609,(((PHOCOENA_PHOCOENA_PHOCOENIDAE_CETARTIODACT
YLA:4.209557657,PHOCOENA_SINUS_PHOCOENIDAE_CETARTIODACTYLA:4.2095576
57):4.369660942,(MONODON_MONOCEROS_MONODONTIDAE_CETARTIODACTYLA:
3.650888315,DELPHINAPTERUS_LEUCAS_MONODONTIDAE_CETARTIODACTYLA:3.6
50888315):4.928330284):0.8936330968,((((LAGENORHYNCHUS_OBLIQUIDENS_DELPHI
NIDAE_CETARTIODACTYLA:1.492640075,CEPHALORHYNCHUS_COMMERSONII_DE
LPHINIDAE_CETARTIODACTYLA:1.492640075):1.761456936,(STENO_BREDANENSIS_
DELPHINIDAE_CETARTIODACTYLA:2.962812676,(TURSIOPS_TRUNCATUS_DELPHIN
IDAE_CETARTIODACTYLA:1.992121799,TURSIOPS_ADUNCUS_DELPHINIDAE_CETA
RTIODACTYLA:1.992121799):0.9706908777):0.291284335):0.1433581353,(GRAMPUS_GRI
SEUS_DELPHINIDAE_CETARTIODACTYLA:2.759884854,(GLOBICEPHALA_MELAS_D
ELPHINIDAE_CETARTIODACTYLA:1.151120046,PEPONOCEPHALA_ELECTRA_DELPH
INIDAE_CETARTIODACTYLA:1.151120046):1.608764808):0.637570293):0.9562416286,OR
CINUS_ORCA_DELPHINIDAE_CETARTIODACTYLA:4.353696775):5.11915492):2.328324
39):1.375649992,(MESOPLODON_EUROPAEUS_ZIPHIIDAE_CETARTIODACTYLA:8.671
930775,ZIPHIUS_CAVIROSTRIS_ZIPHIIDAE_CETARTIODACTYLA:8.671930775):4.50489
5303):2.896171545):0.8240690962,((BALAENOPTERA_ACUTOROSTRATA_BALAENOPT
ERIDAE_CETARTIODACTYLA:3.645278109,(MEGAPTERA_NOVAEANGLIAE_BALAEN
OPTERIDAE_CETARTIODACTYLA:2.897280621,ESCHRICHTIUS_ROBUSTUS_ESCHRI
CHTIIDAE_CETARTIODACTYLA:2.897280621):0.7479974884):6.581663002,(EUBALAEN
A_JAPONICA_BALAENIDAE_CETARTIODACTYLA:5.233017063,((BALAENOPTERA_P
HYSALUS_BALAENOPTERIDAE_CETARTIODACTYLA:3.776696046,(BALAENOPTERA
_MUSCULUS_BALAENOPTERIDAE_CETARTIODACTYLA:3.065548431,BALAENOPTE
RA_EDENI_BALAENOPTERIDAE_CETARTIODACTYLA:3.065548431):0.7111476148):0.7
832718542,BALAENOPTERA_BONAERENSIS_BALAENOPTERIDAE_CETARTIODACTY
LA:4.559967901):0.6730491624):4.993924049):6.670125608):21.99584993):6.698601499,(TR
AGULUS_JAVANICUS_TRAGULIDAE_CETARTIODACTYLA:32.48562495,(((OKAPIA_J
OHNSTONI_GIRAFFIDAE_CETARTIODACTYLA:8.48859835,GIRAFFA_CAMELOPARD
ALIS_GIRAFFIDAE_CETARTIODACTYLA:8.48859835):9.203145059,ANTILOCAPRA_A
MERICANA_ANTILOCAPRIDAE_CETARTIODACTYLA:17.69174341):0.3461326384,((((R
ANGIFER_TARANDUS_CERVIDAE_CETARTIODACTYLA:7.193694272,(ODOCOILEUS_
VIRGINIANUS_CERVIDAE_CETARTIODACTYLA:1.214598919,ODOCOILEUS_HEMION

33
US_CERVIDAE_CETARTIODACTYLA:1.214598919):5.979095352):2.464200093,(ALCES_
ALCES_CERVIDAE_CETARTIODACTYLA:8.92581323,(HYDROPOTES_INERMIS_CERVI
DAE_CETARTIODACTYLA:4.953253571,CAPREOLUS_PYGARGUS_CERVIDAE_CETAR
TIODACTYLA:4.953253571):3.972559659):0.7320811338):0.5183453195,((MUNTIACUS_R
EEVESI_CERVIDAE_CETARTIODACTYLA:2.094982925,(MUNTIACUS_MUNTJAK_CER
VIDAE_CETARTIODACTYLA:1.711908567,MUNTIACUS_CRINIFRONS_CERVIDAE_CE
TARTIODACTYLA:1.711908567):0.3830743579):5.612244095,((ELAPHURUS_DAVIDIANU
S_CERVIDAE_CETARTIODACTYLA:3.13316532,CERVUS_ELAPHUS_CERVIDAE_CETA
RTIODACTYLA:3.13316532):3.028250317,AXIS_PORCINUS_CERVIDAE_CETARTIODAC
TYLA:6.161415637):1.545811382):2.469012664):6.682789881,(MOSCHUS_MOSCHIFERUS
_MOSCHIDAE_CETARTIODACTYLA:15.87389142,(((TRAGELAPHUS_IMBERBIS_BOVI
DAE_CETARTIODACTYLA:7.638980196,(TRAGELAPHUS_SCRIPTUS_BOVIDAE_CETA
RTIODACTYLA:6.255275284,TRAGELAPHUS_STREPSICEROS_BOVIDAE_CETARTIOD
ACTYLA:6.255275284):1.383704912):3.885758855,(SYNCERUS_CAFFER_BOVIDAE_CET
ARTIODACTYLA:9.124847622,(((BOS_GRUNNIENS_BOVIDAE_CETARTIODACTYLA:1.
024955171,BISON_BISON_BOVIDAE_CETARTIODACTYLA:1.024955171):1.396174272,(B
OS_TAURUS_BOVIDAE_CETARTIODACTYLA:1.462451983,BOS_FRONTALIS_BOVIDA
E_CETARTIODACTYLA:1.462451983):0.9586774597):1.457175631,BOS_GAURUS_BOVID
AE_CETARTIODACTYLA:3.878305075):5.246542547):2.39989143):2.217253104,(((((RAPHI
CERUS_CAMPESTRIS_BOVIDAE_CETARTIODACTYLA:8.952966363,MADOQUA_KIRK
II_BOVIDAE_CETARTIODACTYLA:8.952966363):0.7176637132,(SAIGA_TATARICA_BO
VIDAE_CETARTIODACTYLA:8.064439274,(ANTIDORCAS_MARSUPIALIS_BOVIDAE_
CETARTIODACTYLA:6.088484777,LITOCRANIUS_WALLERI_BOVIDAE_CETARTIODA
CTYLA:6.088484777):1.975954497):1.606190802):2.140660659,(((NEOTRAGUS_PYGMAE
US_BOVIDAE_CETARTIODACTYLA:10.12761972,SYLVICAPRA_GRIMMIA_BOVIDAE_
CETARTIODACTYLA:10.12761972):0.1425136992,OREOTRAGUS_OREOTRAGUS_BOVI
DAE_CETARTIODACTYLA:10.27013342):0.6857275707,(((HIPPOTRAGUS_NIGER_BOVI
DAE_CETARTIODACTYLA:4.817540361,HIPPOTRAGUS_EQUINUS_BOVIDAE_CETART
IODACTYLA:4.817540361):2.187866441,ORYX_GAZELLA_BOVIDAE_CETARTIODACT
YLA:7.005406802):2.915342467,((CONNOCHAETES_TAURINUS_BOVIDAE_CETARTIOD
ACTYLA:5.66535062,DAMALISCUS_LUNATUS_BOVIDAE_CETARTIODACTYLA:5.6653
50621):4.064703729,(PANTHOLOPS_HODGSONII_BOVIDAE_CETARTIODACTYLA:8.84
0622944,((((OVIS_NIVICOLA_BOVIDAE_CETARTIODACTYLA:0.923598201,OVIS_CAN
ADENSIS_BOVIDAE_CETARTIODACTYLA:0.923598201):1.339469932,(OVIS_ARIES_BO
VIDAE_CETARTIODACTYLA:1.741888876,OVIS_AMMON_BOVIDAE_CETARTIODACT
YLA:1.741888876):0.5211792569):4.47380876,(CAPRA_SIBIRICA_BOVIDAE_CETARTIOD
ACTYLA:4.179617769,(CAPRA_IBEX_BOVIDAE_CETARTIODACTYLA:2.535727895,CA
PRA_AEGAGRUS_BOVIDAE_CETARTIODACTYLA:2.535727895):1.643889874):2.557259
124):0.3956708203,(AMMOTRAGUS_LERVIA_BOVIDAE_CETARTIODACTYLA:6.625320
552,OREAMNOS_AMERICANUS_BOVIDAE_CETARTIODACTYLA:6.625320552):0.50722
71603):1.708075231):0.8894314058):0.1906949196):1.035111725):0.8554297405):0.13558568
36,((KOBUS_LECHE_BOVIDAE_CETARTIODACTYLA:4.59777652,KOBUS_ELLIPSIPRY
MNUS_BOVIDAE_CETARTIODACTYLA:4.59777652):4.139531267,REDUNCA_REDUNC
A_BOVIDAE_CETARTIODACTYLA:8.737307787):3.209568632):0.5656480495,AEPYCER
OS_MELAMPUS_BOVIDAE_CETARTIODACTYLA:12.51252447):1.229467687):2.1318992
73):0.9851381364):1.178846482):14.44774891):13.1058932):10.83657808):3.305044978):2.26

34
6748734):2.699680101):1.218068824,(((CRASEONYCTERIS_THONGLONGYAI_CRASEON
YCTERIDAE_CHIROPTERA:41.45291744,(RHINOLOPHUS_SINICUS_RHINOLOPHIDAE
_CHIROPTERA:27.12215773,HIPPOSIDEROS_ARMIGER_HIPPOSIDERIDAE_CHIROPTE
RA:27.12215773):14.33075971):6.278137197,(CYNOPTERUS_BRACHYOTIS_PTEROPODI
DAE_CHIROPTERA:20.4652001,(ROUSETTUS_LESCHENAULTII_PTEROPODIDAE_CHI
ROPTERA:18.85760229,MACROGLOSSUS_SOBRINUS_PTEROPODIDAE_CHIROPTERA
:18.85760229):1.60759781):27.26585454):3.376527137,((TADARIDA_BRASILIENSIS_MOL
OSSIDAE_CHIROPTERA:38.84662166,(MINIOPTERUS_SCHREIBERSII_VESPERTILIONI
DAE_CHIROPTERA:35.18680361,((NYCTICEIUS_HUMERALIS_VESPERTILIONIDAE_C
HIROPTERA:14.7340079,(MYOTIS_LUCIFUGUS_VESPERTILIONIDAE_CHIROPTERA:7.
515808084,MYOTIS_BRANDTII_VESPERTILIONIDAE_CHIROPTERA:7.515808085):7.218
199817):8.24173049,(LASIURUS_BOREALIS_VESPERTILIONIDAE_CHIROPTERA:21.515
63322,((PIPISTRELLUS_PIPISTRELLUS_VESPERTILIONIDAE_CHIROPTERA:19.557674
54,EPTESICUS_FUSCUS_VESPERTILIONIDAE_CHIROPTERA:19.55767453):1.352729024
,ANTROZOUS_PALLIDUS_VESPERTILIONIDAE_CHIROPTERA:20.91040356):0.6052296
572):1.460105173):12.21106521):3.659818054):9.416742027,((PTERONOTUS_PARNELLII_
MORMOOPIDAE_CHIROPTERA:32.43221174,(MACROTUS_CALIFORNICUS_PHYLLOS
TOMIDAE_CHIROPTERA:24.67217485,(MICRONYCTERIS_HIRSUTA_PHYLLOSTOMID
AE_CHIROPTERA:22.53043489,(ARTIBEUS_JAMAICENSIS_PHYLLOSTOMIDAE_CHIR
OPTERA:20.52168271,ANOURA_CAUDIFER_PHYLLOSTOMIDAE_CHIROPTERA:20.521
6827):2.008752192):2.14173996):7.760036889):4.346499691,NOCTILIO_LEPORINUS_NOC
TILIONIDAE_CHIROPTERA:36.77871144):11.48465225):2.84421809):14.81005709):3.88935
0612):2.62732779,((((NYCTICEBUS_COUCANG_LORISIDAE_PRIMATES:17.90620044,OT
OLEMUR_GARNETTII_GALAGIDAE_PRIMATES:17.90620044):26.23173245,(DAUBENT
ONIA_MADAGASCARIENSIS_DAUBENTONIIDAE_PRIMATES:33.75239367,((INDRI_IN
DRI_INDRIIDAE_PRIMATES:10.35644819,PROPITHECUS_COQUERELI_INDRIIDAE_PR
IMATES:10.35644818):9.346121928,((CHEIROGALEUS_MEDIUS_CHEIROGALEIDAE_PR
IMATES:12.38270917,MICROCEBUS_MURINUS_CHEIROGALEIDAE_PRIMATES:12.382
70917):6.512119678,(LEMUR_CATTA_LEMURIDAE_PRIMATES:9.152171277,EULEMUR_
MACACO_LEMURIDAE_PRIMATES:9.152171277):9.742657571):0.8077412649):14.049823
55):10.38553923):13.24350194,(((((CEBUS_ALBIFRONS_CEBIDAE_PRIMATES:1.7235452
91,CEBUS_CAPUCINUS_CEBIDAE_PRIMATES:1.723545291):10.65127951,CALLITHRIX_
JACCHUS_CALLITRICHIDAE_PRIMATES:12.3748248):0.2957299318,AOTUS_NANCYM
AAE_AOTIDAE_PRIMATES:12.67055473):1.342429422,(ATELES_GEOFFROYI_ATELIDA
E_PRIMATES:9.093860866,ALOUATTA_PALLIATA_ATELIDAE_PRIMATES:9.093860866):
4.919123286):13.67223752,(((HYLOBATES_MOLOCH_HYLOBATIDAE_PRIMATES:4.5775
37574,NOMASCUS_LEUCOGENYS_HYLOBATIDAE_PRIMATES:4.577537574):6.2036378
03,(((PAN_TROGLODYTES_HOMINIDAE_PRIMATES:1.724853266,PAN_PANISCUS_HO
MINIDAE_PRIMATES:1.724853266):2.72314225,HOMO_SAPIENS_HOMINIDAE_PRIMAT
ES:4.447995516):0.6435955992,GORILLA_GORILLA_HOMINIDAE_PRIMATES:5.0915911
15):5.689584262):5.917951306,((((TRACHYPITHECUS_FRANCOISI_CERCOPITHECIDAE
_PRIMATES:4.75221723,SEMNOPITHECUS_ENTELLUS_CERCOPITHECIDAE_PRIMATE
S:4.75221723):1.635672304,((RHINOPITHECUS_ROXELLANA_CERCOPITHECIDAE_PRI
MATES:1.83894938,RHINOPITHECUS_BIETI_CERCOPITHECIDAE_PRIMATES:1.838949
379):3.135951081,(PYGATHRIX_NEMAEUS_CERCOPITHECIDAE_PRIMATES:4.4530594
02,NASALIS_LARVATUS_CERCOPITHECIDAE_PRIMATES:4.453059401):0.521841059):1.

35
412989074):1.119750864,COLOBUS_ANGOLENSIS_CERCOPITHECIDAE_PRIMATES:7.5
07640399):2.278581625,(((MACACA_NEMESTRINA_CERCOPITHECIDAE_PRIMATES:3.
429801679,((MACACA_MULATTA_CERCOPITHECIDAE_PRIMATES:0.6851823411,MAC
ACA_FUSCATA_CERCOPITHECIDAE_PRIMATES:0.6851823411):1.196443686,MACACA
_FASCICULARIS_CERCOPITHECIDAE_PRIMATES:1.881626028):1.548175651):3.1734137
33,((THEROPITHECUS_GELADA_CERCOPITHECIDAE_PRIMATES:3.731100513,PAPIO_
ANUBIS_CERCOPITHECIDAE_PRIMATES:3.731100513):1.99722178,((MANDRILLUS_SP
HINX_CERCOPITHECIDAE_PRIMATES:2.080843279,MANDRILLUS_LEUCOPHAEUS_C
ERCOPITHECIDAE_PRIMATES:2.080843279):1.994434017,CERCOCEBUS_ATYS_CERCO
PITHECIDAE_PRIMATES:4.075277295):1.653044997):0.8748931195):1.572996215,(((ERYT
HROCEBUS_PATAS_CERCOPITHECIDAE_PRIMATES:3.827355635,CHLOROCEBUS_SA
BAEUS_CERCOPITHECIDAE_PRIMATES:3.827355634):0.7879280385,(CERCOPITHECUS
_NEGLECTUS_CERCOPITHECIDAE_PRIMATES:3.724458383,CERCOPITHECUS_MONA
_CERCOPITHECIDAE_PRIMATES:3.724458383):0.89082529):2.09335908,ALLENOPITHE
CUS_NIGROVIRIDIS_CERCOPITHECIDAE_PRIMATES:6.708642753):1.467568874):1.610
010397):6.91290466):10.98609498):29.69621317):10.0030487,(((ORYCTOLAGUS_CUNICU
LUS_LEPORIDAE_LAGOMORPHA:8.134735543,(LEPUS_AMERICANUS_LEPORIDAE_L
AGOMORPHA:3.768477375,LEPUS_TIMIDUS_LEPORIDAE_LAGOMORPHA:3.76847737
5):4.366258169):22.69708713,OCHOTONA_PRINCEPS_OCHOTONIDAE_LAGOMORPHA:
30.83182267):33.2377761,(((((XERUS_INAURIS_SCIURIDAE_RODENTIA:21.93137986,M
ARMOTA_MONAX_SCIURIDAE_RODENTIA:21.93137986):1.524282169,(SCIURUS_VUL
GARIS_SCIURIDAE_RODENTIA:8.870042841,SCIURUS_CAROLINENSIS_SCIURIDAE_
RODENTIA:8.87004284):14.58561919):19.83267411,(GRAPHIURUS_MURINUS_GLIRIDA
E_RODENTIA:23.07085132,GLIS_GLIS_GLIRIDAE_RODENTIA:23.07085132):20.2174848
1):11.65772354,((THRYONOMYS_SWINDERIANUS_THRYONOMYIDAE_RODENTIA:17.
70944603,PETROMUS_TYPICUS_PETROMURIDAE_RODENTIA:17.70944603):17.656648
3,(((DASYPROCTA_PUNCTATA_DASYPROCTIDAE_RODENTIA:19.66240464,(CAVIA_A
PEREA_CAVIIDAE_RODENTIA:14.15153389,DOLICHOTIS_PATAGONUM_CAVIIDAE_R
ODENTIA:14.15153389):5.510870749):9.726471246,COENDOU_PREHENSILIS_ERETHIZO
NTIDAE_RODENTIA:29.38887589):1.081626225,(CHINCHILLA_LANIGERA_CHINCHILL
IDAE_RODENTIA:28.38487906,MYOCASTOR_COYPUS_MYOCASTORIDAE_RODENTI
A:28.38487907):2.085623047):4.895592226):19.57996533):4.56286807,(((CASTOR_CANAD
ENSIS_CASTORIDAE_RODENTIA:46.65179866,(DIPODOMYS_ORDII_HETEROMYIDAE
_RODENTIA:5.093209271,DIPODOMYS_STEPHENSI_HETEROMYIDAE_RODENTIA:5.0
93209271):41.5585894):9.132746136,((ZAPUS_HUDSONIUS_DIPODIDAE_RODENTIA:21.
24272273,JACULUS_JACULUS_DIPODIDAE_RODENTIA:21.24272273):29.51526029,((((R
ATTUS_NORVEGICUS_MURIDAE_RODENTIA:6.71134861,RATTUS_RATTUS_MURIDA
E_RODENTIA:6.71134861):18.37218667,(UROMYS_CAUDIMACULATUS_MURIDAE_RO
DENTIA:24.91130559,((MASTOMYS_COUCHA_MURIDAE_RODENTIA:23.28022013,(MU
S_SPRETUS_MURIDAE_RODENTIA:22.69412458,APODEMUS_SYLVATICUS_MURIDAE
_RODENTIA:22.69412458):0.5860955514):0.7993061946,ARVICANTHIS_NILOTICUS_MU
RIDAE_RODENTIA:24.07952633):0.8317792662):0.1722296806):18.38697726,(PSAMMOM
YS_OBESUS_MURIDAE_RODENTIA:36.99411843,(ACOMYS_RUSSATUS_MURIDAE_R
ODENTIA:11.74118596,ACOMYS_CAHIRINUS_MURIDAE_RODENTIA:11.74118596):25.2
5293247):6.476394107):1.068092127,(CRICETOMYS_GAMBIANUS_NESOMYIDAE_ROD
ENTIA:43.42493655,(SIGMODON_HISPIDUS_CRICETIDAE_RODENTIA:37.21164474,((N

36
EOTOMA_LEPIDA_CRICETIDAE_RODENTIA:25.05916574,(((PEROMYSCUS_MANICUL
ATUS_CRICETIDAE_RODENTIA:1.839804535,PEROMYSCUS_POLIONOTUS_CRICETID
AE_RODENTIA:1.839804535):2.685687587,PEROMYSCUS_LEUCOPUS_CRICETIDAE_R
ODENTIA:4.525492122):7.54062027,PEROMYSCUS_EREMICUS_CRICETIDAE_RODENT
IA:12.06611239):12.99305335):9.462848672,(((MICROTUS_ARVALIS_CRICETIDAE_RODE
NTIA:7.017706291,MICROTUS_AGRESTIS_CRICETIDAE_RODENTIA:7.017706292):4.52
7006731,ARVICOLA_AMPHIBIUS_CRICETIDAE_RODENTIA:11.54471302):1.937676735,
ONDATRA_ZIBETHICUS_CRICETIDAE_RODENTIA:13.48238976):21.03962465):2.689630
339):6.213291798):1.113668111):6.219378359):5.026561786):0.8393244665,PEDETES_CAPE
NSIS_PEDETIDAE_RODENTIA:56.62386927):2.88505847):4.560671027):3.31488477):5.049
833733):1.551388369,((MYRMECOPHAGA_TRIDACTYLA_MYRMECOPHAGIDAE_PILO
SA:58.14657783,DASYPUS_NOVEMCINCTUS_DASYPODIDAE_CINGULATA:58.1465778
3):14.82885348,(((MICROGALE_TALAZACI_TENRECIDAE_AFROSORICIDA:36.2898605
2,ECHINOPS_TELFAIRI_TENRECIDAE_AFROSORICIDA:36.28986051):21.73254843,CHR
YSOCHLORIS_ASIATICA_CHRYSOCHLORIDAE_AFROSORICIDA:58.02240894):4.62088
1431,(((DUGONG_DUGON_DUGONGIDAE_SIRENIA:37.43241219,ELEPHAS_MAXIMUS
_ELEPHANTIDAE_PROBOSCIDEA:37.4324122):6.356241737,HETEROHYRAX_BRUCEI_
PROCAVIIDAE_HYRACOIDEA:43.78865394):13.66005434,ORYCTEROPUS_AFER_ORYC
TEROPODIDAE_TUBULIDENTATA:57.44870827):5.194582105):10.33214093):1.01027433):
83.25205117,((MONODELPHIS_DOMESTICA_DIDELPHIDAE_DIDELPHIMORPHIA:26.2
5982614,DIDELPHIS_VIRGINIANA_DIDELPHIDAE_DIDELPHIMORPHIA:26.25982614):4
2.52387203,((SARCOPHILUS_HARRISII_DASYURIDAE_DASYUROMORPHIA:18.641153
58,ANTECHINUS_FLAVIPES_DASYURIDAE_DASYUROMORPHIA:18.64115358):37.0673
6858,(PHASCOLARCTOS_CINEREUS_PHASCOLARCTIDAE_DIPROTODONTIA:44.5716
0159,(((MACROPUS_GIGANTEUS_MACROPODIDAE_DIPROTODONTIA:1.816883337,M
ACROPUS_FULIGINOSUS_MACROPODIDAE_DIPROTODONTIA:1.816883337):4.607144
204,(MACROPUS_RUFUS_MACROPODIDAE_DIPROTODONTIA:5.873321851,MACROP
US_EUGENII_MACROPODIDAE_DIPROTODONTIA:5.873321851):0.5507056892):30.5520
3535,PSEUDOCHIROPS_CUPREUS_PSEUDOCHEIRIDAE_DIPROTODONTIA:36.9760628
9):7.595538702):11.13692056):13.07517602):88.45405864):27.05196289);

37
Supplementary Figure 1.3
Double click on the Excel spreadsheet and scroll through table
geographyorder N_speciesmean_xcount_per_species whatis_x whatis_y phylolm_xestimate phylolm_r.squared phylolm_p.value
total_genome all_orders 268 6.08399 log_hklk log_SSD -0.0806 0.00418 0.29168
total_genome all_orders 268 6.89244 log_mvdp log_SSD 0.01098 0.0001 0.86854
total_genome all_orders 268 4.34969 log_p21 log_SSD -0.10408 0.00999 0.1025
total_genome all_orders 268 3.9669 log_c3 log_SSD 0.0475 0.00173 0.4981
total_genome all_orders 268 5.25011 log_pb2 log_SSD -0.00783 5.35E-05 0.90515
total_genome all_orders 268 5.97366 log_pem1 log_SSD -0.07509 0.00172 0.4992
total_genome all_orders 268 6.26023 log_pem2 log_SSD -0.15976 0.00919 0.11747
total_genome all_orders 268 6.7211 log_psa1 log_SSD 0.01056 0.00025 0.79747
total_genome all_orders 268 5.50923 log_psa3 log_SSD -0.0029 1.94E-06 0.9819
total_genome all_orders 268 6.11254 log_sc12 log_SSD 0.00663 2.31E-05 0.93763
total_genome all_orders 268 6.13622 log_slp2 log_SSD -0.04009 0.00111 0.58754
total_genome all_orders 268 6.75967 log_slp3 log_SSD -0.15159 0.00495 0.25123
total_genome all_orders 268 6.46708 log_sre1 log_SSD 0.03249 0.00076 0.65296
total_genome all_orders 268 5.83498 log_sre2 log_SSD -0.00675 3.20E-05 0.9266
total_genome all_orders 268 6.13806 log_ccnd1log_SSD 0.00377 5.86E-06 0.96855
total_genome all_orders 268 6.13827 log_hk2 log_SSD -0.10547 0.00443 0.2775
total_genome all_orders 268 6.65673 log_are log_SSD -0.00847 7.07E-05 0.89099
total_genome all_orders 268 5.36402 log_selarelog_SSD 0.1103 0.00398 0.30348
total_genome all_orders 268 5.80912 log_ere log_SSD -0.01579 8.94E-05 0.87754
total_genome all_orders 268 9.12334 log_total_ARElog_SSD 0.02584 0.00042 0.73879
total_genome all_orders 268 9.16095 log_combined_total_are_ere log_SSD 0.02598 0.00041 0.74295
total_genome all_orders 268 9.44485 positive_effect_AREs:mvdp_c3_psa1_sre1_selare_combined_total_are_ere log_SSD 0.03613 0.00097 0.61199
total_genome all_orders 268 8.67998 negative_effect_AREs:hklk_p21_pb2_pem1_pem2_psa3_sc12_slp2_slp3_sre2_ccnd1_hk2_are log_SSD -0.06348 0.00163 0.5107
total_genome all_orders 268 -0.32347 scale_center_PC1_28percent log_SSD 0.01072 0.00188 0.47915
total_genome all_orders 268 -0.06749 scale_center_PC2_23percent log_SSD -0.00913 0.00143 0.53801
total_genome all_orders 268 -0.30817 scale_center_PC3_14percent log_SSD 0.01841 0.00241 0.42383
total_genome all_orders 268 -0.70059 scale_center_PC4_11percent log_SSD -0.0117 0.00196 0.47086
total_genome all_orders 268 922.498 not_scale_center_PC1_76percent log_SSD -3.59E-06 0.00375 0.31785
total_genome all_orders 268 1222.05 not_scale_center_PC2_15percent log_SSD 6.84E-06 0.00334 0.34584
total_genome CARNIVORA 48 5.98429 log_hklk log_SSD -0.45653 0.02952 0.24291
total_genome CARNIVORA 48 6.79262 log_mvdp log_SSD -0.1613 0.00452 0.64982
total_genome CARNIVORA 48 4.24363 log_p21 log_SSD -0.36121 0.03285 0.21764
total_genome CARNIVORA 48 4.01758 log_c3 log_SSD -0.28804 0.01775 0.36665
total_genome CARNIVORA 48 5.17832 log_pb2 log_SSD -0.19695 0.00629 0.59221

38
Supplementary Figure 1.5
Double click on the Excel spreadsheet and scroll through table
order N
Species
Mean
SSD
Mean
ARE
Mean
body size
p.value,
ARE or
ERE
coefficien
t,
ARE or
ERE
p.value,
body size
ALL 268 0.209 7.943 2.096 0.126 0.061 **0.001
CARNIVORA 48 0.375 7.947 3.154 **0.01 -0.502 **0
CETARTIODACTYLA 83 0.195 7.759 5.203 0.34 0.07 **0.006
PRIMATES 42 0.327 7.94 1.798 0.318 -0.113 *0.054
CHIROPTERA 21 -0.058 7.586 -3.974 **0.035 0.243 0.714
RODENTIA 41 0.073 8.375 -1.449 0.059 0.07 0.702
ALL 268 0.209 8.452 2.096 *0.074 0.072 **0.001
CARNIVORA 48 0.375 8.461 3.154 **0.008 -0.876 **0
CETARTIODACTYLA 83 0.195 8.272 5.203 0.649 0.032 **0.005
PRIMATES 42 0.327 8.422 1.798 0.472 -0.076 **0.052
CHIROPTERA 21 -0.058 8.044 -3.974 **0.016 0.237 0.758
RODENTIA 41 0.073 8.851 -1.449 0.063 0.073 0.707
ALL 268 0.209 4.59 2.096 0.222 0.051 **0.001
CARNIVORA 48 0.375 4.709 3.154 *0.08 -0.333 **0.017
CETARTIODACTYLA 83 0.195 4.408 5.203 0.389 0.06 **0.006
PRIMATES 42 0.327 4.684 1.798 0.391 -0.091 **0.048
CHIROPTERA 21 -0.058 4.335 -3.974 **0.002 0.329 0.481
RODENTIA 41 0.073 4.905 -1.449 0.766 0.018 0.827
ALL 268 0.209 5.167 2.096 0.144 0.061 **0.001
CARNIVORA 48 0.375 5.303 3.154 **0.047 -0.658 **0.008
CETARTIODACTYLA 83 0.195 5.051 5.203 0.708 0.025 **0.005
PRIMATES 42 0.327 5.183 1.798 0.468 -0.077 **0.05
CHIROPTERA 21 -0.058 4.858 -3.974 **0.002 0.273 0.691
RODENTIA 41 0.073 5.407 -1.449 0.737 0.022 0.84
ERE
Supplemental File 4. Repeat of Table 1, but with the 1o Kb and 100 Kb cutoffs. Results of phylogenetically controlled linear ARE

39
Supplementary Figure 1.5
are.pos05 BP
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
positive regulation of phospholipid transport (GO:2001140)14 7 1.3151071282354547
+ 1.2095251121017345E-4 5.322760290556901 0.030535552010437235
regulation of phospholipid transport (GO:2001138) 15 7 1.4090433516808443 +
2.083905062804792E-4 4.967909604519774 0.04719432053999088
regulation of lipid transport (GO:0032368) 158 30 14.841923304371559 +
1.7156026719444207E-4 2.0213013761608485 0.041281689293662625
regulation of transport (GO:0051049) 1846 220 173.40626848018923 +
1.7172408193337846E-4 1.268696927326672 0.04068539787344659
positive regulation of biological process (GO:0048518) 6391 674 600.346404039485
+ 2.0117948318993487E-4 1.1226851622079022 0.0469418794109848
response to pheromone (GO:0019236) 105 32 9.86330346176591 +
1.05347772930652E-9 3.2443491294823015 2.3176510044743444E-6
response to organic substance (GO:0010033) 2499 290 234.74662239002865 +
9.538080288241725E-5 1.2353745372240907 0.025769550252442555
positive regulation of cholesterol transport (GO:0032376) 46 13 4.321066278487923
+ 2.2406182175248248E-4 3.008516685966944 0.04929360078554614
regulation of cholesterol transport (GO:0032374) 71 19 6.669471864622663 +
2.0415814609237894E-5 2.8488012822698905 0.007311710348424734
regulation of sterol transport (GO:0032371) 71 19 6.669471864622663 +
2.0415814609237894E-5 2.8488012822698905 0.007145535113233264
positive regulation of sterol transport (GO:0032373) 46 13 4.321066278487923
+ 2.2406182175248248E-4 3.008516685966944 0.048599324718144085
regulation of MAP kinase activity (GO:0043405) 166 31 15.593413091934677 +
2.3168365485358805E-4 1.9880189037019749 0.04955455951035078
regulation of cellular metabolic process (GO:0031323) 5927 653 556.7599963608243
+ 8.300167173839891E-7 1.172857253157974 5.32594060321393E-4
regulation of metabolic process (GO:0019222) 6551 712 615.3761997907474 +
1.3649153488018213E-6 1.1570158225848004 8.084498604441557E-4
regulation of primary metabolic process (GO:0080090) 5594 637 525.4792339535096
+ 6.870093784069944E-9 1.2122267805093834 8.13841879035978E-6
regulation of macromolecule metabolic process (GO:0060255) 6031 653
566.5293635991447 + 1.0445343168175211E-5 1.1526322234235307
0.004468285688608285
regulation of nitrogen compound metabolic process (GO:0051171) 5404 614
507.6313514988855 + 2.246114706687511E-8 1.2095391629910943
2.161885405186729E-5
regulation of vesicle-mediated transport (GO:0060627) 609 88 57.20716007824228
+ 4.1136207517228276E-5 1.5382689838062635 0.013771686864463378
carbohydrate derivative biosynthetic process (GO:1901137)558 80 52.41641268252741
+ 1.2444539142630645E-4 1.5262395098371042 0.030910629483308378

40
positive regulation of transcription by RNA polymerase II (GO:0045944) 1278 159
120.05049356320794 + 2.055705358918678E-4 1.324442701406177
0.04725054108559349
positive regulation of DNA-templated transcription (GO:0045893)1609 198
151.1433835236319 + 6.251335872883694E-5 1.310014341243339
0.018513571623540167
regulation of DNA-templated transcription (GO:0006355) 3185 370 299.186871673566
+ 5.698393141206005E-6 1.2366852794386518 0.0029251751458190824
regulation of RNA biosynthetic process (GO:2001141) 3196 371 300.22017013146524
+ 5.882687645637897E-6 1.2357597420504443 0.002922367411058826
regulation of macromolecule biosynthetic process (GO:0010556) 4863 541
456.81185461492976 + 4.230795563344727E-6 1.1842950101547534
0.002246698333638234
regulation of cellular biosynthetic process (GO:0031326) 5021 561
471.65377791930126 + 1.4079729382391758E-6 1.1894317956592
8.030660462549373E-4
regulation of biosynthetic process (GO:0009889) 5065 566 475.78697175089843
+ 1.15248870249612E-6 1.1896080254512165 7.0993304073761E-4
regulation of RNA metabolic process (GO:0051252) 3505 414 329.24646317609063
+ 1.8208507990554916E-7 1.257416696314155 1.5578390169696985E-4
regulation of nucleobase-containing compound metabolic process (GO:0019219) 3806
448 357.5212664331529 + 7.230302087594965E-8 1.2530723122277938
6.549803067586026E-5
regulation of gene expression (GO:0010468) 4736 525 444.8819542373653 +
9.941895189053222E-6 1.1800883245533669 0.004503093703277047
positive regulation of RNA biosynthetic process (GO:1902680) 1612 198
151.42519219396806 + 6.412905247440296E-5 1.3075763492931345
0.018633724681241616
positive regulation of RNA metabolic process (GO:0051254) 1757 214
165.04594459354956 + 5.936927174613882E-5 1.2966086535904118
0.017927191860598782
positive regulation of nucleobase-containing compound metabolic process (GO:0045935)
1973 240 185.33616885775373 + 2.087855765821459E-5
1.2949442166585465 0.0071451063985889934
positive regulation of nitrogen compound metabolic process (GO:0051173) 2999 344
281.7147341127235 + 4.500461503653683E-5 1.221093391098082
0.014438980657555564
positive regulation of metabolic process (GO:0009893) 3808 421 357.70913888004367
+ 1.5030376901743167E-4 1.1769338667670457 0.03674092131537218
regulation of transcription by RNA polymerase II (GO:0006357) 2470 304
232.02247191011236 + 3.450235854683209E-7 1.3102179176755449
2.6566816081060706E-4
regulation of cellular component organization (GO:0051128) 2595 307
243.76449984078607 + 1.0267590913011043E-5 1.2594122614265653
0.004517740001724858

41
negative regulation of cellular biosynthetic process (GO:0031327) 2124 251
199.52053859800756 + 8.949591589701326E-5 1.2580158502163672
0.024611376871678643
negative regulation of biosynthetic process (GO:0009890) 2147 254 201.6810717372515
+ 8.172007253486986E-5 1.2594141721485355 0.02288162030976356
negative regulation of biological process (GO:0048519) 5522 614 518.7158258654415
+ 6.058344925676492E-7 1.183692436943838 4.442786278829428E-4
negative regulation of cellular metabolic process (GO:0031324) 2587 301
243.01301005322296 + 4.953602452690215E-5 1.2386168128779491
0.015568464851312105
negative regulation of cellular process (GO:0048523) 5040 570 473.4385661647637
+ 1.9543183161362925E-7 1.2039576847688227 1.584026424657837E-4
transport (GO:0006810) 3520 400 330.6555065277714 +
2.0406981689244693E-5 1.2097182478538413 0.007482559952723054
establishment of localization (GO:0051234)3791 422 356.11222308147205 +
7.793901770867967E-5 1.1850197006673764 0.02222705319840124
localization (GO:0051179) 4409 492 414.16480917072283 +
1.1207133906064383E-5 1.1879328931521864 0.004664590869010581
cellular component organization (GO:0016043) 5478 597 514.5826320338443 +
1.472423926705687E-5 1.1601635244477801 0.005967191702965152
cellular component organization or biogenesis (GO:0071840) 5691 620
534.5910476277123 + 9.118016173864773E-6 1.1597650255298817
0.004255074214470227
nitrogen compound metabolic process (GO:0006807) 6351 681 596.5889551016695
+ 2.035752806214094E-5 1.1414894529583528 0.007646486150170012
primary metabolic process (GO:0044238) 6912 737 649.2871764545331 +
1.6313373279184324E-5 1.1350909531656352 0.006441690987165092
G protein-coupled receptor signaling pathway (GO:0007186) 1904 131
178.85456944002183 - 5.457131680923056E-5 0.7324386534274727
0.01680796557724301
detection of chemical stimulus involved in sensory perception of smell (GO:0050911) 548
16 51.47705044807351 - 2.91351484559286E-9
0.31081811915660734 4.4868128622130045E-6
detection of chemical stimulus involved in sensory perception (GO:0050907) 605 21
56.83141518446072 - 1.7294464009322274E-8 0.3695139375262642
1.90239104102545E-5
detection of stimulus involved in sensory perception (GO:0050906) 697 23
65.47354774143656 - 3.5478807197124296E-10 0.3512869058330236
9.106227180595235E-7
sensory perception (GO:0007600) 1669 91 156.77955693035528 -
1.1018522853704655E-9 0.5804328177839161 2.121065649338146E-6
nervous system process (GO:0050877) 2224 150 208.91416094254652 -
3.3457190555795905E-6 0.7179982406326755 0.001840145480568775
system process (GO:0003008) 2798 208 262.8335532002002 -
1.0127121682593427E-4 0.7913753684316192 0.026889254122748067

42
detection of stimulus (GO:0051606) 805 32 75.61865987353863 -
3.7007226339252226E-9 0.42317597341073504 4.749260713537369E-6
detection of chemical stimulus (GO:0009593) 637 23 59.83737433471319 -
1.76797796545957E-8 0.38437515442012155 1.815124044538492E-5
sensory perception of chemical stimulus (GO:0007606) 1262 54 118.54751398808169
- 2.3567884760034335E-12 0.45551355893830847
1.8147271265226438E-8
sensory perception of smell (GO:0007608) 1162 46 109.15389164354274 -
6.316204808024769E-13 0.421423362075073 9.726955404358144E-9
immunoglobulin production (GO:0002377) 253 4 23.765864531683576 -
6.173563932048339E-7 0.16830862578836048 4.3214947524338365E-4
production of molecular mediator of immune response (GO:0002440) 320 9
30.05959150252468 - 6.633170229355355E-6 0.29940526634382564
0.0031922131728772643
immune system process (GO:0002376) 2468 182 231.83459946322156 -
2.1312221007138105E-4 0.7850424415570146 0.0475664063057865
immunoglobulin mediated immune response (GO:0016064)288 3 27.053632352272214
- 2.9780663071066903E-9 0.11089083938660209
4.1692928299493666E-6
B cell mediated immunity (GO:0019724) 293 3 27.523313469499158 -
2.0316697020463197E-9 0.10899850424348602 3.476412601279258E-6
adaptive immune response based on somatic recombination of immune receptors built from
immunoglobulin superfamily domains (GO:0002460) 389 6 36.54119092025656
- 2.0023381645661752E-10 0.16419826089121545
6.167201546863819E-7
adaptive immune response (GO:0002250) 558 14 52.41641268252741 -
8.634086214280388E-11 0.2670919142214932 4.4321642566639324E-7
immune response (GO:0006955) 1701 105 159.78551608060775 -
7.55195770786522E-7 0.6571309000687532 5.056528204396713E-4
lymphocyte mediated immunity (GO:0002449) 374 5 35.13214756857572 -
1.392014690554145E-10 0.14231979386515778 5.359256558633458E-7
leukocyte mediated immunity (GO:0002443) 437 18 41.05012964563526 -
4.252677130129338E-5 0.43848826192230766 0.013934303788083363
immune effector process (GO:0002252) 597 28 56.0799253968976 -
1.794193463828986E-5 0.4992873974391733 0.006907644835741597
are.pos05 MF
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
sphingolipid floppase activity (GO:0046623) 4 4 0.37574489378155845
+ 7.765838239923133E-5 10.645520581113802 0.02774068359846828
floppase activity (GO:0140328) 7 6 0.6575535641177274 +
4.394196209379965E-6 9.124731926668971 0.003662562540518201
phosphatidylcholine floppase activity (GO:0090554) 6 5 0.5636173406723377
+ 4.028000406169803E-5 8.871267150928167 0.01831275457386835

43
pheromone binding (GO:0005550) 91 30 8.548196333530456 +
4.1608124322175446E-10 3.5095122794880664 6.936074324506647E-7
binding (GO:0005488) 14170 1421 1331.076286221171 +
1.2246743440392751E-5 1.067557145078526 0.0061245963945404155
pheromone receptor activity (GO:0016503) 99 30 9.299686121093574 +
4.033327595917201E-9 3.2259153276102426 5.042667826795481E-6
transmembrane signaling receptor activity (GO:0004888) 2205 160 207.1293726970841
- 2.1658589186598283E-4 0.772464078448167 0.04923391114644455
RNA polymerase II transcription regulatory region sequence-specific DNA binding
(GO:0000977)1322 165 124.18368739480508 + 1.4349083465374046E-4
1.3286769257819797 0.037768298110702946
transcription cis-regulatory region binding (GO:0000976) 1421 178 133.48337351589865
+ 6.170058359034198E-5 1.3334994112865985 0.02373573988733079
transcription regulatory region nucleic acid binding (GO:0001067) 1426 179
133.9530546331256 + 5.1907303164617884E-5 1.3362890491019428
0.021632368593854503
nucleic acid binding (GO:0003676)3387 378 318.16198880953465 +
1.771721438461437E-4 1.1880740418249238 0.042192280541645935
organic cyclic compound binding (GO:0097159) 5485 615 515.2401855979621 +
1.5979888120303012E-7 1.1936180779188674 1.5983084097927073E-4
sequence-specific double-stranded DNA binding (GO:1990837) 1477 182
138.74380202884046 + 1.2254368775727552E-4 1.3117703085732646
0.03830256140463343
double-stranded DNA binding (GO:0003690) 1577 192 148.13742437337942 +
1.3777776860396916E-4 1.2960938183727648 0.03827925671046943
DNA binding (GO:0003677) 2308 277 216.80480371195927 +
1.1732545129227114E-5 1.2776469674906943 0.006519384243473866
sequence-specific DNA binding (GO:0043565) 1577 193 148.13742437337942 +
1.1303870481892254E-4 1.302844307010123 0.037687104186628774
ion binding (GO:0043167) 5415 580 508.6646499567848 +
1.542677815367331E-4 1.1402404315874433 0.03857465877326011
protein binding (GO:0005515) 9535 978 895.68189055179 +
1.300770792164823E-4 1.0919055194891765 0.038265616068331065
olfactory receptor activity (GO:0004984) 1169 47 109.81144520766047 -
1.079783121423079E-12 0.4280063877778859 5.399995390236818E-9
antigen binding (GO:0003823) 287 2 26.95969612882682 -
3.238988607608597E-10 0.07418481241194287 8.099091013325296E-7
are.pos05 CC
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
secretory granule (GO:0030141) 494 71 46.40449438202247 +
3.0810035536031643E-4 1.5300242130750605 0.043838279134125024
intracellular membrane-bounded organelle (GO:0043231) 11653 1262
1094.6388118091252 + 7.60563856938718E-15 1.1528916994220904
3.030086406043852E-12

44
membrane-bounded organelle (GO:0043227) 12147 1321 1141.0433061911478 +
4.029385167611689E-17 1.1577124135713617 2.6755117512941613E-14
organelle (GO:0043226) 13112 1428 1231.6917618159487 +
6.599767181698988E-21 1.159380978480057 1.3146736225944385E-17
cellular anatomical entity (GO:0110165) 19044 1863 1788.921439294 +
2.2986104607351725E-7 1.0414096220654805 5.723540047230579E-5
intracellular organelle (GO:0043229) 12792 1388 1201.632170313424 +
1.0198322516455529E-18 1.155095572747495 1.0157529226389707E-15
intracellular anatomical structure (GO:0005622) 14420 1516 1354.5603420825184 +
1.2671393418521082E-15 1.1191823301642525 6.310353922423499E-13
cytoplasm (GO:0005737) 11670 1229 1096.2357276076968 +
6.932665838912228E-10 1.1211092368627988 2.3016450585188597E-7
microtubule cytoskeleton (GO:0015630) 1353 166 127.09571032161216 +
3.0147163673654835E-4 1.3061023033739032 0.046194730798400335
intracellular non-membrane-bounded organelle (GO:0043232) 4724 527
443.7547195560206 + 4.392397400580135E-6 1.1875930030158708
9.721839579950699E-4
non-membrane-bounded organelle (GO:0043228) 4725 527 443.84865577946596 +
4.405632852580665E-6 1.18734166058137 8.776020642340685E-4
organelle membrane (GO:0031090) 2583 294 242.63726515944137 +
3.260815850311932E-4 1.2116852693950668 0.04330363449214245
nucleoplasm (GO:0005654) 3700 406 347.56402674794157 +
3.7856151704186133E-4 1.1681300961978929 0.03968918641828357
nuclear lumen (GO:0031981) 4170 453 391.7140517672747 +
3.992360023631266E-4 1.1564558329123626 0.03976390583536741
intracellular organelle lumen (GO:0070013) 4632 499 435.1125869990447 +
3.5293834421205747E-4 1.146829613552631 0.04394082385440116
organelle lumen (GO:0043233) 4633 499 435.20652322249015 +
3.5393843081380855E-4 1.146582078561577 0.041473256128300384
membrane-enclosed lumen (GO:0031974)4633 499 435.20652322249015 +
3.5393843081380855E-4 1.146582078561577 0.03916918634339481
nucleus (GO:0005634) 7212 786 677.4680434881499 +
1.3348293963174374E-7 1.160202326227877 3.798543082091908E-5
are.neg05 BP
No enrichment
are.neg05 MF
No enrichment
are.neg05 CC
No enrichment
ere.pos05 BP
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr

45
oxidative demethylation (GO:0070989) 15 7 0.5540645043897557 +
4.5394078827864435E-7 12.63390804597701 0.006990688139491123
ere.pos05 MF
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
testosterone 6-beta-hydroxylase activity (GO:0050649) 8 5 0.2955010690078697
+ 3.4667206200039404E-6 16.920412561576356 0.017337069820639706
estrogen 16-alpha-hydroxylase activity (GO:0101020) 17 6 0.6279397716417231
+ 2.1755432158221573E-5 9.555056505360765 0.02719972905581652
nucleosomal DNA binding (GO:0031492) 54 11 1.9946322158031207 +
3.655308624024421E-6 5.514801131180441 0.009140099214373065
nucleic acid binding (GO:0003676)3387 170 125.10776509120684 +
1.9375279624106645E-5 1.3588285257598962 0.032298591133385776
ere.pos05 CC
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
CENP-A containing nucleosome (GO:0043505) 12 8 0.44325160351180454
+ 1.4537412083729744E-9 18.048440065681447 2.895852487078965E-6
CENP-A containing chromatin (GO:0061638) 12 8 0.44325160351180454
+ 1.4537412083729744E-9 18.048440065681447 1.4479262435394824E-6
chromosome, centromeric core domain (GO:0034506) 13 8 0.4801892371377883
+ 3.656991866027435E-9 16.660098522167488 2.428242599042217E-6
nucleosome (GO:0000786) 104 16 3.8415138971023066 +
1.3198590242552695E-6 4.165024630541872 6.572897940791242E-4
ere.neg05 BP
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
collagen catabolic process (GO:0030574) 35 8 0.665514261019879 +
2.3992775545135973E-7 12.020779220779222 0.00369488743395094
collagen metabolic process (GO:0032963) 61 9 1.1598962834917892 +
2.1567535789092706E-6 7.759314456035767 0.016607002557601384
ere.neg05 MF
term number_in_reference number_in_list expected plus_minus pValue
fold_enrichment fdr
acetylcholine receptor inhibitor activity (GO:0030550) 14 4 0.2662057044079516
+ 1.1089143186402823E-4 15.025974025974026 0.055456805075200524
acetylcholine receptor regulator activity (GO:0030548) 26 5
0.49438202247191015 + 1.1480995931291572E-4 10.113636363636363
0.05219678241126286
neurotransmitter receptor regulator activity (GO:0099602) 26 5
0.49438202247191015 + 1.1480995931291572E-4 10.113636363636363
0.04784705054365762

46
cysteine-type endopeptidase activator activity involved in apoptotic process (GO:0008656)
27 6 0.5133967156439067 + 9.620620341741393E-6
11.686868686868687 0.02405636116452435
peptidase activator activity involved in apoptotic process (GO:0016505) 29 6
0.5514261019878998 + 1.4949726336597142E-5 10.880877742946709
0.018690895352330577
peptidase activator activity (GO:0016504) 68 8 1.2929991356957649 +
4.358614583375533E-5 6.18716577540107 0.036329052552435064
cysteine-type endopeptidase regulator activity involved in apoptotic process (GO:0043028)
62 9 1.1789109766637857 + 2.481045697011155E-6
7.6341642228739 0.012407709530752787
metalloendopeptidase activity (GO:0004222) 119 10 2.2627484874675887 +
9.349217255626545E-5 4.419404125286478 0.05195048388376484
endopeptidase activity (GO:0004175) 440 23 8.366464995678479 +
1.3692463124530391E-5 2.749070247933884 0.02282533602859216
peptidase activity (GO:0008233) 645 28 12.264477095937771 +
5.20521846123107E-5 2.283016208597604 0.037187567892309405
hydrolase activity (GO:0016787) 2392 72 45.48314606741573 +
9.051109256256174E-5 1.58300395256917 0.05658074673817141
metallopeptidase activity (GO:0008237) 199 14 3.7839239412273122 +
2.9431390762362294E-5 3.6998629511192322 0.029437277040514763
ere.neg05 CC
No enrichment

47
Supplementary Figure 1.6

48
Chapter 3: Manuscript Status and Contribution
The status of the manuscript currently titled “Androgen receptor (AR) Co-immuno precipitation
and sequencing (ChIP-seq) and single-cell RNA-sequencing identify major differences in cells
that give rise to the baculum (penis bone) compared to cells that give rise to forelimb and
hindlimb skeletal system” is in prep and almost ready for submission.
My contribution to this project has involved the majority of the manuscript. I was involved in
data collection, single-cell RNA sequencing, AR ChIP-seq tissue preparation, sequencing and
data analysis, and writing of the manuscript.

49
Chapter 3: Androgen receptor (AR) Co-immuno precipitation and sequencing (ChIP-seq)
and single-cell RNA-sequencing identify major differences in cells that give rise to the
baculum (penis bone) compared to cells that give rise to forelimb and hindlimb skeletal
system.
Abstract
The baculum, a bone in the penis of many mammal species, is unusual compared to the rest of
the skeleton. It shows an astonishing level of morphological divergence across species and
requires androgen to form. Its reliance on androgen makes it unlike other bones such as those in
the forelimb and hindlimb, which form even in the absence of androgen signaling. To better
understand the molecular basis of this developmental difference, we used ChIP-seq to compare
androgen receptor (AR) binding sites in developing penis, forelimb, and hindlimb tissue. We
identified roughly 9K AR binding sites from developing penis, which were significantly enriched
near genes with known roles in cartilage and bone formation. In contrast, we identified no AR
binding sites from developing forelimb or hindlimb. To elucidate possible downstream effects on
gene expression, we compared single-cell RNA-seq in these same three tissues. We identified
chondrocytes from all three tissue types, but the chondrocytes from the developing penis formed
a distinct cluster, with 139 (48) genes significantly up- (down-) regulated compared to
forelimb+hindlimb chondrocytes. Over half of these genes had at least one AR peak identified in
our ChIP-seq data. One of the most highly upregulated genes was Ar, providing a potential link
between androgen signaling and baculum development.

50
Introduction
While much is known about the developmental genetic basis of bone and cartilage, our
knowledge of how these programs vary across different bone systems is incomplete. In addition,
most of what we know about bone and cartilage development focuses on limb systems [151-
155]. These studies have hinted at different genetic mechanisms of development across skeletal
elements. For example, the T-box transcription factor Tbx5 is expressed in forelimb, while Tbx4,
Pitx1, Pitx2, is expressed in hindlimb [153, 156-160]. Interestingly, while Tbx5/4 are expressed
differently, they both target the same fibroblast growth factor Fgf10, which is necessary for
maintaining the outgrowth of limb buds [161], and there many “classic” bone genes such as
Runx2, Osterix, Msx2, and Hoxd13 show conserved patterns of expression in osteoblasts (the
bone forming cells), regardless of tissue origin [153, 162-165]. The downstream genetic
networks also show interesting variation across pectoral / pelvic girdles. Deletion of transcription
factors Alx1/4 caused a reduction/absence of the pubis bone formation, with little effect on
pectoral girdle morphology. However, if Tbx15 and Gli3 are also deleted, blade patterning in the
pectoral girdle is completely abolished [166].
One theme to emerge over the past few years is that hindlimbs and genitals share many
aspects of gene expression, to the exclusion of forelimbs. This may be related to the fact that
hindlimbs and external genitalia evolved after forelimbs and thus may have coopted different
genetic machinery [156]. Deleting the crucial homeobox genes HoxD11-13 drastically reduced
the size of both digits and the baculum in mice [167, 168]. Deletion of a conserved HLEB
enhancer region of Tbx4 in mice also caused a size reduction in the pelvic bone and baculum of
male mice [169]. In a fascinating study, Lozovska et al. [170] found that by conditionally
knocking out the Tgfbr1 gene a cascade of chromatin accessibility changes forced male mice to

51
develop an extra set of hindlimbs, in some cases with full bone systems, where their genitals
would have normally developed.
In sum, we have many examples of both conserved and unique developmental genetic
pathways shared across bone systems. However, this variation has not been exhaustively
documented, but rather mostly focused on limb systems. The baculum, a bone found in the penis
of about 85% of mammalian species, is unique in three main ways that make it an ideal model
system to expand our knowledge of how bone and cartilage formation varies across the skeleton.
First, the baculum is extremely morphologically diverse – so much so that mammologists have
used the baculum to distinguish otherwise outwardly identical species [171]. This morphological
variation is genetically determined and probably driven by its impact on reproductive fitness
[172-174]. In other words, the genetics of the baculum may offer insight into the development of
a bone that evolves much more rapidly than the rest of the skeleton. Second, the baculum has
been gained approximately 11 times and lost 10 times across the mammalian phylogeny [175].
This evolutionary pattern allows us to ask an important question about novel bone cell
development such as, do novel bone systems employ similar or different gene pathways than do
other bone systems?
Third – and the main motivation of the current manuscript – is that the baculum requires
androgen to develop [176]. Seminal studies have shown that neonatally castrated male rats failed
to fully form a baculum, but baculum development can be rescued by treatment with exogenous
testosterone via a renal capsule [177, 178]. The baubellum, which is a bone that forms in the
clitoris of some mammal species, is more evolutionarily and developmentally labile than the
baculum, and there are many species that have a baculum but lack a baubellum (and no species
that have a baubellum without a baculum) [179]. But even in species which lack a baubellum –

52
for example, rats – bone formation can be experimentally induced by administering testosterone
[178, 180, 181], suggesting the genetic pathways that lead to bone development in this tissue are
normally latent but can be re-activated. Similar studies have garnered similar results in longtailed weasels. Immature castrated long-tailed weasels failed to develop adult type bacula even
when they were allowed to age way into adulthood [182]. Weasels are Carnivorans which
showcase an independent gain of the baculum from rodents [175] showcasing a possible
conservation in the baculum’s androgen reliance for proper development and maturation. These
studies suggest that for genital bones, androgen signaling is an important regulator of its
development [183].
The requirement of androgen is unusual for bone development; however, it is not unusual
for bone maintenance. Reducing or removing androgen often modifies bone patterning [184],
including digit lengths [185] and pelvic shapes. These studies have shown that androgens give
males an anti-fracture protection while estrogens can in old age negate that [186, 187]. However,
to our knowledge removal of androgen does not result in the wholesale absence of a bone, as it
does in the baculum. The reliance on androgen leads to several critical questions. What genes are
targeted by androgen signaling during baculum formation? How does androgen signaling
translate into downstream effects on gene expression profiles? How does both androgen
signaling and gene expression vary across different bone systems?
We attempt to answer these questions by first characterizing Androgen receptor (Ar)
binding across developing penis, forelimb, and hindlimb tissues in mice, using chromatin
immunoprecipitation sequencing (ChIP-seq) with a reliable antibody for AR. We identify
approximately 9K AR binding sites in developing mouse penis, while we find none in either
forelimb or hindlimb. Interestingly, the 9K AR binding sites in developing mouse penis are

53
enriched near genes with known roles in bone and cartilage development, perhaps revealing the
molecular basis of the baculum’s unique reliance on androgen. To better understand the effect of
androgen signaling on downstream expression, we also sequenced mRNA at single cell
resolution (scRNA-seq) from these same three tissues, followed by computational integration of
the datasets. Most cell types can be found in all three tissues, but chondrocytes – the cartilageforming cells – show striking divergence. We find chondrocytes across all three tissues, but the
chondrocytes found in developing mouse penis form a unique and distinct cluster compared to
those of the forelimb and hindlimb. Ar is one of the most highly up-regulated genes in the penis
chondrocytes, suggesting this important transcription factor might be driving major differences
in gene expression in these unique chondrocytes. Our study links androgen signaling and gene
expression profiles across three important bone systems.
Materials and Methods
3.1 Animals
All protocols and personnel for mice were approved under the University of Southern
California’s Institute for Animal Care and Use Committee, protocol #21218 and for rats under
protocol #20584. All mice were of the C57BL/6J genotype and rats were Crl:CD(SD).
3.2 Determining sampling timepoint for sampling baculum precursors
In order to sample chondrocytes and osteoblasts in early stages of differentiation, we
sought to sample bacula, hindlimbs, and forelimbs at a timepoint where ossification had just
begun. To determine the best development timepoint for the commencement of baculum bone

54
development in laboratory mice, we performed histological staining on developing mouse
penises, ranging from P1 (postnatal day one) through P10 (postnatal day ten). After euthanasia of
male neonates, we fixed tissues in 10% Neutral Buffered Formalin Solution (Sigma-Aldrich
#HT5012) for 3-5 days. After fixation, tissues were placed in a tissue embedding cassette and
transferred to 70% EtOH until tissue infiltration processing. Tissues were then infiltrated with
paraffin wax using the Translational Pathology Core Laboratory (TPCL) at the University of
California, Los Angeles and embedded in Paraplast (McCormick Scientific #39501006) using
metal base molds. Mouse penises were embedded in a vertical longitudinal orientation. Tissues
were cut into thin sections of 5 and 10 microns using Microm HM310 Rotary Microtome at the
USC School of Pharmacy. Sections were then deparaffinized with Xylenes and rehydrated using
graduated amounts of EtOH starting at 100% to 50%. Each developmental stage was stained
with Alcian Blue to observe cartilage development and Alizarin Red S for bone development.
Through these histological samples, we determined that ossification in the mouse penis began
around P6 (Supplementary Figure 1).
Three P6 male mice from two different litters (total n=3) were sampled. Sex was
determined by observing anogenital distance, scrotal coloration, nipple absence, and testes
presence after dissection. Neonatal mice were decapitated using a sterile razor blade, and penises
were dissected in ice cold 1x DPBS for ChIP-seq or DMEM w/o Phenol Red (Hyclone
#SH30284.02) for single-cell RNA sequencing. Dissections were made under a LEICA M165 C
microscope. To eliminate some of the extraneous tissue and increase the chance of sampling
chondrocytes and osteoblasts, we removed prepuce from each penis, focusing on the glans penis
for downstream experiments.

55
For AR ChIP-seq data only, we also sampled P13 (13 days old) male mice penises (n=6)
for comparison. Penis tissue and preparation for ChIP-seq data was prepared the same as above.
Prepuce was removed, and sex was determined via anal genital distance and testes presence. At
this age the baculum is still not fully formed allowing us to compare the AR binding in the
baculum at different developmental stages.
For AR ChIP-seq data only, we collected data from P2 (2 day old) rat penises (n=2). P2
was the timepoint chosen based on studies done by Murakami and Mizuno (1984) [178]. They
observed that bone was well established in P3 (3 days old) rat penis tissue. In lieu of that
information, we chose P2 as the timing for when bone begins forming in rats. Penis tissue was
dissected in the same way as for the P6 mouse penis tissue. Prepuce was removed as well
focusing on the glans penis for the AR ChIP-seq experiments. Sex in rats was confirmed via the
presence of testes after dissection. P2 rat penis tissue was processed for ChIP-seq the same as for
mice tissues.
3.3 Determining sampling timepoint for sampling forelimb and hindlimb precursors
To determine the best development timepoint for sampling forelimbs and hindlimbs of
mice, we consulted Theiler stages. Bone development begins in the long bones of forelimbs and
hindlimbs according to the Theiler stages at E15.0 (embryonic day 15) [188].
To sample forelimb and hindlimb tissue from E15 male mice, one male and one female were
paired for two weeks, then separated. Pregnant female mice were euthanized 15 days after
mating, as judged from the presence of a copulatory plug. Embryos (n=6) from two different
litters were dissected out, and males determined by the presence of testes. If the embryo was
male, the forelimbs and hindlimbs were used for experiments.

56
3.4 Androgen Receptor Chromatin Immunoprecipitation Sequencing (AR ChIP-seq)
We incubated the harvested penises, forelimbs, and hindlimbs in Trypsin-EDTA for 5
minutes at 37°C to loosen up the tissue. We inactivated Trypsin-EDTA using DMEM + 10%
FBS, then cross-linked DNA to bound proteins by adding 1% formaldehyde, incubated on a
horizontal nutator for 20 minutes at room temperature. We quenched the crosslinking reaction
using 2.5 M Glycine, spun down tissues at 4,000 rpm for 5 minutes at 4°C. We washed tissues
twice in ice-cold PBS, then removed the tissue pellet. We sheared chromatin to a size of 100-300
bp in a bath-based sonicator (Diagenode Bioruptor 300).
Of three different antibodies tested, one (AR rabbit polyclonal, Millipore catalog #06-
680; Lot# 2967860) produced robust results. We pre-blocked this antibody in PBS/0.5% BSA
before incubating overnight with 100 µg of sheared chromatin. We added the reaction to a
Protein G Agarose Column (Active Motif, catalog # 53037) and incubated for 3 hours. After
washing, we eluted chromatin bound to AR antibody from the column and reversed protein-DNA
crosslinks by incubating at 65°C overnight. After treating solution with RNase and proteinase K,
we purified deproteinated DNA with MicroChIP DiaPure columns (Diagenode, catalog #
C03040001).
We produced ChIP and input chromatin control libraries using the NEBNext Ultra II
Library Prep Kit (NEB, catalog # E7645S). We performed qPCR to assess enrichment of a
positive control site over a negative control site in AR ChIP-seq libraries compared to the input
library. As a positive control, we focused on intron5 of Fkbp5, which was bound by AR in
multiple tissues in previous studies [118, 189]. As a negative control, we focused on an intronic
region of Stra8 that lacked matches to known AR binding motifs. The primers used in qPCR

57
were as follows: FKBP5-F: 5’- ACCCCCATTTTAATCGGAGAAC-3’; FKBP5-R: 5’-
TTTTGAAGAGCACAGAACACCCT-3’; STRA8control-F: 5’-
GGCAGAAGGGTTCATGTGTT-3’; STRA8control-R: 5’- AAATGCTCTCACTTGGCTTGA3’. We sequenced libraries on the NextSeq 2000 platform at the Georgia Genomics and
Bioinformatics Core. Sequencing reads from the ChIP and input control libraries were aligned to
the mouse genome (mm10) using BOWTIE 2 v2.4.5 with options “--end-to-end --very-sensitive”
[190]. We only kept uniquely aligning reads with fewer than three sequence mismatches to the
reference genome.
We identified AR peaks using MACS2 v2.7.1 [191] with the options “--bdg --SPMR --
qvalue=0.01” and generated fold-enrichment over input control. We performed de novo motif
analysis from sequences within 50 bp of AR peak summits, using HOMER
(http://homer.ucsd.edu/), with 100k random genomic sequences as the background. HOMER
uses hypergeometric optimization to detect short sequence motifs that are enriched in sample vs.
control [192]. To test enrichment of Gene Ontology annotations with associated AR ChIP-seq
peaks, we used the R package CHIPENRICH [193]. In our case, this method assigns AR peaks
based on the nearest transcription start site found in the mouse reference genome (mm10), and
test whether the functions associated with those genes are significantly enriched or depleted
compared to random expectations.
3. 5 Single-cell dissociation and Library construction
3.5.1 Baculum Dissociation

58
We dissociated tissues using the gentleMACS Octo Dissociator using the soft tumor
program 37C_m_TDK1. We placed the cell suspension in a MACS SmartStrainer (70 µM)
placed in a 15 mL conical tube, then washed it in 10 mL of DMEM. We then centrifuged the
sample at 300 G for 7 min. at 4°C, removed the supernatant, and resuspended cells in 100 µL of
PBS w/ 0.04% BSA. We processed cell suspensions with a 10X Single Cell 3’ v3.1 kit, aiming to
isolate ~10,000 cells, then prepared libraries according to the 10X Genomics protocol, and
sequenced using a 200-cycle P3 NextSeq 2000 run at the Molecular Genomics Core (MGC) at
USC Norris Comprehensive Cancer Center, Keck School of Medicine USC aiming for 1.2 billion
reads per sample.
3.5.2 Forelimb and Hindlimb Dissociation
We randomly placed forelimbs and hindlimbs from 6 animals into two Eppendorf tubes
filled with ice cold DMEM. We then dissociated cells on a gentleMACS Octo Dissociator.
Dissociation, filtration, processing, and library prep followed the above baculum protocol.
3.5.3 Read Mapping
We performed alignment, filtering, barcode counting, and Unique Molecular Identifier
(UMI) counting using the function cellranger count in the Cell Ranger v7.0.0 software created
by 10X Genomics, all aligned to the mm10 version of the mouse genome (NCBI accession
number: GCA_000001635.2).
3.5.4 Seurat v5.1.0 Quality Control Metrics and Normalization

59
Using Seurat (v 5.1.0) we did quality control on our samples. For the P6 mouse baculum
tissue sample, excluded cells with fewer than 50 reads as well as cells that had more than 5%
reads derived from mitochondria as these reads are indicators of stressed or dying cells. For the
P6 mouse penis tissue samples cells we kept cells that had more than 100 but less than 11,000
genes detected. For the E15 mouse hindlimb tissue samples, we kept cells that had more than 100
but less than 9500 genes detected. For the E15 mouse forelimb tissue samples, we kept cells that
had greater than 100 and less than 9000 genes detected and less than 10% mitochondrial reads.
All three samples were normalized using the SCTransform function from the sctransform v2
package. Within the SCTransform function we implemented the “glmGamPoi” method and
regressed out the mitochondrial percentage. We performed cell clustering and dimensional
reduction using the Seurat package in R, in combination with custom scripts. We ran a principal
component analysis on the SCTransformed data using the RunPCA function, using the 3000 most
variably expressed genes. We clustered cells using the FindClusters function using 50 Principal
Components at a resolution of 1. We used the RunUMAP function to project the data into the
two-dimensional space of a uniform manifold approximation and projection graph (UMAP) for
visualization.
3.5.5 Comparing Chondrocytes Between Developing P6 Baculum and Developing E15 Hindlimb
and Forelimb Bone Systems.
To compare the gene expression profiles of cell types from the three tissues (mouse penis,
forelimbs, and hindlimbs) we took a data integration approach via the Seurat package (v. 5.1.0)
[194]. Integrating data is a strategy whereby we are able to merge multiple datasets to better
analyze shared cell types. Integration uses shared sources of variance between the different
datasets to create UMAP space where shared and not shared cell types can be analyzed. We

60
integrated the data across tissue type (penis, forelimb, hindlimb) using a custom R script and the
IntegrateData function in Seurat (v 5.1.0) [194]. To further compare the gene expression profiles
of different cell types in the integrated data sets, we used the FindAllMarkers function in the
Seurat package (v 5.1.0) for each of the integrated clusters. We output the top 20 marker genes
using the following parameters: an adjust p-value < 0.01, log2 fold change greater than 1,
percentage of cells in the first group that detected said gene greater than 50%, and percentage of
cells in the second group that detect said gene less than 10%. Using those marker genes, we were
able to identify cell types via primary literature research and secondarily via ChatGPT (v. GPT3.5).
Results
3.6 Inferred ARE’s cluster near genes involved in bone and cartilage formation, but only in
developing mouse and rat penises.
We identified a total of 9,681 unique peaks across two biological replicates of P6 mouse
penises. From two P6 mouse penis replicates, we generated 31,804,830 and 41,150,993 total
reads, with 25,534,883 and 32,551,378 uniquely aligning to the reference genome, respectively.
The corresponding control libraries yielded 46,942,969 and 54,275,840 total reads, with
36,655,357 and 42,073,083 uniquely aligning to the genome, respectively. We observed 88-fold
enrichment at the positive control Fkbp5 site and no enrichment at the negative control site Stra8
locus [118, 189], providing confidence that our ChIP-seq data identified bona fide AR binding
sites (Figure 2.1).
Using HOMER (http://homer.ucsd.edu/), we identified de novo motif enrichment on the
sequences within 50 bp of the AR peak summits. The top three hits were for ARE and
glucocorticoid binding elements (GRE) (Figure 2.1). The canonical ARE and GRE are very

61
similar in sequence, it seems likely that what HOMER identified as GRE enrichment is probably
ARE enrichment. The alternative – that our AR ChIP-seq has actually targeted and enriched for
GR bound to DNA, seems less likely. Our results suggest that we have identified bona fide AR
binding sites.
Of 24,001 protein coding genes annotated in mm10, 4,604 were associated with at least one AR
peak. Gene Ontology analyses revealed 11 terms related to osteo- or chondrogenesis were
significantly over-represented among these genes (Table 2.1). 616 of the 9,681 AR ChIP-seq
peaks occurred near a gene associated with functions related to osteo- or chondrogenesis.
Assigning AR peaks to a gene is somewhat subjective. Many studies have assumed that
AR binding sites within 50 kb of a transcription start site modify that gene’s expression. But
other studies have shown that AR binding can influence the expression of genes hundreds of kb
away [195-199]. Therefore, we repeated the analysis using different cutoffs between each peak
and their closest transcription start site. We reran CHIPENRICH with a cutoff of 113 kb (75%
quantile) as well as a peak score that was greater than 181 (25% quantile) where score is -
10*log10qvalue, as calculated by MACS2 [191]. This analysis returned 5,476 AR ChIP-seq
peaks which is nearly half of our original number of AR ChIP-seq peaks. All terms in Table 1
remain significantly enriched. We explored an even more stringent cutoff and only included
peaks that were associated with genes that had at least two peaks, reducing our analysis to 3,757
AR ChIP-seq peaks. All terms remained significantly enriched except for GO:0005201
(extracellular matrix structural constituent). In short, the enrichment genes associated with osteoand chondrogenesis was robust to cutoffs.
We also collected AR ChIP-seq data from P13 day old mouse penises, when the baculum
is much further along in development. We generated 34 million reads from the control library

62
and 29 million reads from the AR ChIP-seq library. We observed 64-fold enrichment against the
known AR binding site, Fkbp5 [118, 189], validating the effectiveness of our antibody. As with
the P6 mouse data, the top de novo motif observed (ACABNNTGTTCY) was very similar to the
canonical Androgen Response Element (ARE) motif (AGAACANNNTGTTCT). We identified
3,893 AR ChIP-seq peaks and performed several analysis to confirm true positives. GO
enrichment analyses uncovered almost identical results as those found in mouse P6 and P13 data.
The fact that AR binding is still apparent after 7 more days of development suggests that
androgen signaling during baculum development is a sustained process and could explain why
experimental or genetic manipulation of androgen has such drastic effects on baculum
morphology.
We also compared the inferred ARE’s from mice to those from P2 rat penises for an
interspecific comparison. For input (control) libraries, we generated a total of 44,528,151
(49,294,048) input reads, with 39,161,658 (43,158,380) uniquely aligning to the rat reference
genome. Interestingly, the same GO term enrichment results emerged. The most recent common
ancestor of these two rodent species existed 8-14 million years ago [200].
To compare bone development across different bone systems, we repeated AR ChIP-seq
on E15 mouse hindlimbs. The two input E15 hindlimb libraries yielded 34,460,578 and
33,743,452 total reads, of which 28,599,585 and 29,043,883 aligned uniquely. The two control
E15 hindlimb libraries yielded 29,959,621 and 38,370,600, of which 25,731,006 and 33,242,439
uniquely aligned. We found zero AR ChIP-seq peaks in the developing E15 mouse forelimbs and
hindlimbs, across two biological replicates each. This might explain why experimental or genetic
manipulation of androgen signaling does not influence the skeletal elements as dramatically as
the baculum.

63
In sum, AR ChIP-seq data strongly suggested we identified bona fide AR binding sites, as
all datasets showed significant enrichment over our positive control site (Fkbp5) and no
enrichment over our negative control site (Stra8). In addition, de novo motif searching showed
enrichment of sequence nearly identical to the canonical AR binding motif. For three datasets –
mouse P6, mouse P13, and rat P2 penises, GO enrichment showed that AR tended to bind near
genes with known influence on bone and cartilage development, regardless of cutoffs. This
enrichment was not observed in mouse E15 forelimbs and hindlimbs, suggesting that androgen
signaling is a unique aspect of baculum formation.
3.7 Analysis of single-cell RNA-sequencing of developing P6 mouse penis and E15 mouse
hindlimb and forelimb via Integration
To compare gene expression profiles between developing bone systems (P6 penis, E15
hindlimb, and E15 forelimb) we used single-cell RNA-sequencing (scRNA-seq). scRNA-seq
allows us to computationally dissect different cell types based on genes that mark the cell types
sequenced. After QC and filtering, the P6 mouse penis sequencing data analyzed included
247,746,250 reads from 19,576 cells averaging 9,606 reads per cell. The E15 hindlimbs
consisted of 148,734,246 reads from 20,511 cells averaging 7,251 reads per cell; E15 forelimb
data consisted of 142,700,891 reads from 22, 055 cells averaging 6,470 reads per cell. Using the
Seurat package (v.5.3.0) in R, we integrated the three datasets to further analyze the overlap of
cell types from the three datasets. After integration, we identified 40 cell clusters. Using the
FindMarkers function in the Seurat package, we collapsed these into 10 cell types (Figure 2.2A).
The majority of cells from all three datasets were identified as mesenchyme, an intuitive cell type
as these tissues will be actively differentiating.

64
Of most interest to the current study, we also identified chondrocytes (cartilage-forming
cells) from all three tissues, suggesting we identified precursors to the baculum and limb bones.
We note that the proximal end of the baculum forms from endochondral ossification, as limb
bones do [201]. However, the chondrocytes from penis fell into a distinct cluster compared to
the chondrocytes from forelimbs+hindlimbs. We identified differentially expressed genes
between the two groups of chondrocytes (forelimb+hindlimb versus penis) using the
FindMarkers() function in Seurat. We identified 139 (48) over- (under-) expressed genes in the
penis chondrocytes compared to the forelimb+hindlimb chondrocytes. Up-regulated genes were
significantly differentially regulated, as judged from Wilcoxon Rank Sum tests implemented in
the FindMarkers function, with a Benjamini-Hochberg-corrected p-value of <0.01. Up- (down-)
regulated genes further showed a log2 fold change of more (less) than 1 (-1), expression in more
(less) than 50% of penis chondrocytes, and expression in less (more) than 50% of FL+HL
chondrocytes. The second-most upregulated gene in penis chondrocytes vs. FL+HL chondrocytes
was Ar. Furthermore, 79 of 139 up-regulated genes showed at least one AR peak in our ChIP-seq
data collected from developing penis, significantly more than expected from the genome, where
4,604 of 25,190 genes had an AR peak (2=133.8, df = 1, p-value < 2.2e-16). Recall that no AR
peaks were detected in developing hindlimb, so these genes offer a potential link between bone
formation and androgen signaling in the baculum.
In addition to Ar, some of the top up-regulated genes in penis chondrocytes compared to
forelimb+hindlimb chondrocytes included Sox9, Sox5, and Sox6, all genes that are important in
early differentiation of chondrocytes. Interestingly, two classic markers of later-differentiated
chondrocytes, Col2a1, Runx2 were not up-regulated in penis chondrocytes. Therefore, it is
possible that one reason penis chondrocytes are distinct from forelimb+hindlimb chondrocytes is

65
that they are in a different state of differentiation. We deem this explanation unlikely, as all
tissues were sampled at a developmental timepoint where bone has just begun to form. In other
words, the developmental ages of these three tissues should be relatively equivalent. In addition,
the forelimb and hindlimb samples should surely include early-differentiating chondrocytes.
Although the gene expression differences between penis chondrocytes versus forelimb+hindlimb
chondrocytes could be partly driven by variation in developmental timing, at least some of this
difference seems to be due to inherently different gene expression profiles of the two groups of
cells. This is especially interesting given that these differentially regulated genes observed in our
scRNA-seq data are significantly more likely to have at least one AR peak detected in our ChIPseq data.
Discussion
Deciphering the molecular mechanisms of bone development is an important undertaking
towards understanding the evolution of morphological diversity. The baculum is unusual in that
it requires androgen to develop, making it unlike any other bones where sex hormones are not
required. Our study sheds light on the molecular mechanisms that underlie the development of
this unusual bone. We have shown that AR binding sites are abundant in developing penis and
are enriched near genes that influence bone and cartilage formation. In contrast, we detected no
AR binding sites from developing forelimbs and hindlimbs. Concomitant with this difference in
AR binding sites between the tissues, the chondrocytes identified from the penis were very
distinct from the chondrocytes in forelimbs and hindlimbs. These clusters separated by
differential regulation of hundreds of genes, including Ar itself. For over half of these
differentially expressed genes (79/139), we detected an AR peak in our ChIP-seq data, a
significantly higher number than expected from the genomic background.

66
Some genes that were both differentially regulated between the two chondrocyte groups
and had at least one AR peak nearby and are involved in matrix development, remodeling
processes, and tissue differentiation during bone development (Adamts12, Aspn, Fbn1, Lox,
Lrp5, Slc8a1) [202-207]. Our study indicates that some of these fundamental molecular
processes are regulated by androgen signaling.
Some of these differentially expressed genes were “classic” chondrocyte genes. For
example, Sox6, Sox9, Col2a1, and Runx2 are all upregulated early in chondrocyte development
[162, 208, 209], and all were identified as chondrocyte markers in our data (Table 2.2).
However, many of these classic genes were also significantly differentially expressed in the two
subsets of chondrocytes (in some cases not enough to meet our threshold as a marker
distinguishing the two chondrocyte clusters).
Taken together, it appears that we have identified the precursors to both baculum and
limb bones from our single cell RNA-seq data, but that the chondrocytes from penis have
diverged in their expression profiles compared to forelimb+hindlimb chondrocytes. An
alternative hypothesis is that we are simply sampling chondrocytes of different developmental
ages across these tissues. Within developing tissues, cell types are shuffling throughout different
stages, and it is possible that the two chondrocyte clusters observed here simply reflect different
stages along a continuum of differentiation, rather than being unique in their gene expression per
se. This does not seem a likely alternative, for two reasons. First, we sampled each tissue at a
developmental timepoint right before bone begins to form. Thus, each tissue was sampled at a
different chronological age to standardize their developmental age. Second, because all three
tissues were sampled at a relatively early stage, before bone has begun to develop, all tissues
should have had “young chondrocytes”, even if some tissues were developmentally older than

67
others. In that case, we would expect one tissue’s chondrocytes to be a subset of another tissue’s
chondrocytes. In other words, the “young chondrocytes” from all tissues should still cluster
together. But this is not what we observe – instead, we observe two distinct clusters of
chondrocytes, separated by differential expression of hundreds of genes.
Studies have shown that estrogens and androgens allow for the proper maintenance of
bone and cartilage during puberty and throughout an organism’s lifespan. Losing one or both of
these hormones can lead to defects in bone remodeling, osteoblastogenesis and
osteoclastogenesis [186, 187]. Previous studies documented the requirement of androgen
signaling for baculum development [177, 178, 180, 181]; our study uncovers the first molecular
links between those observations and the underlying genetic regulation responsible for baculum
formation.
In addition to its requirement for androgen, the baculum is unusual in that it has arisen
multiple times during mammalian evolution. Our study offers the first glimpse into the genetic
networks that are deployed during the formation of a novel bone in the baculum. Studies on
multiple bone systems indicate that gene pathways used to make bone/cartilage share a lot of
similarities [162, 210-215]. The similarity in gene expression of typical cartilage genes between
independently evolved bone systems (baculum and hindlimb/forelimb bones) that we observe in
our data suggests that the evolution of cartilage has been able to maintain several of the major
genes in cartilage development, however our data suggests some difference in the evolution of
baculum chondrocytes that differs from forelimb+hindlimb chondrocytes.
We expected to find AR peaks in the developing forelimb and hindlimb. Zheng and Cohn
[185] found that there is a small window of time during digit development that is regulated by
androgen and estrogen signaling. Disrupting androgen and estrogen signaling at timepoints

68
E12.5 and E15.5 in mice causes a disruption in digit length ratio of mouse hindlimbs that they
did observe until E17 in mice hindlimbs showcasing important timepoints for proper androgen
and estrogen regulation. Our hindlimb data came from E15 individuals, which is very close to
these developmental timepoints that showed an influence of androgen. It is possible that we
didn’t observe any AR binding via ChIP-seq because our timepoint was earlier than when digits
are fully forming [185]. Our data suggests that the initiation of bone formation in the mouse
hindlimb might not require androgen signaling while the initiation of the baculum might require
androgen signaling, at least in early chondrocytes.
Taken together, our study offers the first insights into how an otherwise conserved bone
developmental program can vary in the baculum, where androgen signaling is required. Future
work should focus on the link between androgen signaling and bone development, which seems
to be important for the baculum but not the forelimbs or hindlimbs. Ultimately, our study offers a
launching pad to experimentally manipulate candidate genes, to mechanistically validate their
role in bone development. The strong expression of Ar, and the large number of AR binding sites
near genes involved in bone and cartilage development, suggest that we may be able to disrupt
the baculum’s formation through manipulation of androgen signaling.
Limitations of the study
We employed scRNA-seq data because we expected the cells of interest (osteoblasts and
chondrocytes) to be rare, at least in the developing mouse penis. Our strategy was to
computationally dissect these cells and compare them to similar cells in other tissue systems.
However, the read coverage per cell is still small relative to the total number of RNA molecules
expected to be in a cell. Because scRNA-seq can be highly variable source of data, we cannot

69
rule out that more subtle differences could have been detected with more coverage. Nevertheless,
the coverage was high enough to detect major differences among chondrocytes.
Any developmental study is potentially compromised by uncertainty in sampling the
correct developmental timepoint. We collected our samples at a standard timepoint right before
bone formation in these three different systems. Nevertheless, it is possible that we would have
observed a single chondrocyte cluster – instead of two – if we had sampled our tissues at
different timepoints.
Acknowledgements
We would like to thank the UCLA Translational Pathology Core Laboratory and USC
School of Pharmacy Histology Core for processing our histology tissue. We want to thank Dr.
Douglas Menke for his help with AR ChIP-seq. All AR ChIP-seq was performed in the lab of Dr.
Douglas Menke, Department of Genetics, University of Georgia, by Sungdae Park. We would
like to thank Dr. John Carpten for letting us use his Miltenyi Octo-Dissociator when we needed
to dissociate cells. All scRNA-seq data was generated through the MGC at USC Norris
Comprehensive Cancer Center. We would like to thank Stephanie Tring for her help at the MGC
more specifically. Special thanks to Ricardo Chavez and Beah Tolentino for helping us track
down animal documentation.

70
Figure Legends
Figure 2.1
A) The top four enriched motifs found among manually curated transcription factor binding sites.
The top three hits include known motifs for Androgen Receptor Element (ARE) and
Glucocoriticoid Receptor Element (GRE) which is known to share a high degree of sequence
similarity. B) Reads mapped from either Control or AR ChIP-seq libraries, with associated AR
peak calls. Left panel) After ChIP-seq, we mapped 64 times as many reads to an intron of Fkbp5,
an AR-responsive gene known to be bound by AR. Right panel) We identified an AR binding
region in Bmpr1b – a gene expressed in early digit formation.
Figure 2.2
A) Integrated Data of three tissue types. Note that chondrocytes from MP6 form a distinct cluster
compared to chondrocytes from FL+HL. MP6=mouse penis from P6; FL=forelimb from E15
mouse; HL=hindlimb from E15. B-E) Gene expression FeaturePlots of Chondrocyte specific
genes: B) Col2a1 C) Runx2 D) Sox9 E) Sox5
Supplementary Figure 2.1
Alizaren S. Red bone staining for bone to corroborate the stage at which the baculum in male
mice begin to develop bone – 6 days after birth.

71
Table 2.1. Significantly over-represented Gene Ontology (GO) terms associated with
bone development, identified with CHIPENRICH. Rank=ranked by FDR. FDR=False
Discovery Rate [216]. N.Genes=number of genes in the genome with that GO
annotation. N.with.Peaks=number of those genes that have at least one AR peak.
Odds.Ratio=Increased likelihood that an AR peak is associated with a GO annotation,
compared to random. BP=Biological Process. CC=Cellular Component. MF=Molecular
Function. Numbers in parentheses indicate total number of significantly enriched terms
identified. Every term in this table was also identified after repeating the analysis in rat
penis (see text).
Rank Description FDR N.Ge
nes
N.with.P
eaks
Odds.R
atio
GO Biological Process (151 significant terms
total):
31 extracellular matrix organization
(GO:0030198) 0.001 200 76 2.13
34 extracellular structure organization
(GO:0043062) 0.001 201 76 2.11
42 positive regulation of ossification
(GO:0045778) 0.003 96 49 2.81
74 cartilage development (GO:0051216) 0.013 179 77 1.92
79 regulation of ossification (GO:0030278) 0.014 218 82 1.84
89 pos. regulation of osteoblast
differentiation (GO:0045669) 0.017 67 33 2.87
97 ossification (GO:0001503) 0.021 390 133 1.57
GO Molecular Function (10 significant terms
total):
4
extracellular matrix structural constituent
(GO:0005201) 0.015 43 23 3.91
GO Cellular Component (9 significant terms
total):
1 extracellular matrix (GO:0031012) 1.00E04 498 167 1.8
3
proteinaceous extracellular matrix
(GO:0005578)
1.00E04 343 139 1.93
8 extracellular space (GO:0005615) 0.005 1427 347 1.36
Table 2.1

72
Figure 2.1
A
B

73
ABC

74
Figure 2.2
D
E

75
Table 2.2
Top 10 Marker genes used to identify the 10 different cell types found in our 3 datasets (P6 Baculum, E15
Hindlimb, E15 Forelimb). Marker genes were identified using the following parameters: Up- (down-) regulated
genes further showed a log2 fold change of more (less) than 1, expression in more (less) than 50% of penis
chondrocytes, and expression in less (more) than 50% of FL+HL chondrocytes
Upregulated Downregulated
gene avg_log2FC pct.1 pct.2 p_val_adj gene avg_log2FC pct.1 pct.2 p_val_adj
Mesenchyme Mesenchyme
Twist1 2 0.7 0.29 0 Sox6 -2.21 0.1 0.36 0
Il11ra1 1.89 0.67 0.26 0 Chd7 -2.93 0.05 0.29 0
Mfap2 1.45 0.94 0.54 0 Msi2 -1.38 0.34 0.58 0
Cdh11 1.38 0.84 0.45 0 Itga6 -2.57 0.08 0.3 0
Tmem132c 3.06 0.47 0.09 0 Cux1 -1.2 0.28 0.49 0
Pdgfra 1.71 0.57 0.19 0 Stard13 -2.35 0.13 0.33 0
Dcn 2.57 0.79 0.41 0 Elmo1 -2.21 0.15 0.35 0
Prrx1 1.17 0.74 0.37 0 Sox5 -1.91 0.17 0.37 0
Col6a2 1.17 0.79 0.43 0 Lcp1 -5.66 0.01 0.2 0
Col5a2 1.14 0.91 0.55 0 Tspo -1.82 0.14 0.33 0
Keratinocyte Keratinocyte
Krt15 7.79 0.83 0.06 0 Vim -2.34 0.57 0.97 0
Krt5 7.57 0.78 0.04 0 Lgals1 -2.35 0.58 0.96 0
Lgals7 6.47 0.86 0.13 0 Fbn2 -2.53 0.43 0.81 0
Sfn 6.9 0.75 0.02 0 Fstl1 -2.28 0.5 0.85 0
Krt14 7.63 0.74 0.04 0 Mfap2 -2.25 0.46 0.8 0
S100a14 7.67 0.7 0.01 0 Grb10 -1.83 0.5 0.83 0
Perp 6.85 0.69 0.02 0 Tcf4 -2.23 0.57 0.88 0
Fxyd3 6.96 0.63 0.02 0 Col1a2 -2.6 0.58 0.89 0
Dsp 6.73 0.62 0.02 0 Rrbp1 -1.82 0.61 0.91 0
Krtdap 6.5 0.59 0.04 0 Col3a1 -2.66 0.61 0.9 0
Tenocyte Tenocyte
Col11a1 2.08 0.95 0.35 0 Celf2 -2.43 0.25 0.68 0
Tnmd 6.53 0.62 0.03 0 Zeb2 -2.29 0.22 0.62 0
Scx 4.84 0.63 0.05 0 Akap12 -2.15 0.19 0.42 0
Fmod 3.2 0.62 0.11 0 Junb -1.06 0.58 0.76 0
Mkx 3.1 0.56 0.09 0 Tmem132c -2.34 0.12 0.34 0
Thbs2 1.98 0.69 0.23 0 Nrp1 -1.83 0.19 0.41 0
Kera 3.22 0.47 0.08 0 Klf4 -1.15 0.44 0.64 0
Zfp185 3.72 0.43 0.05 0 Osr2 -3.24 0.03 0.23 0
Col12a1 1.68 0.92 0.55 0 Vcan -1.2 0.53 0.69 0
Khdrbs2 3.61 0.26 0.03 0 Fyn -1.32 0.26 0.48 0

76
White blood White blood
Tyrobp 5.76 0.91 0.04
0 Nedd4
-2.34 0.42 0.98
0
Fcer1g 5.32 0.92 0.05
0 Rbms3
-2.68 0.29 0.84
0
Lyz2 9.17 0.82 0.02
0 Mdk
-2.88 0.33 0.88
0
Laptm5 5.32 0.82 0.03
0 Grb10
-2.52 0.3 0.85
0
Lst1 5.57 0.81 0.05
0 Auts2
-2.71 0.39 0.94
0
Ptpn18 4.79 0.81 0.06
0 Mfap2
-2.82 0.27 0.81
0
Ms4a6c 8.37 0.75
0
0 Nfib
-2.52 0.35 0.89
0
Spi1 6.17 0.77 0.02
0 Fstl1
-2.74 0.33 0.86
0
Coro1a 5.17 0.78 0.04
0 Serpinh1
-2.47 0.39 0.93
0
Csf1r 6.34 0.77 0.03
0 Fbn2
-2.76 0.3 0.82
0
Chondrocytes Chondrocytes
Sox5 3.42 0.86 0.17
0 Celf2
-2.49 0.3 0.71
0
Sox6 3.48 0.75 0.14
0 Zeb2
-1.59 0.35 0.65
0
Col9a1 4.61 0.63 0.08
0 Tmem132c
-2.98 0.08 0.36
0
Col2a1 5.01 0.66 0.16
0 Lama2
-2.12 0.14 0.42
0
Col9a3 4.41 0.62 0.12
0 Nxn
-1.56 0.27 0.55
0
Sox9 2.87 0.7 0.2
0 Igf1
-2.18 0.2 0.46
0
Mia 4.3 0.54 0.06
0 Dpysl3
-1.64 0.26 0.52
0
Kcnq5 2.89 0.61 0.16
0 Ank3
-1.78 0.18 0.44
0
Col9a2 4.19 0.54 0.09
0 Rbp1
-1.69 0.17 0.43
0
Acan 4.65 0.48 0.05
0 Kctd12
-1.36 0.26 0.52
0
Muscle Muscle
Tnnt1 5.62 0.69 0.06
0 Col1a1
-2.93 0.42 0.9
0
Acta2 5.74 0.67 0.06
0 Plpp3
-2.83 0.18 0.66
0
Vgll2 6.39 0.57 0.01
0 Cdh11
-2.39 0.26 0.73
0
Rbm24 4.96 0.58 0.04
0 Mfap2
-2.34 0.36 0.82
0
Tpm2 3.34 0.85 0.33
0 Ifitm2
-1.88 0.48 0.95
0
Pdgfa 2.88 0.67 0.15
0 Col5a2
-1.92 0.36 0.81
0
Plxna2 2.77 0.64 0.14
0 Dcn
-3.47 0.23 0.68
0
Eya4 2.25 0.67 0.16
0 Selenom
-2.09 0.23 0.67
0
Tnik 3.31 0.61 0.11
0 Igfbp4
-2.41 0.35 0.78
0
Actc1 6.46 0.55 0.06
0 Fstl1
-1.68 0.44 0.87
0
Endothelial Endothelial
Cdh5 7.37 0.78 0.01
0 Meg3
-2.79 0.43 0.91
0
Emcn 7.79 0.77 0.01
0 Mdk
-2.3 0.41 0.88
0
Flt1 5.95 0.79 0.05
0 Gpc3
-2.71 0.32 0.78
0
Cd93 7.25 0.73 0.01
0 Fbn2
-2.68 0.37 0.82
0
Egfl7 5.28 0.77 0.07
0 Rian
-2.64 0.23 0.68
0
Plxnd1 5.28 0.77 0.07
0 Ptprd
-2.67 0.35 0.78
0
Esam 6.56 0.71 0.02
0 Gas1
-2.29 0.32 0.74
0

77
Cldn5 6.48 0.69 0.01
0 Lrp1
-2.76 0.11 0.53
0
Ecscr 5.6 0.7 0.03
0 Col6a1
-2.52 0.31 0.73
0
Ctla2a 5.81 0.7 0.04
0 Adgrl3
-3.07 0.25 0.66
0
Immune Immune
Arhgap15 5.62 0.96 0.05
0 Serpinh1
-3.15 0.23 0.92
0
Srgn 6.22 0.99 0.09
0 Nfib
-3.17 0.22 0.89
0
Rac2 5.51 0.92 0.03
0 Sparc
-3.8 0.26 0.93
0
Laptm5 4.19 0.88 0.04
0 Fstl1
-3.4 0.2 0.86
0
Coro1a 4.23 0.88 0.04
0 Mdk
-3.48 0.22 0.88
0
Fcer1g 4.21 0.9 0.06
0 Mfap2
-3.43 0.16 0.81
0
Samsn1 6.1 0.85 0.02
0 Nfia
-3.01 0.23 0.88
0
Dock2 5.02 0.86 0.03
0 Marcks
-2.81 0.25 0.9
0
Tyrobp 3.77 0.88 0.05
0 Meg3
-2.95 0.27 0.9
0
Arhgdib 4.05 0.96 0.15
0 Tcf4
-2.8 0.26 0.89
0
Glial Glial
Plp1 8.28 0.93 0.02
0 Ifitm2
-2.17 0.54 0.92
0
Sox10 7.25 0.82 0.01
0 Igf2
-3.77 0.16 0.74
0
Erbb3 5.34 0.82 0.03
0 Ptprd
-2.74 0.25 0.77
0
Zfp536 4.19 0.91 0.14
0 Igfbp4
-2.75 0.26 0.76
0
Fign 4.33 0.87 0.12
0 H19
-3.19 0.24 0.72
0
Ednrb 6.5 0.78 0.03
0 Gpc3
-2.47 0.31 0.77
0
Gpm6b 3.89 0.92 0.17
0 Fbn2
-2.32 0.4 0.81
0
Foxd3 9.2 0.75
0
0 Gas1
-2.69 0.25 0.74
0
Cdh19 8.96 0.75
0
0 Dlk1
-3.62 0.15 0.66
0
Mpz 11.24 0.74 0.01
0 Ebf1
-3.18 0.18 0.68
0
Red Blood Red Blood
Hba
-a2 7.48
1 0.42
0 Zbtb20
-1.11 0.57 0.67
0
Alas2 6.39 0.56 0.02
0 Nfkb1
-1.46 0.19 0.35
0
Hba
-a1 7.61
1 0.46
0 Iqgap1
-1.01 0.25 0.41
0
Hbb
-bt 7.4 0.99 0.47
0 Fnbp1
-1.02 0.16 0.29
0
Snca 5.61 0.52 0.02
0 Nfe2l2
-1.15 0.17 0.3
0
Mkrn1 3.5 0.58 0.14
0 Sat1
-1.06 0.19 0.31
0
Slc25a37 3.99 0.47 0.06
0 Ankrd44
-1.07 0.17 0.3
0
Bpgm 3.96 0.45 0.08
0 Osbpl8
-1.1 0.26 0.39
0
Ube2l6 4.08 0.31 0.03
0 Cgnl1
-1.1 0.14 0.26
0
Tent5c 2.85 0.36 0.09
0 Ar
-1.15 0.16 0.28
0

78
Supplementary Figure 2.1
PROXIMAL GLANS DISTAL

79
Chapter 4: Conclusion
Introduction
The second chapter of my thesis involved understanding the development of the bone/cartilage
forming cells in the baculum which naturally leads us to wanting to disrupt the baculum’s
development in efforts to better understand the function of this morphologically diverse bone.
While there is a consensus that the baculum is a sexually selected trait (residual testes mass is
significantly correlated with baculum presence [217]), a mechanical function for the baculum has
been difficult to pinpoint because function of the baculum does not seem to be equal across all
mammals that have a baculum. Several hypothesis have been presented for the function of the
baculum - 1) the baculum protects the urethra, 2) the baculum functions in the context of sperm
competition, 3) the baculum functions to stimulate the female’s ovulation timing, 4) the baculum
signals male quality, and 5) the baculum provides rigidity and support for the penis during
intromission.
The complexity of pinpointing the function of the baculum begins with the first hypothesis
mentioned above. Brassey et al. [92] used finite element analysis (FEA) [218] simulations to
estimate deformation, strain and stress on the bacula of 74 Carnivorans when subjected to load.
Their analysis suggested a role for the baculum in reducing malformations in the urethra. Similar
results in Chiropterans, Carnivorans, and Primates showed that the baculum protects the urethra
during copulation, but not in all Chiropteran and Carnivoran species creating a contradiction for
the first hypothesis [219-224]. To address the second hypothesis, Stockley et al. [225] found that
the size (width) of the baculum explained much of the variance in reproductive success for the

80
house mouse. In fact, several studies in mice have come to a similar conclusion where the shape,
relative thickness, resulted in higher paternity but only in situations where competition was
present suggesting an important role in reproductive success for the baculum [30, 172, 173, 226].
However, traditional length and width measurements is not enough for the understanding the
critical role of the baculum. The baculum is morphologically hyper diverse which lead Brassey et
al. [92] to use (FEA) to quantify 3D performance of the baculum in Carnivorans. They found a
significant correlation between the baculum’s robustness under dorsoventral bending and
intromission duration supporting the idea that longer more complex bacula should occur in
mammals with prolonged intromission. Of course, this doesn’t seem to hold for every
Carnivoran. Studies have also shown that baculum length negatively correlates with sexual size
dimorphism in pinnipeds but not in Carnivorans meaning that in some animals the baculum is
influenced/influences sperm competition while in others it does not [30, 227]. Studies in some
Chiropterans, Primates, and Carnivorans contradict this hypothesis as well [224, 227, 228]. The
third hypothesis is important because it considers the female’s point of view and raises questions
of possible conflict. Studies in mice have found that species with induced ovulation have longer
bacula allowing the baculum to stimulate the female, but this does not hold true for carnivorans
[223, 224, 229]. The fourth hypothesis suggests that females would choose males with high
quality bacula leading to high quality bacula displaying positive allometry. This holds true for
muskrats and seals [230-232] but not for martens, mice, and bats [233-235]. A fifth hypothesis
has been proposed suggesting that the baculum functions as a support mechanism for
intromission in general. Artificial erections via 10% formalin in bats suggest that the baculum
and the corpus cavernosa (CC) work in consort to provide support for the penile shaft and distal
tip [219]. The hypotheses for the function of the baculum have just as many supports as they do

81
contradictions making the function of the baculum a conundrum. Altogether, these studies bring
up several hypothesis for the functional role of the baculum which needs to be further
investigated, but to do so, we need to “break” the baculum.
The disruption of the baculum requires two puzzle pieces of information, gene pathways and
how they are regulated. To uncover the first part of the puzzle, we have to consider how the
baculum is made. The baculum is made using the two known bone forming techniques,
endochondral bone formation and intramembranous bone formation. Endochondral ossification
requires a cartilage scaffold to be made prior to bone deposition of the skeleton. On the other
hand, intramembranous ossification does not rely on a cartilage scaffold. Rather, mesenchymal
stem cells (MSCs) differentiate into osteoblasts (bone forming cells) and bone is made [210].
The proximal base of the baculum is made using endochondral ossification while the distal tip is
made using intramembranous ossification [176-178].
Bone formation in any capacity requires a complex pathway of genes, the most important being
Runx2. Runx2 is a transcription factor well known to influence the differentiation of
mesenchymal stem cells (MSCs) into osteoblasts (bone-forming cells) and maturing
chondrocytes into hypertrophic chondrocytes. Runx2 seems to be a master regulator of multiple
genes including Ihh, Sp7, Col10a1, Spp1, Ibsp, Mmp13, and Vegfa. [162, 236]. Studies have
shown that knocking down Runx2 inhibits the formation of bone causing mice to be born with
skeletons made of pure cartilage [237]. Knowing the gene pathways of bone formation gives us
one piece of the puzzle to disrupt the development of the baculum. In the second chapter of this
thesis titled “Androgen receptor (AR) Co-immuno precipitation and sequencing (ChIP-seq) and
single-cell RNA-sequencing identify major differences in cells that give rise to the baculum
(penis bone) compared to cells that give rise to forelimb and hindlimb skeletal system”, we found

82
that several bone genes in the developing baculum had androgen receptor binding locations
while developing hindlimb bone genes did not. This implies that bone genes in the developing
baculum are regulated by androgen receptor binding giving us the second piece of the puzzle.
With both puzzle pieces in mind, we were able to generate a genetically modified male mouse
that has a severely disrupted baculum to the point where only a small misshapen distal tip is the
only portion that forms. We implemented the cre-lox system to knock out androgen receptors in
cells expressing Runx2, thus disrupting the formation of the baculum. Past experiments have
disrupted androgen signaling in male mice via castration. While those studies have shown that
the baculum is disrupted/not developed, several off-target effects confound the function of the
baculum. Many of the off-target effects arise in penile development (e.g. disruption of penile
spine development, penis size, male feminization), overall body size, and male mating behavior
[58, 180, 181, 238]. In our experiments, we minimized the off-target effects by targeting only the
baculum allowing us to study the mechanical function of the baculum.
Materials and Methods
4.1 Genetic Disruption of the Baculum’s Development using the Cre-Lox System to knock out
the Androgen Receptor in Runx2 Expressing Cells
4.1.1 Cre-Lox System Design and PCR
Because the baculum is highly dependent on androgen signaling, we aimed to disrupt the
development of the baculum. We chose to knock out the androgen receptor (Ar) in bone cells
using a Cre-Lox system. We used B6.129S1-Artm2.1Reb/J female mice which have Exon 1 of their

83
X-linked androgen receptor floxed (FAr) (strain #: 018450). Female FAr mice were maintained
as homozygous, and male FAr mice were maintained as hemizygous. Animals were genotyped
according to protocols established by Jackson Laboratories. The genotyping was done using 2
primer pairs. The first primer pair, Ar P1 (5′-CAGCACCCTACACTAGAATACTG-3′) and Ar P2
(5′-AATGACCTGAGAGTGCTTCCTCC-3′), was used to amplify the 5’ LoxP site of the FAr
allele. The second primer pair, Ar P3 (5′-AGGGCACAGAGTAAGCAGTTTGC-3′) and Ar P4
(5′-TCCAGATGTAGGACAGACCTTCC-3′), was used to amplify the 3’ LoxP site of the FAr
allele. Both primer sets produce a 200- and 220- bp control and FAr allele respectively [239]. We
used a touch down PCR to amplify the FAr alleles.
The PCR Mix recipe is the following (total of 25 µL per rxn):
1. 14.75 µL of Nuclease Free H2O
2. 2.5 µL of 10X PCR Reaction Buffer
3. 0.5 µL of 10 mM dNTPs
4. 1 µL of 10 µM Forward Primer
5. 1µL of 10 µM Reverse Primer
6. 0.25 µL of Taq Polymerase
7. 5 µL of DNA
The PCR program is the following:
1. 94ºC for 2 minutes
2. 94ºC for 20 seconds
3. 65ºC for 15 seconds
4. 68ºC for 10 seconds

84
repeat steps 2 – 4 for 10 cycles decreasing the temperature in step 3 by -0.5C every cycle
5. 94ºC for 15 seconds
6. 56ºC for 15 seconds
7. Extension: 72ºC for 10 seconds
repeat steps 5-7 for 28 cycles
8. Final Extension: 72ºC for 2 minutes
9. Indefinite hold at 4ºC
To target the Ar in bone genes, we used the 777 cbfa-cre (B6) (Runx2-cre (B6)) GVO strain from
Dr. Jan Tuckermann at Ulm University in Germany [240]. The Runx2-cre mice were maintained
as heterozygous because the strain is homozygous lethal. Animals were genotyped according to
the protocols established by the Tuckermann lab. The Tuckermann lab uses a 3-primer strategy
for amplifying the transgene and the wildtype gene. The three primers are the following: cbfa_24
(5′-CCAGGAAGACTGCCAGAAGG -3′), cbfa_25 (5′-TGGCTTGCAGGTACAGGAG-3′), and
cbfa-30 (5′-GGAGCTGCCGAGTCAATAAC-3′). The combination of cbfa_24 and cbfa_25
primers generate a transgene product of 600 bp while the combination of cbfa_24 and cbfa_30
generates a wildtype product of 780 pb.
The PCR mix recipe is the following (total of 25 µL per rxn):
1. 16.8 µL of Nuclease Free H2O
2. 2.5 µL of 10X PCR Reaction Buffer
3. 0.5 µL of 10mM dNTPs
4. 0.75 µL of Primer cbfa_24
5. 0.75 µL of Primer cbfa_25

85
6. 0.75 µL of Primer cbfa_30
7. 0.75 µL of MgCl2
8. 0.2 µL of Taq Polymerase
9. 2 µL of DNA
The PCR program is the following:
1. 95ºC for 5 minutes
2. 95ºC for 1 minute
3. 59ºC for 30 seconds
4. 72ºC for 1 minute
Repeat steps 2-4 for 35 cycles
5. 72ºC for 10 minutes
6. Indefinite hold at 15ºC
We crossed homozygous FAr females with Runx2-cre heterozygous males. This produced male
animals that were hemizygous for the FAr allele and heterozygous for the Runx2-cre allele and
male animals that were hemizygous for the FAr allele and wildtype (WT) for the Runx2 allele
which served as a control for our experiments. For the purposes of this study, we only cared
about the male animals. Males that were hemizygous for the FAr allele and heterozygous for the
Runx2-cre allele should have their Ar knocked out only in cells expressing Runx2.

86
4.2 Histology Analysis of FArRunx2-cre Mutant Baculum
To examine the status of the baculum of the FArRunx2-cre male mice, we did histology using Von
Kossa (VK) staining to quantify mineralization (bone) and Safranin-O/Fast Green (SafO/FG)
staining for cartilage. Adult male specimens (n=2), 12 weeks of age, were euthanized, and their
penises dissected. The penises were fixed in formalin for 3-5 days. Fixed specimens were
shipped to, and all paraffin embedding, tissue sectioning, and tissue staining was done by the
histology core at the Indiana University School of Medicine.
4.3 MicroCT Scanning and Bone Morphometric Analysis of FArRunx2-cre Mutant Baculum
To further assess the bone shape and location of the baculum of the FArRunx2-cre mice, we had our
specimens scanned using the PaleoCT (a micro-X-ray computed tomography (CT) scanner) at
the University of Chicago. The PaleoCT setup allows us to scan our small samples at high
resolutions. Segmentation of the data was done using 3D Slicer (slicer.org) [241] and a custom
Python and R script.
4.4 Assessing Male Mouse Mating Behavior
Male mouse mating behavior has been well documented. When a female mouse in estrus is
placed with a male mouse, the male mouse begins by smelling the female’s anogenital region.
Once he has investigated the female thoroughly, the male will place his paws on the female’s
flanks and restrain the female. He then begins thrusting his pelvis in hopes of inserting his penis
into the female’s vaginal canal. After several intromissions the male will ejaculate and freeze for
~20-25 seconds [58]. The male then dismounts the female and begins to groom himself. Both

87
animals then don’t interact with each other for several minutes. For my experiments, FArRunx2-cre
male mice were placed with WT females in the same cage for two weeks at a time. After mating
event/ejaculation has occurred, males leave a copulatory plug in the female’s vaginal canal [242-
244]. I used the presence or absence of the copulatory plug as evidence for a mating
event/ejaculation. I checked for copulatory plug presence once in the morning and once at night
for the duration of two weeks. After two weeks males and females were separated, and females
left alone to see if they give birth.
Results
4.5 Androgen Receptor Knockout via Cre-Lox in Runx2 Expressing Cells Disrupts the
Baculum’s Development in Mice
To visually assess the baculum of the mutant mice we used bone and cartilage staining alongside
microCT scanning. VK staining of the FAr/Runx2-cre male mice penis showed very minimal
staining for mineralized tissue resulting in an abnormally underdeveloped baculum (Figure 1C).
SafO/FG staining showed the cartilaginous tip that is part of the baculum known as the MUMP is
present albeit misshapen. (Figure 1D). Gross dissections of the FArRunx2-cre mouse penis produced
a small distal tip-like bone structure that could be dissected out. Histological VK and SafO/FG
staining and gross dissections of the FAr/Runx2-WT control mouse penis produced a normal
baculum (Figure 1A) and MUMP (Figure 1B). Apart from the underdeveloped baculum, the rest
of the FArRunx2-cre male mouse penis developed normally including the penile spines which are
known to be highly androgen sensitive (Figure 1A-D – Red Arrows). MicroCT scanning of the
FArRunx2-cre mouse (Figure 1F) shows a better picture of the reduced baculum and its position in
the glans penis in relation to a wildtype mouse (Figure 1E). While a normally formed baculum

88
spans the entire glans of the adult mouse penis, the reduced baculum is located and possibly
overlaps with the distal tip of the baculum.
4.6 Mating Behavior of the FArRunx2-cre Mouse with the Underdeveloped Baculum
All male FArRunx2-cre mice behaving normally. Multiple mating attempts were observed. Males
investigated the female’s anogenital region as observed in their WT counterparts. FArRunx2-cre
male mice attempted mounting females as per normal behavior. After two weeks of checking for
copulatory plugs morning and night, no plugs were observed. After females were split from the
males to wait for possible pregnancy, none of these instances resulted in pups being born. In one
case, a FArRunx2-Cre male was left with two females for ~3 months. Even then, no pups resulted
from those crossing. However, in only one other case, a FArRunx2-cre male mouse was weaned at 3
weeks (early November 2023) with his 2 sisters. They reached sexual maturity in late December
2023 at 8 weeks old while being housed in the same cage. They were all paired together until
March 2024, ~2 months after reaching sexual maturity. They were unpaired because pups were
born to this specific crossing. One of the two females gave birth to 6 pups, 4 females and 2
males. Male FArRunx2-cre mice in any other pairing scenario have still not been able to produce
progeny as of submitting my thesis. Because progeny has only been produced in this one
extraneous case, continuous breeding of this genotype has to be carried out by crossing FAr
homozygous female mice and Runx2-cre heterozygous male mice.
Discussion
The baculum is a sexually selected structure important for reproductive success. Several
hypothesis have been suggested for the function of the baculum, but studies show contradictory

89
evidence. By knocking down the androgen receptor in Runx2 expressing cells in male mice via a
cre-lox system, we managed to severely disrupt the development of the baculum. After checking
body size, femur and humerus length, sperm production and motility, testes size, penile
morphology, the FArRunx2-cre male mice normal in all other aspects except for the development of
the baculum. All of the FArRunx2-cre mice analyzed so far appear to express no significant
variation in the reduced baculum’s morphology, but more still need to be analyzed to get a better
view about possible morphological diversity. This work presents a new model organism with a
severely disrupted baculum allowing us to study the mechanical function of the baculum without
any other confounding factors. Further characterization of this model requires single-cell RNA
sequencing to further validate the decrease/disappearance of osteoblasts and chondrocytes, and
Ar ChIP-sequencing to validate the decrease/disappearance of androgen receptor binding in the
developing mouse baculum.
While the mutant mice may behave completely normal, we find that the FArRunx2-cre male mice
have an extremely reduced ability to procreate in a lab setting. The lack of copulatory plug
presence and sired offspring implies the importance of the baculum’s role in intromission.
Studies in bats suggested that the baculum and CC are required for penile shaft and distal tip
support [219]. In rats, the proximal base of the baculum sits in a depression in the CC separated
via a layer of fibrocartilage creating a support unit [245]. Our FArRunx2-cre mice lack a proximal
base and histology shows that the CC appears to be present in the penile shaft without connecting
to any structure possibly reducing the support unit. The misshapen distal tip of the baculum that
does form does not seem to interact with the CC. We hypothesize that without the baculum’s
proximal base, the shaft of the mouse penis does not have enough support for proper
intromission leading to a severely decreased reproductive success rate.

90
If the baculum is indeed needed for support and intromission, aside from its other functions, then
animals like rabbits, which don’t have a baculum [175], have figured out a distinct way to get the
same amount of penile shaft and tip support for proper intromission. It is possible that the penile
support in baculumless animals comes from morphological variation in penile structure, possibly
in erectile tissue. Studies focusing on the development of the rabbit penis have shown major
differences between adult rabbit penises and adult mice/rat penises. For example, the glans of an
adult rabbit penis is rudimentary, tapered, and covered with one prepuce tissue, while in adult
mice/rats have a well-developed glans and penile body and covered by two prepuce tissues
[246]. Studies have shown that in rabbits there is a different distribution of nerve types between
CC and corpus spongiosum (CS) when compared to cats, rats, monkeys, and humans [247].
While the differences in the nerve amounts in the penis of rabbits might explain how they deal
with not having a baculum, it doesn’t explain why the innervation of human penises, which lack
a baculum as well, are not more similar to rabbit penises. One way to test whether erectile tissue
can rescue the lack of support provided by the baculum would be to give the FArRunx2-cre male
mice a vasodilator that would allow more blood flow into the existing erectile tissue of their
penis. Hypothetically, this could restore the FArRunx2-cre male mouse’s ability to intromit due to
the increase in blood flow and support.
The FArRunx2-cre male mice still make a misshapen distal tip of the baculum This data suggests
that proximal base of the baculum is androgen sensitive while the distal tip is not entirely
sensitive to androgen signaling. Studies have shown that the baculum develops the proximal base
using endochondral ossification first and later after going into and continuing a little bit after
puberty the distal portion of the baculum gets made using intramembranous ossification.
Considering that the FArRunx2-cre mice are completely missing the proximal base of the baculum,

91
we have possibly disrupted the endochondral ossification of the baculum meaning that the
endochondral ossification portion of the baculum is androgen sensitive. One possible way of
testing this hypothesis is to use the same cre-lox system with the FAr and the cre enzyme driven
by a Col10a1 promoter which is a known marker of hypertrophic chondrocytes responsible for
endochondral ossification. Presumably, this version of the cre-lox design would generate the
same phenotype as the FArRunx2-cre. There are a versions of the FAr cre-lox system that can be
used to generate varying phenotypes of disrupted bacular development. One other version would
be to disrupt the baculum by disrupting chondrocyte development all together by using cre
driven by a Col2a1 promoter. To further address the many functions of the baculum, it would be
important to heavily manipulate the shape of the baculum using other cre driven genes in effort
to better understand how shape may affect reproduction in animals with bacula. Regardless, there
are many avenues left to explore.

92
Figure Legends
Figure 3.1.
Wildtype Baculum vs. FArRunx2-cre Reduced-Baculum – Bone stain, Cartilage stain, and bone
microCT scanning images. A and C) Von Kossa staining of calcium deposits in the adult mouse
penis, wildtype vs mutant respectively. A) shows a normal stained baculum while C) shows very
small piece of bone stained, albeit reduced/malformed. B and D) Safranin-O/Fast Green staining
of cartilage in the adult mouse penis, wildtype vs mutant respectively. B) shows the normal
cartilage tip of the baculum (the MUMP) stained while D) shows the cartilage tip stained also
reduced/malformed. E and F) shows a microCT scanning image of the bone part of the baculum,
wildtype vs. mutant respectively. Yellow circles in all panels show where the bone and cartilage
tip are present. Red circles in E and F show where the bone is present in the microCT scanned
images. Red arrows in A-D show the penile spines the develop normally.

93
Figure 3.1
E F

94
References
1. Darwin, C., On the origin of species: facsimile of the first edition. 1859.
2. Andersson, M. and Y. Iwasa, Sexual selection. Trends Ecol Evol, 1996. 11(2): p. 53-8.
3. Lande, R., Sexual Dimorphism, Sexual Selection, and Adaptation in Polygenic
Characters. Evolution, 1980. 34(2): p. 292-305.
4. Shutler, D., Sexual selection: when to expect trade-offs. Biol Lett, 2011. 7(1): p. 101-4.
5. Johnston, S.E., et al., Life history trade-offs at a single locus maintain sexually selected
genetic variation. Nature, 2013. 502(7469): p. 93-5.
6. Owens, I.P.F. and I.R. Hartley, Sexual dimorphism in birds: why are there so many
different forms of dimorphism? Proceedings of the Royal Society of London Series B:
Biological Sciences, 1998. 265(1394): p. 397-407.
7. Ohlsson, T., et al., Pheasant sexual ornaments reflect nutritional conditions during early
growth. Proc Biol Sci, 2002. 269(1486): p. 21-7.
8. Malo, A.F., et al., Antlers honestly advertise sperm production and quality. Proc Biol Sci,
2005. 272(1559): p. 149-57.
9. Emlen, D.J., J. Hunt, and L.W. Simmons, Evolution of sexual dimorphism and male
dimorphism in the expression of beetle horns: phylogenetic evidence for modularity,
evolutionary lability, and constraint. Am Nat, 2005. 166 Suppl 4: p. S42-68.
10. Min, S.E., Variation in Sexually Dimorphic Traits of Male Pronghorns. Journal of
Mammalogy, 1997. 78(1): p. 31-47.
11. Thoren, S., P. Lindenfors, and P.M. Kappeler, Phylogenetic analyses of dimorphism in
primates: evidence for stronger selection on canine size than on body size. Am J Phys
Anthropol, 2006. 130(1): p. 50-9.
12. Davis, E.B., K.A. Brakora, and A.H. Lee, Evolution of ruminant headgear: a review.
Proc Biol Sci, 2011. 278(1720): p. 2857-65.
13. Chen, I.P., et al., Sexual selection and the evolution of complex color patterns in dragon
lizards. Evolution, 2012. 66(11): p. 3605-14.
14. Stuart-Fox, D.M. and T.J. Ord, Sexual selection, natural selection and the evolution of
dimorphic coloration and ornamentation in agamid lizards. Proc Biol Sci, 2004.
271(1554): p. 2249-55.
15. Kodric-Brown, A., Sexual Dichromatism and Temporary Color Changes in the
Reproduction of Fishes. American Zoologist, 1998. 38(1): p. 70-81.
16. Brzozowski, F., et al., Sexually dimorphic levels of color trait integration and the
resolution of sexual conflict in Lake Malawi cichlids. J Exp Zool B Mol Dev Evol, 2012.
318(4): p. 268-78.

95
17. Mitoyen, C., C. Quigley, and L. Fusani, Evolution and function of multimodal courtship
displays. Ethology, 2019. 125(8): p. 503-515.
18. Ralls, K., Sexual Dimorphism in Mammals: Avian Models and Unanswered Questions.
The American Naturalist, 1977. 111(981): p. 917-938.
19. McPherson, F.J. and P.J. Chenoweth, Mammalian sexual dimorphism. Anim Reprod Sci,
2012. 131(3-4): p. 109-22.
20. Ralls, K. and S.L. Mesnick, Sexual Dimorphism, in Encyclopedia of Marine Mammals.
2009, American Press. p. 1005-1011.
21. Le Boeuf, B.J., Male-male Competition and Reproductive Success in Elephant Seals.
American Zoologist, 1974. 14(1): p. 163-176.
22. Dean, M.D., K.G. Ardlie, and M.W. Nachman, The frequency of multiple paternity
suggests that sperm competition is common in house mice (Mus domesticus). Molecular
Ecology, 2006. 15(13): p. 4141-4151.
23. Dean, M.D. and M.W. Nachman, Faster fertilization rate in conspecific versus
heterospecific matings in house mice. Evolution, 2009. 63(1): p. 20-28.
24. Poiani, A., Complexity of seminal fluid: A review. Behavioral Ecology and Sociobiology,
2006. 60(3): p. 289-310.
25. Price, C.S.C., Conspecific sperm precedence in Drosophila. Nature, 1997. 388(6643): p.
663-666.
26. Laws, R.M., The Elephant Seal (Mirounga leonina Linn.) The Physiology of
Reproduction. Scientific Reports, 1956. 15: p. 1-80.
27. Emlen, D.J., et al., A mechanism of extreme growth and reliable signaling in sexually
selected ornaments and weapons. Science, 2012. 337(6096): p. 860-4.
28. Clutton_Brock, T.H., The functions of antlers. Behaviour, 1982. 79(2/4): p. 108-125.
29. Jennings, D.J., et al., Effect of body weight, antler length, resource value and experience
on fight duration and intensity in fallow deer. Animal Behaviour, 2004. 68(1): p. 213-
221.
30. Fitzpatrick, J.L., et al., Male contest competition and the coevolution of weaponry and
testes in pinnipeds. Evolution, 2012. 66(11): p. 3595-604.
31. Wedell, N., M.J.G. Gage, and G.A. Parker, Sperm competition, male prudence and sperm
limited females. Trends in Ecology and Evolution, 2002. 17(7): p. 313-320.
32. Wigby, S. and T. Chapman, Sperm competition. Current Biology, 2004. 14(3): p. R100-
R103.
33. Wigby, S., et al., Seminal Fluid Protein Allocation and Male Reproductive Success.
Current Biology, 2009. 19(9): p. 751-757.

96
34. Pizzari, T. and G.A. Parker, Sperm competition and sperm phenotype, in Sperm Biology:
An Evolutionary Perspective, D.J. Hosken, S. Pitnick, and T.R. Birkhead, Editors. 2009,
Elsevier Science. p. 207-245.
35. Gomendio, M. and E.R. Roldan, Sperm competition influences sperm size in mammals.
Proc Biol Sci, 1991. 243(1308): p. 181-5.
36. Lande, R. and S.J. Arnold, Evolution of mating preference and sexual dimorphism. J
Theor Biol, 1985. 117(4): p. 651-64.
37. Houck, L.D. and S.J. Arnold, Courtship and Mating Behavior, in Reproductive Biology
and Phylogeny of Urodela. 2003. p. 383-424.
38. Neunuebel, J.P., et al., Female mice ultrasonically interact with males during courtship
displays. eLife, 2015. 4(MAY): p. 1-24.
39. Girard, M.B., M.M. Kasumovic, and D.O. Elias, Multi-modal courtship in the peacock
spider, Maratus volans (O.P.-Cambridge, 1874). PLoS One, 2011. 6(9): p. e25390.
40. Ligon, R.A., et al., Evolution of correlated complexity in the radically different courtship
signals of birds-of-paradise. PLoS Biol, 2018. 16(11): p. e2006962.
41. Borgia, G., Sexual Selection in Bowerbirds. Scientific American, 1986. 254(6): p. 92-
101.
42. Miles, M.C., F. Goller, and M.J. Fuxjager, Physiological constraint on acrobatic
courtship behavior underlies rapid sympatric speciation in bearded manakins. Elife,
2018. 7.
43. Barske, J., et al., Female choice for male motor skills. Proc Biol Sci, 2011. 278(1724): p.
3523-8.
44. French, A.R. and T.B. Smith, Importance of Body Size in Determining Dominance
Hierarchies among Diverse Tropical Frugivores1. Biotropica, 2005. 37(1): p. 96-101.
45. Neff, B.D. and T.E. Pitcher, Genetic quality and sexual selection: an integrated
framework for good genes and compatible genes. Mol Ecol, 2005. 14(1): p. 19-38.
46. Brown, J.H. and B.A. Maurer, Body size, ecological dominance and Cope's rule. Nature,
1986. 324(6094): p. 248-250.
47. Miller, E.J., et al., Dominance, body size and internal relatedness influence male
reproductive success in eastern grey kangaroos (Macropus giganteus). Reprod Fertil
Dev, 2010. 22(3): p. 539-49.
48. Andersson, M., Sexual selection. Vol. 72. 1994: Princeton University Press.
49. Kang, J.H., et al., Transcriptomics of two evolutionary novelties: how to make a spermtransfer organ out of an anal fin and a sexually selected "sword" out of a caudal fin. Ecol
Evol, 2015. 5(4): p. 848-64.

97
50. Emlen, D.J., et al., Insulin signaling and limb-patterning: candidate pathways for the
origin and evolutionary diversification of beetle 'horns'. Heredity (Edinb), 2006. 97(3): p.
179-91.
51. Gunter, H.M., et al., Identification and characterization of gene expression involved in
the coloration of cichlid fish using microarray and qRT-PCR approaches. J Mol Evol,
2011. 72(2): p. 127-37.
52. Pointer, M.A., et al., A novel method for screening a vertebrate transcriptome for genes
involved in carotenoid binding and metabolism. Mol Ecol Resour, 2012. 12(1): p. 149-
59.
53. Baker, B.S., B.J. Taylor, and J.C. Hall, Are Complex Behaviors Specified by Dedicated
Regulatory Genes? Reasoning from Drosophila. Cell, 2001. 105(1): p. 13-24.
54. Demir, E. and B.J. Dickson, fruitless splicing specifies male courtship behavior in
Drosophila. Cell, 2005. 121(5): p. 785-94.
55. Ito, H., et al., Sexual orientation in Drosophila is altered by the satori mutation in the
sex-determination gene fruitless that encodes a zinc finger protein with a BTB domain.
Proceedings of the National Academy of Sciences, 1996. 93(18): p. 9687-9692.
56. Ryner, L.C., et al., Control of male sexual behavior and sexual orientation in Drosophila
by the fruitless gene. Cell, 1996. 87(6): p. 1079-89.
57. Clyne, J.D. and G. Miesenbock, Sex-specific control and tuning of the pattern generator
for courtship song in Drosophila. Cell, 2008. 133(2): p. 354-63.
58. Hull, E.M. and J.M. Dominguez, Sexual behavior in male rodents. Horm Behav, 2007.
52(1): p. 45-55.
59. Le Roith, D., Seminars in medicine of the Beth Israel Deaconess Medical Center. Insulinlike growth factors. N Engl J Med, 1997. 336(9): p. 633-40.
60. Baker, J., et al., Role of insulin-like growth factors in embryonic and postnatal growth.
Cell, 1993. 75(1): p. 73-82.
61. Warren, I.A., et al., A general mechanism for conditional expression of exaggerated
sexually-selected traits. Bioessays, 2013. 35(10): p. 889-99.
62. Ranz, J.M., et al., Sex-dependent gene expression and evolution of the Drosophila
transcriptome. Science, 2003. 300(5626): p. 1742-5.
63. Yang, X., et al., Tissue-specific expression and regulation of sexually dimorphic genes in
mice. Genome Res, 2006. 16(8): p. 995-1004.
64. Xiong, L., et al., Direct androgen receptor control of sexually dimorphic gene expression
in the mammalian kidney. Dev Cell, 2023.
65. Kitano, J., et al., Differences in the contributions of sex linkage and androgen regulation
to sex-biased gene expression in juvenile and adult sticklebacks. J Evol Biol, 2020.

98
66. Lindenfors, P., J.L. Gittleman, and K.E. Jones, Sexual size dimorphism in mammals.
Book: Sex, Size, and Gender Roles, 2007: p. 16-26.
67. Fairbairn, D.J., W.U. Blanckenhorn, and T. Székely, Sex, size and gender roles:
evolutionary studies of sexual size dimorphism. 2007: Oxford University Press, USA.
68. Reeve, J.P. and D.J. Fairbairn, Predicting the evolution of sexual size dimorphism.
Journal of Evolutionary Biology, 2001. 14(2): p. 244-254.
69. Rice, W.R. and B. Holland, The enemies within: intergenomic conflict, interlocus contest
evolution (ICE), and the intraspecic Red Queen. Behavioral Ecology and Sociobiology,
1997. 41: p. 1-10.
70. Speakman, J.R., Body size, energy metabolism and lifespan. J Exp Biol, 2005. 208(Pt 9):
p. 1717-30.
71. Pennell, T.M. and E.H. Morrow, Two sexes, one genome: the evolutionary dynamics of
intralocus sexual conflict. Ecol Evol, 2013. 3(6): p. 1819-34.
72. McIntyre, L.M., et al., Sex-specific expression of alternative transcripts in Drosophila.
Genome Biol, 2006. 7(8): p. R79.
73. Bachtrog, D., A dynamic view of sex chromosome evolution. Curr Opin Genet Dev, 2006.
16(6): p. 578-85.
74. Badyaev, A., Growing apart: an ontogenetic perspective on the evolution of sexual size
dimorphism. 2002. 17(8): p. 369-378.
75. Claessens, F., et al., Selective DNA binding by the androgen receptor as a mechanism for
hormone-specific gene regulation. Journal of Steroid Biochemistry & Molecular Biology,
2001. 76: p. 23-30.
76. Bourdeau, V., et al., Mechanisms of primary and secondary estrogen target gene
regulation in breast cancer cells. Nucleic Acids Res, 2008. 36(1): p. 76-93.
77. McEwan, I.J., Molecular mechanisms of androgen receptor-mediated gene regulation:
structure-function analysis of the AF-1 domain. Endocr Relat Cancer, 2004. 11(2): p.
281-93.
78. Sanchez, R., et al., Diversity in the mechanisms of gene regulation by estrogen receptors.
Bioessays, 2002. 24(3): p. 244-54.
79. Verrijdt, G., A. Haelens, and F. Claessens, Selective DNA recognition by the androgen
receptor as a mechanism for hormone-specific regulation of gene expression. Mol Genet
Metab, 2003. 78(3): p. 175-85.
80. Robins, D.M., Androgen receptor and molecular mechanisms of male-specific gene
expression. Novartis Found Symp, 2005. 268: p. 42-52; discussion 53-6, 96-9.
81. Ellegren, H. and J. Parsch, The evolution of sex-biased genes and sex-biased gene
expression. Nat Rev Genet, 2007. 8(9): p. 689-98.

99
82. Ingleby, F.C., I. Flis, and E.H. Morrow, Sex-biased gene expression and sexual conflict
throughout development. Cold Spring Harb Perspect Biol, 2014. 7(1): p. a017632.
83. Parsch, J. and H. Ellegren, The evolutionary causes and consequences of sex-biased gene
expression. Nat Rev Genet, 2013. 14(2): p. 83-7.
84. Hau, M., Regulation of male traits by testosterone: implications for the evolution of
vertebrate life histories. Bioessays, 2007. 29(2): p. 133-44.
85. Hale, M.D., et al., Ontogenetic Change in Male Expression of Testosterone-Responsive
Genes Contributes to the Emergence of Sex-Biased Gene Expression in Anolis sagrei.
Front Physiol, 2022. 13: p. 886973.
86. Mayne, B.T., et al., Large Scale Gene Expression Meta-Analysis Reveals Tissue-Specific,
Sex-Biased Gene Expression in Humans. Front Genet, 2016. 7: p. 183.
87. Naqvi, S., et al., Conservation, acquisition, and functional impact of sex-biased gene
expression in mammals. Science, 2019. 365(6450).
88. Smith, J.D., K.F. McManus, and H.B. Fraser, A novel test for selection on cis-regulatory
elements reveals positive and negative selection acting on mammalian transcriptional
enhancers. Mol Biol Evol, 2013. 30(11): p. 2509-18.
89. Hahn, M.W., Detecting natural selection on cis-regulatory DNA. Genetica, 2007. 129(1):
p. 7-18.
90. Anderson, A.P. and A. Jones, The relationship between sexual dimorphism and androgen
response element proliferation in primate genomes. Evolution, 2024. 78(1): p. 195-208.
91. Silva, M. and J.A. Downing, Handbook of Mammalian Body Masses. 1st ed. 1995: CRC
Press.
92. Brassey, C.A., J.D. Gardiner, and A.C. Kitchener, Testing hypotheses for the function of
the carnivoran baculum using finite-element analysis. Proceedings of the Royal Society
B: Biological Sciences, 2018. 285(1887): p. 20181473-20181473.
93. Weckerly, F.W., Sexual-Size Dimorphism: Influence of Mass and Mating Systems in the
Most Dimorphic Mammals. Journal of Mammalogy, 1998. 79(1): p. 33-52.
94. Johnson, S.E., et al., Morphological Variation in Populations of Eulemur albocollaris
and E. fulvus rufus. International Journal of Primatology, 2005. 26(6): p. 1399-1416.
95. Bayart, F. and B. Simmen, Demography, range use, and behavior in black lemurs
(Eulemur macaco macaco) at Ampasikely, northwest Madagascar. Am J Primatol, 2005.
67(3): p. 299-312.
96. Liker, A., et al., Evolution of large males is associated with female-skewed adult sex
ratios in amniotes. Evolution, 2021. 75(7): p. 1636-1649.
97. Trillmich, F., et al., Age at maturity in cavies and guinea‐pigs (Cavia aperea and Cavia
aperea f. porcellus): influence of social factors. Journal of Zoology, 2006. 268(3): p. 285-
294.

100
98. Nazarova, G.G. and L.P. Proskurnyak, Protein content in urine of male and female water
vole (Arvicola amphibius) at the period of spring growth and sexual maturation. Journal
of Evolutionary Biochemistry and Physiology, 2013. 49(3): p. 360-364.
99. Zoonomia, C., A comparative genomics multitool for scientific discovery and
conservation. Nature, 2020. 587(7833): p. 240-245.
100. Christmas, M.J., et al., Evolutionary constraint and innovation across hundreds of
placental mammals. Science, 2023. 380(6643): p. eabn3943.
101. Foley, N.M., et al., A genomic timescale for placental mammal evolution. Science, 2023.
380(6643): p. eabl8189.
102. De Bruyn, R., R. Bollen, and F. Claessens, Identification and characterization of
androgen response elements. Methods Mol Biol, 2011. 776: p. 81-93.
103. Gruber, C.J., et al., Anatomy of the estrogen response element. Trends Endocrinol Metab,
2004. 15(2): p. 73-8.
104. Pettersson, K., et al., Mouse estrogen receptor beta forms estrogen response elementbinding heterodimers with estrogen receptor alpha. Mol Endocrinol, 1997. 11(10): p.
1486-96.
105. Shaffer, P.L., et al., Structural basis of androgen receptor binding to selective androgen
response elements. Proc Natl Acad Sci U S A, 2004. 101(14): p. 4758-63.
106. Roche, P.J., S.A. Hoare, and M.G. Parker, A consensus DNA-binding site for the
androgen receptor. Mol Endocrinol, 1992. 6(12): p. 2229-35.
107. Ikeda, K., K. Horie-Inoue, and S. Inoue, Identification of estrogen-responsive genes
based on the DNA binding properties of estrogen receptors using high-throughput
sequencing technology. Acta Pharmacol Sin, 2015. 36(1): p. 24-31.
108. Jones, B.G., et al., Complex sex-biased antibody responses: estrogen receptors bind
estrogen response elements centered within immunoglobulin heavy chain gene
enhancers. Int Immunol, 2019. 31(3): p. 141-156.
109. Pages, H., et al., Package ‘Biostrings’. Bioconductor, 2013. 18129.
110. Revell, L.J., phytools: an R package for phylogenetic comparative biology (and other
things). Methods in Ecology and Evolution, 2011. 3(2): p. 217-223.
111. Kirilenko, B.M., et al., Integrating gene annotation with orthology inference at scale.
Science, 2023. 380(6643): p. eabn3107.
112. Wilson, S., J. Qi, and F.V. Filipp, Refinement of the androgen response element based on
ChIP-Seq in androgen-insensitive and androgen-responsive prostate cancer cell lines.
Sci Rep, 2016. 6: p. 32611.
113. Ho, L. and C. Ane, A linear-time algorithm for Gaussian and non-Gaussian trait
evolution models. Syst Biol, 2014. 63(3): p. 397-408.

101
114. Rensch, B., Evolution above the species level. 196, New York: Colombia University
Press.
115. Upham, N.S., J.A. Esselstyn, and W. Jetz, Inferring the mammal tree: Species-level sets
of phylogenies for questions in ecology, evolution, and conservation. PLoS Biol, 2019.
17(12): p. e3000494.
116. Mi, H., et al., Large-scale gene function analysis with the PANTHER classification
system. Nat Protoc, 2013. 8(8): p. 1551-66.
117. Denayer, S., et al., The rules of DNA recognition by the androgen receptor. Mol
Endocrinol, 2010. 24(5): p. 898-913.
118. Sahu, B., et al., Androgen receptor uses relaxed response element stringency for selective
chromatin binding and transcriptional regulation in vivo. Nucleic Acids Res, 2014.
42(7): p. 4230-40.
119. Nelson, C.C., et al., Determinants of DNA sequence specificity of the androgen,
progesterone, and glucocorticoid receptors: evidence for differential steroid receptor
response elements. Mol Endocrinol, 1999. 13(12): p. 2090-107.
120. Umesono, K. and R.M. Evans, Determinants of target gene specificity for steroid/thyroid
hormone receptors. Cell, 1989. 57(7): p. 1139-46.
121. John-Alder, H.B., et al., Development of sexual size dimorphism in lizards: testosterone
as a bipotential growth regulator. Sex, size and gender roles: Evolutionary studies of
sexual size dimorphism, 2007: p. 195-204.
122. Cheverud, J.M., M.M. Dow, and W. Leutenegger, The Quantitative Assessment of
Phylogenetic Constraints in Comparative Analyses: Sexual Dimorphism in Body Weight
among Primates. Evolution, 1985. 39(6): p. 1335-1351.
123. Kuparinen, A., E. Yeung, and J.A. Hutchings, Correlation between body size and
longevity: New analysis and data covering six taxonomic classes of vertebrates. Acta
Oecologica, 2023. 119.
124. Bonduriansky, R., et al., Sexual selection, sexual conflict and the evolution of ageing and
life span. Functional Ecology, 2008. 22(3): p. 443-453.
125. Gillooly, J.F., et al., Effects of size and temperature on developmental time. Nature, 2002.
417(6884): p. 70-3.
126. Gillooly, J.F., et al., Effects of size and temperature on metabolic rate. Science, 2001.
293(5538): p. 2248-51.
127. Gaillard, J.-M., et al., Senescence in natural populations of mammals: a reanalysis.
Evolution, 1994: p. 509-516.
128. Promislow, D.E.L., Senescence in Natural Populations of Mammals: A Comparative
Study. Evolution, 1991. 45(8): p. 1869-1887.
129. Peters, R.H., The ecological implications of body size. Vol. 2. 1986: Cambridge
university press.

102
130. Schmidt-Nielsen, K., Scaling: why is animal size so important? 1984: Cambridge
university press.
131. Curlis, J.D., C.L. Cox, and R.M. Cox, Sex-Specific Population Differences in Resting
Metabolism Are Associated with Intraspecific Variation in Sexual Size Dimorphism of
Brown Anoles. Physiol Biochem Zool, 2021. 94(4): p. 205-214.
132. Wiley, R.H., Evolution of social organization and life-history patterns among grouse.
The Quarterly Review of Biology, 1974. 49(3): p. 201-227.
133. Teder, T. and W. Blanckenhorn, Sexual size dimorphism requires a corresponding sex
difference in development time: a meta-analysis in insects. Functional Ecology, 2014.
28(2): p. 479-486.
134. Leigh, S.R., Patterns of variation in the ontogeny of primate body size dimorphism.
Journal of Human Evolution, 1992. 23: p. 27-50.
135. Shea, B.T., The ontogeny of sexual dimorphism in the African apes. Am J Primatol, 1985.
8(2): p. 183-188.
136. Kozlowski, J., Sexual Size Dimorphism: A Life History Perspective. Oikos, 1989. 54(2):
p. 253-255.
137. Diaz, J.M.V., The Role of Gene Duplication in Mediating Peto's Paradox in Afrotheria
and Chiroptera, in Department of Human Genetics. 2020, University of Chicago:
Chicago, Illinois.
138. Jurgens, K.D. and J. Prothero, Scaling of maximal lifespan in bats. Comp Biochem
Physiol A Comp Physiol, 1987. 88(2): p. 361-7.
139. Podlutsky, A.J., et al., A new field record for bat longevity. J Gerontol A Biol Sci Med
Sci, 2005. 60(11): p. 1366-8.
140. Seim, I., et al., Genome analysis reveals insights into physiology and longevity of the
Brandt's bat Myotis brandtii. Nat Commun, 2013. 4: p. 2212.
141. Lu, D., C.Q. Zhou, and W.B. Liao, Sexual size dimorphism lacking in small mammals.
North-Western Journal of Zoology, 2014. 10(1): p. 53-59.
142. Tombak, K.J., S. Hex, and D.I. Rubenstein, New estimates indicate that males are not
larger than females in most mammal species. Nat Commun, 2024. 15(1): p. 1872.
143. Nelson, A.C., et al., Protein pheromone expression levels predict and respond to the
formation of social dominance networks. J Evol Biol, 2015. 28(6): p. 1213-24.
144. Cox, R.M. and R. Calsbeek, Sexually antagonistic selection, sexual dimorphism, and the
resolution of intralocus sexual conflict. Am Nat, 2009. 173(2): p. 176-87.
145. Anunciado, R., et al., Quantitative Trait Loci for Body Weight in the Intercross between
SM/J and A/J Mice. Experimental Animals 50, 2001. 4: p. 319-324.
146. Chan, Y.F., et al., Parallel selection mapping using artificially selected mice reveals body
weight control loci. Curr Biol, 2012. 22(9): p. 794-800.

103
147. Ruzicka, F., et al., Genome-wide sexually antagonistic variants reveal long-standing
constraints on sexual dimorphism in fruit flies. PLoS Biol, 2019. 17(4): p. e3000244.
148. Innocenti, P. and E.H. Morrow, The sexually antagonistic genes of Drosophila
melanogaster. PLoS Biol, 2010. 8(3): p. e1000335.
149. Long, T.A. and W.R. Rice, Adult locomotory activity mediates intralocus sexual conflict
in a laboratory-adapted population of Drosophila melanogaster. Proc Biol Sci, 2007.
274(1629): p. 3105-12.
150. Pischedda, A. and A.K. Chippindale, Intralocus sexual conflict diminishes the benefits of
sexual selection. PLoS Biol, 2006. 4(11): p. e356.
151. Bininda-Emonds, O.R.P., et al., Forelimb-hindlimb developmental timing changes across
tetrapod phylogeny. BMC Evolutionary Biology, 2007. 7(1): p. 182-182.
152. Coates, M.I., Limb Evolution: Fish fins or tetrapod limbs - a simple twist of fate? Current
Biology, 1995. 5(8): p. 844-848.
153. Gibson-Brown, J.J., et al., Evidence of a role for T-box genes in the evolution of limb
morphogenesis and the specification of forelimb/hindlimb identity. Mechanisms of
Development, 1996. 56(1-2): p. 93-101.
154. Hinchliffe, J.R., Developmental basis of limb evolution. International Journal of
Developmental Biology, 2002. 46(7): p. 835-845.
155. King, M., et al., T-Genes and Limb Bud Development. American Journal of Medical
Genetics, 2006. 140(A): p. 1407-1413.
156. Sears, K.E., T.D. Capellini, and R. Diogo, On the serial homology of the pectoral and
pelvic girdles of tetrapods. Evolution, 2015. 69(10): p. 2543-2555.
157. Naiche, L.A. and V.E. Papaioannou, Loss of Tbx4 blocks hindlimb development and
affects vascularization and fusion of the allantois. Development, 2003. 130(12): p. 2681-
93.
158. Minguillon, C., et al., Hox genes regulate the onset of Tbx5 expression in the forelimb.
Development, 2012. 139(17): p. 3180-8.
159. Marcil, A., et al., Pitx1 and Pitx2 are required for development of hindlimb buds.
Development, 2003. 130(1): p. 45-55.
160. Wang, J.S., et al., PITX1 promotes chondrogenesis and myogenesis in mouse hindlimbs
through conserved regulatory targets. Developmental Biology, 2018. 434(1): p. 186-195.
161. Ohuchi, H., et al., The mesenchymal factor, FGF10, initiates and maintains the
outgrowth of the chick limb bud through interaction with FGF8, an apical ectodermal
factor. Development, 1997. 124(11): p. 2235-44.
162. Komori, T., Runx2, an inducer of osteoblast and chondrocyte differentiation. Histochem
Cell Biol, 2018. 149(4): p. 313-323.

104
163. Matsubara, T., et al., BMP2 regulates Osterix through Msx2 and Runx2 during osteoblast
differentiation. J Biol Chem, 2008. 283(43): p. 29119-25.
164. Takarada, T., et al., Genetic analysis of Runx2 function during intramembranous
ossification. Development, 2016. 143(2): p. 211-8.
165. Villavicencio-Lorini, P., et al., Homeobox genes d11-d13 and a13 control mouse autopod
cortical bone and joint formation. Journal of Clinical Investigation, 2010. 120(6): p.
1994-2004.
166. Kuijper, S., et al., Function and regulation of Alx4 in limb development: complex genetic
interactions with Gli3 and Shh. Dev Biol, 2005. 285(2): p. 533-44.
167. Zákány, J., et al., Regulation of number and size of digits by posterior Hox genes: A dosedependent mechanism with potential evolutionary implications. Proceedings of the
National Academy of Sciences of the United States of America, 1997. 94(25): p. 13695-
13700.
168. Kondo, T., et al., Of fingers, toes and penises. Nature, 1997. 390(6655): p. 29.
169. Infante, C.R., et al., Shared Enhancer Activity in the Limbs and Phallus and Functional
Divergence of a Limb-Genital cis-Regulatory Element in Snakes. Developmental Cell,
2015. 35(1): p. 107-119.
170. Lozovska, A., et al., Tgfbr1 controls developmental plasticity between the hindlimb and
external genitalia by remodeling their regulatory landscape. Nat Commun, 2024. 15(1):
p. 2509.
171. Burt, W., Bacula of North American Mammals. Bacula of North American Mammals,
2017(113): p. 7-108.
172. André, G.I., R.C. Firman, and L.W. Simmons, Phenotypic plasticity in genitalia:
baculum shape responds to sperm competition risk in house mice. Proc Biol Sci, 2018.
285(1882).
173. Simmons, L.W. and R.C. Firman, Experimental evidence for the evolution of the
mammalian baculum by sexual selection. Evolution, 2014. 68(1): p. 276-283.
174. Schultz, N.G., et al., The Genetic Basis of Baculum Size and Shape Variation in Mice.
G3, 2016. 6(May): p. 1141-1151.
175. Schultz, N.G., et al., The Baculum was Gained and Lost Multiple Times during
Mammalian Evolution. Integr Comp Biol, 2016. 56(4): p. 644-656.
176. Williams-Ashman, H.G. and A.H. Reddi, Differentiation of mesenchymal tissues during
phallic morphogenesis with emphasis on the os penis: Roles of androgens and other
regulatory agents. Journal of Steroid Biochemistry and Molecular Biology, 1991. 39(6):
p. 873-881.
177. Murakami, R. and T. Mizuno, Proximal-distal sequence of development of the skeletal
tissues in the penis of rat and the inductive effect of epithelium. Journal of embryology
and experimental morphology, 1986. 92: p. 133-143.

105
178. Murakami, R. and T. Mizuno, Histogenesis of the Os Penis and Os Clitoridis in Rats:
chondrogenesis/bone formation/testosterone/genital tubercle. Development, Growth &
Differentiation, 1984. 26(5): p. 419-426.
179. Lough-Stevens, M., N.G. Schultz, and M.D. Dean, The baubellum is more
developmentally and evolutionarily labile than the baculum. Ecology and Evolution,
2018. 8(2): p. 1073-1083.
180. Glucksmann, A., et al., The effect of neonatal treatment of male mice with antiandrogens
and of females with androgens on the development of the os penis and os clitoridis.
Journal of Anatomy, 1976. 121(2): p. 363-370.
181. Glucksmann, A. and C.P. Cherry, The hormonal induction of an os clitoridis in the
neonatal and adult rat. J Anat, 1972. 112(Pt 2): p. 223-31.
182. Wright, P.L., Development of the Baculum of the Long-tailed Weasel. Proceedings of the
Society for Experimental Biology and Medicine, 1950. 75(3): p. 820-822.
183. Yonezawa, T., et al., Distribution of Aromatase and Sex Steroid Receptors in the
Baculum During the Rat Life Cycle: Effects of Estrogen During the Early Development of
the Baculum1. Biology of Reproduction, 2011. 85(1): p. 105-112.
184. Chiang, C., et al., Mineralization and bone resorption are regulated by the androgen
receptor in male mice. J Bone Miner Res, 2009. 24(4): p. 621-31.
185. Zheng, Z. and M.J. Cohn, Developmental basis of sexually dimorphic digit ratios. Proc
Natl Acad Sci U S A, 2011. 108(39): p. 16289-94.
186. Vanderschueren, D., et al., Sex steroid actions in male bone. Endocr Rev, 2014. 35(6): p.
906-60.
187. Vanderschueren, D., et al., Androgens and bone. Endocr Rev, 2004. 25(3): p. 389-425.
188. Theiler, K., The House Mouse: Atlas of Embryonic Development. Vol. 1. 1989, New
York: Springer-Verlag New York Inc. 178.
189. Pihlajamaa, P., et al., Tissue-specific pioneer factors associate with androgen receptor
cistromes and transcription programs. EMBO J, 2014. 33(4): p. 312-26.
190. Langmead, B. and S.L. Salzberg, Fast gapped-read alignment with Bowtie 2. Nat
Methods, 2012. 9(4): p. 357-9.
191. Zhang, Y., et al., Model-based analysis of ChIP-Seq (MACS). Genome Biol, 2008. 9(9):
p. R137.
192. Heinz, S., et al., Simple combinations of lineage-determining transcription factors prime
cis-regulatory elements required for macrophage and B cell identities. Mol Cell, 2010.
38(4): p. 576-89.
193. Welch, R.P., et al., ChIP-Enrich: gene set enrichment testing for ChIP-seq data. Nucleic
Acids Res, 2014. 42(13): p. e105.

106
194. Hao, Y., et al., Dictionary learning for integrative, multimodal and scalable single-cell
analysis. Nat Biotechnol, 2024. 42(2): p. 293-304.
195. Sharma, N.L., et al., The androgen receptor induces a distinct transcriptional program in
castration-resistant prostate cancer in man. Cancer Cell, 2013. 23(1): p. 35-47.
196. Takeda, D.Y., et al., A Somatically Acquired Enhancer of the Androgen Receptor Is a
Noncoding Driver in Advanced Prostate Cancer. Cell, 2018. 174(2): p. 422-432 e13.
197. Stelloo, S., A.M. Bergman, and W. Zwart, Androgen receptor enhancer usage and the
chromatin regulatory landscape in human prostate cancers. Endocr Relat Cancer, 2019.
26(5): p. R267-R285.
198. Carleton, J.B., K.C. Berrett, and J. Gertz, Multiplex Enhancer Interference Reveals
Collaborative Control of Gene Regulation by Estrogen Receptor alpha-Bound
Enhancers. Cell Syst, 2017. 5(4): p. 333-344 e5.
199. Thormann, V., et al., Genomic dissection of enhancers uncovers principles of
combinatorial regulation and cell type-specific wiring of enhancer-promoter contacts.
Nucleic Acids Res, 2018. 46(6): p. 2868-2882.
200. Jacobs, L.L. and D. Pilbeam, Of Mice and Men: Fossil-based Divergence Dates and
Molecular "Clocks". Journal of Human Evolution, 1980. 9(7): p. 551-555.
201. Mackie, E.J., et al., Endochondral ossification: how cartilage is converted into bone in
the developing skeleton. Int J Biochem Cell Biol, 2008. 40(1): p. 46-62.
202. Delhon, L., et al., The critical role of the TB5 domain of fibrillin-1 in endochondral
ossification. Hum Mol Genet, 2022. 31(22): p. 3777-3788.
203. Iftikhar, M., et al., Lysyl oxidase-like-2 (LOXL2) is a major isoform in chondrocytes and
is critically required for differentiation. J Biol Chem, 2011. 286(2): p. 909-18.
204. Papathanasiou, I., K.N. Malizos, and A. Tsezou, Low-density lipoprotein receptor-related
protein 5 (LRP5) expression in human osteoarthritic chondrocytes. J Orthop Res, 2010.
28(3): p. 348-53.
205. Gong, M., et al., Gene expression profiling: identification of gene expression in human
MSC chondrogenic differentiation. Am J Transl Res, 2018. 10(11): p. 3555-3566.
206. Duval, E., et al., Asporin expression is highly regulated in human chondrocytes. Mol
Med, 2011. 17(7-8): p. 816-23.
207. Lin, E.A. and C.J. Liu, The emerging roles of ADAMTS-7 and ADAMTS-12 matrix
metalloproteinases. Open Access Rheumatol, 2009. 1: p. 121-131.
208. Goldring, M.B., K. Tsuchimochi, and K. Ijiri, The control of chondrogenesis. Journal of
Cellular Biochemistry, 2006. 97(1): p. 33-44.
209. Kozhemyakina, E., A.B. Lassar, and E. Zelzer, A pathway to bone: signaling molecules
and transcription factors involved in chondrocyte development and maturation.
Development, 2015. 142(5): p. 817-31.

107
210. Berendsen, A.D. and B.R. Olsen, Bone development. Bone, 2015. 80: p. 14-18.
211. Olsen, B.R., A.M. Reginato, and W. Wang, Bone Development. Annu. Rev. Cell Dev.
Biol, 2000. 16: p. 191-220.
212. Carter, D.R., M. Wong, and T.E. Orr, Musculoskeletal ontogeny, phylogeny, and
functional adaptation. Journal of Biomechanics, 1991. 24(SUPPL. 1): p. 3-16.
213. Bragdon, B., et al., Bone Morphogenetic Proteins: A critical review. Cellular Signalling,
2011. 23(4): p. 609-620.
214. Hall, B.K. and T. Miyake, All for one and one for all: Condensations and the initiation of
skeletal development. BioEssays, 2000. 22(2): p. 138-147.
215. Ducy, P., et al., A Cbfa1-dependent genetic pathway controls bone formation beyond
embryonic development. Genes and Development, 1999. 13(8): p. 1025-1036.
216. Benjamini, Y. and Y. Hochberg, Controlling the False Discovery Rate: A Practical and
Powerful Approach to Multiple Testing. Journal of Royal Statistical Society. Series B
(Methodological), 1995. 57(1): p. 289-300.
217. Orr, T.J. and P.L. Brennan, All Features Great and Small-the Potential Roles of the
Baculum and Penile Spines in Mammals. Integr Comp Biol, 2016. 56(4): p. 635-43.
218. Rayfield, E.J., Finite Element Analysis and Understanding the Biomechanics and
Evolution of Living and Fossil Organisms. Annual Review of Earth and Planetary
Sciences, 2007. 35(1): p. 541-576.
219. Herdina, A.N., et al., Testing hypotheses of bat baculum function with 3D models derived
from microCT. J Anat, 2015. 226(3): p. 229-35.
220. Dixson, A.F., Baculum length and copulatory behavior in primates. American Journal of
Primatology, 1987. 13(1): p. 51-60.
221. Dixson, A.F., Observations on the evolution of the genitalia and copulatory behaviour in
male primates. Journal of Zoology, 1987. 213(3): p. 423-443.
222. Clear, E., et al., Baculum shape complexity correlates to metrics of post-copulatory
sexual selection in Musteloidea. J Morphol, 2023. 284(4): p. e21572.
223. Dixson, A.F., Baculum length and copulatory behaviour in carnivores and pinnipeds
(Grand Order Ferae). Journal of Zoology, 1995. 235(1): p. 67-76.
224. Larivière, S. and S.H. Ferguson, On the evolution of the mammalian baculum: vaginal
friction, prolonged intromission or induced ovulation? Mammal Review, 2002. 32(4): p.
283-294.
225. Stockley, P., et al., Baculum morphology predicts reproductive success of male house
mice under sexual selection. BMC Biology, 2013. 11.
226. Andre, G.I., R.C. Firman, and L.W. Simmons, Baculum shape and paternity success in
house mice: evidence for genital coevolution. Philos Trans R Soc Lond B Biol Sci, 2020.
375(1813): p. 20200150.

108
227. Ramm, S.A., Sexual selection and genital evolution in mammals: A phylogenetic analysis
of baculum length. American Naturalist, 2007. 169(3): p. 360-369.
228. Hosken, D.J., et al., Is the bat os penis sexually selected? Behavioral Ecology and
Sociobiology, 2001. 50(5): p. 450-460.
229. André, G.I., R.C. Firman, and L.W. Simmons, The effect of genital stimulation on
competitive fertilization success in house mice. Animal Behaviour, 2022. 190: p. 93-101.
230. Tasikas, D.E., et al., Baculum variation and allometry in the muskrat (Ondatra
zibethicus): a case for sexual selection. Evolutionary Ecology, 2007. 23(2): p. 223-232.
231. Miller, E.H. and L.E. Burton, It’s all relative: allometry and variation in the baculum (Os
penis) of the harp seal, Pagophilus groenlandicus (Carnivora: Phocidae). Biological
Journal of the Linnean Society, 2001. 72(3): p. 345-355.
232. Miller, E.H., I.L. Jones, and G.B. Stenson, Baculum and testes of the hooded seal
(Cystophora cristata): growth and size-scaling and their relationships to sexual
selection. Canadian Journal of Zoology, 1999. 77(3): p. 470-479.
233. Schulte-Hostedde, A.I., J. Bowman, and K.R. Middel, Allometry of the baculum and
sexual size dimorphism in American martens and fishers (Mammalia: Mustelidae).
Biological Journal of the Linnean Society, 2011. 104: p. 955-963.
234. Ramm, S.A., L. Khoo, and P. Stockley, Sexual selection and the rodent baculum: An
intraspecific study in the house mouse (Mus musculus domesticus). Genetica, 2010.
138(1): p. 129-137.
235. Lupold, S., A.G. McElligott, and D.J. Hosken, Bat genitalia: allometry, variation and
good genes. Biological Journal of the Linnean Society, 2004. 83(4): p. 497-507.
236. Komori, T., Regulation of bone development and extracellular matrix protein genes by
RUNX2. Cell and Tissue Research, 2010. 339(1): p. 189-195.
237. Komori, T., et al., Targeted Disruption of Results in a Complete Lack of Bone Formation
owing to Maturational Arrest of Osteoblasts. Cell, 1997. 89(5): p. 755-764.
238. Zheng, Z., et al., Timing of androgen receptor disruption and estrogen exposure
underlies a spectrum of congenital penile anomalies. PNAS, 2015. 112(52): p. E7194-
E7203.
239. Chakraborty, P., et al., Androgen-dependent sertoli cell tight junction remodeling is
mediated by multiple tight junction components. Mol Endocrinol, 2014. 28(7): p. 1055-
72.
240. Rauch, A., et al., Glucocorticoids suppress bone formation by attenuating osteoblast
differentiation via the monomeric glucocorticoid receptor. Cell Metab, 2010. 11(6): p.
517-31.
241. Fedorov, A., et al., 3D Slicer as an image computing platform for the Quantitative
Imaging Network. Magn Reson Imaging, 2012. 30(9): p. 1323-41.

109
242. Dean, M.D., Genetic Disruption of the Copulatory Plug in Mice Leads to Severely
Reduced Fertility. PLoS Genetics, 2013. 9(1): p. 1-7.
243. Lough-Stevens, M., et al., Male-derived copulatory plugs enhance implantation success
in female Mus musculus. Biol Reprod, 2021. 104(3): p. 684-694.
244. Sutter, A., et al., Function of copulatory plugs in house mice: Mating behavior and
paternity outcomes of rival males. Behavioral Ecology, 2016. 27(1): p. 185-195.
245. Kelly, D.A., Anatomy of the Baculum-Corpus Cavernosum Interface in the Norway Rat
(Rattus norvegicus), and Implications for Force Transfer During Copulation. J Morphol,
2000. 244(1): p. 69-77.
246. Skonieczna, J., et al., Histological and morphometric evaluation of the urethra and penis
in male New Zealand White rabbits. Anat Histol Embryol, 2021. 50(1): p. 136-143.
247. McConnell, J., G.S. Benson, and J. Wood, Autonomic innervation of the mammalian
penis: a histochemical and physiological study. J Neural Transm, 1979. 45(3): p. 227-38.

110
Appendix
Appendix A. Male-derived copulatory plugs enhance implantation success in female Mus
musculus
My Contributions:
This particular study looked at how copulatory plugs in mice affect blastocyst implantation
success. When I joined the lab, Michael Lough-Stevens, PhD. was working on this project, and I
began helping him. I learned how to do several techniques, NSET injections, cardiac puncture,
and enzyme-linked immunosorbent assay (ELISA), that were crucial for this study. My
contributions to this study involved doing all of the cardiac punctures on the female mice to
retrieve blood to do hormone analysis. I also ran the majority of the ELISA hormone processing
on all of the blood samples used. For the purpose of having a double-blind analysis, I, alongside
some of the undergrads in the lab, checked for fertilization of the eggs from females that were
mated with wildtype animals and TGM4 knockout mice which cannot make a copulatory plug. I
also contributed to the editing of the manuscript. Taking into consideration the amount of work I
contributed to this study, I obtained second authorship on this published paper. The paper was
published in 2020.

111

112

113

114

115

116

117

118

119

120

121

122
Appendix B. Complex genetics cause and constrain fungal persistence in different parts of the
mammalian body.
My Contributions:
This study looked at the persistence of the yeast Saccharomyces cerevisiae’s pathogenicity in
different organs in a mouse model. My contributions to this project involved the majority of the
mouse work. I did the injections of the different yeast genotypes into each mouse via a tail
injection. After the injections were finished, I administered post-injection care to the mice
including intraperitoneal (IP) injections of antibiotics. Subsequently, I did the majority of the
mouse tissue (brain, lungs, spleen, liver, testes) dissections. After the mouse work was finished, I
also contributed to the manuscript by writing the mouse methods as well as contributing to the
editing of the manuscript. My contributions placed me as second author on this publication
which was published in 2022. Incredibly, this study has lead to a continuation of this project and
received funding from the army and NIH to continue this research.

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139
Appendix C. The lifespan of corpora lutea in non-pregnant females is positively correlated with
gestation length
My Contributions:
This study is looking at the lifespan of the corpora lute in female animals that are not pregnant.
This phenomenon should be something that doesn’t happen in mammals, yet for some reason,
the lifespan of corpora lutea in non-pregnant mammal females continues and is highly correlated
with gestation length. For this project female body size was a necessary covariate in the
correlation analysis. I contributed to this study by amassing all of the female body size data via
literature research. This paper which I am second author on is currently undergoing review at the
Journal of Mammalogy.

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163
Figure Legends
Figure 1. Corpus luteum lifespan in non-pregnant females was significantly positively correlated
with gestation length. Each point on the plot is a species, placed into one of six groups indicated
by color and image – the red kangaroo represents 23 species of metatherians; the black canid
represents 20 species of Order Carnivora; the green rodent represents 7 species of Order
Rodentia + Order Lagomorpha; the magenta deer represents 10 species of Order Artiodactyla +
Order Perissodactyla; the blue monkey represents 9 species of Order Primates; the orange
anteater represents 3 species of Order Cingulata + Order Pilosa. For both axes, the units are
ln(days). Species images are from phylopic.org (exact file names: PhyloPic.156b515d.SarahWerning.Callitrichoidea_Cebidae_Cebinae_Platyrrhini.png, PhyloPic.570c7d9e.Alexandra-vander-Geer.Rattus_Rattus-exulans.png, PhyloPic.6df900f7.Xavier-AJenkins.Myrmecophagidae_pan-Myrmecophagidae_Tamandua_Tamanduamexicana_Vermilingua.png, PhyloPic.96adba97.Margot-Michaud.Canis_Canis-simensis.png,
PhyloPic.c306572a.Sarah-Werning.Macropus-Macropus.png, PhyloPic.cc03f5c2.FerranSayol.Cervus-elaphus.png)
Figure 2. The phylogenetic distribution of corpus luteum lifespan in non-pregnant females.
Taxon colors and images as in Fig. 1. Branches colored according to the residuals of corpus
luteum lifespan regressed onto gestation length + body mass. For context, most recent common
ancestor of the species shown is roughly 157 million years ago.

164
Figure 1.

165
Figure 2.

Asset Metadata

Creator Ghione, Caleb Roberto (author)

Core Title Understanding the development of sexually selected traits

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Degree Conferral Date 2024-12

Publication Date 03/30/2025

Defense Date 06/21/2024

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

Tag baculum,Development,evolution,Mammals,OAI-PMH Harvest,osteoblasts,sexual size dimorphism,sexually selected traits

Format theses (aat)

Language English

Advisor Dean, Matthew (committee chair), Benayoun, Berenice (committee member), McMahon, Andrew (committee member), Nuzhdin, Sergey (committee member)

Creator Email calebghione@gmail.com,ghione@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC11399BCUC

Unique identifier UC11399BCUC

Identifier etd-GhioneCale-13558.pdf (filename)

Legacy Identifier etd-GhioneCale-13558

Document Type Dissertation

Format theses (aat)

Rights Ghione, Caleb Roberto

Internet Media Type application/pdf

Type texts

Source 20241001-usctheses-batch-1215 (batch), 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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

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

Repository Email uscdl@usc.edu

Abstract (if available)

Abstract This dissertation focused on studying two sexually selected traits – sexual size dimorphism (SSD) and the baculum (penis bone) found in many male mammals. Both character traits have been heavily studied, but their underlying genetics, mechanics, and functions are still shrouded in mystery. In my first chapter, I ask and answer the questions, 1) how does SSD develop and 2) how do two individuals of opposite sex deal with the conflict of body size differences when they share a similar genome? Via a meta-analysis, I test how androgen signaling might regulate SSD via their genomic architecture hormone-DNA binding sites. In my second chapter, I am interested in understanding the development of the baculum, a unique bone that is morphologically hyper diverse, has a unique evolutionary history of independent development/loss across mammal lineages, and is androgen sensitive as it develops. More specifically, I’m ask and answer the questions, 1) how does the baculum develop, 2) Are there any androgen binding sites near bone genes during the baculum’s development, 3) are the genetic and regulatory pathways between the baculum and other bone systems that do not have a unique evolutionary history like fore- and hindlimbs the same or different, and 4) can we disrupt the development of the baculum to understand its function better. I conclude my thesis by presenting a genetically engineered mouse via the cre-lox system that has a drastically developmentally reduced baculum as a novel model for studying the function of the baculum which has been disputed among scientists.