My research aims to discover more about the processes that lead to the genetic basis of phenotypic and genetic variation in order to understand the evolution of adaptation. During my graduate work, I utilized wolves as a study system for this effort, but I have also worked on evolutionary genetics of dog domestication, phylogeography of masked flowerpiercers, and population genetics of ringtails. Through these research experiences I have realized the power of combining cutting-edge molecular techniques with field-based research to understand ecology and evolution of natural populations.
Read more below and see my publications.
Local adaptation in North American gray wolves
For my PhD, I aimed to uncover and characterize genetic mechanisms of adaptive variation. Wild canids are an excellent system in which to study how non-model organisms adapt to their environment since they are found throughout the world and have diverse morphological, physiological, and behavioral adaptations. Additionally, novel genomic resources have been developed for many canid species and the well-annotated dog genome is available. At UCLA, I had a golden opportunity to conduct research in this system since Dr. Bob Wayne is an expert on canid evolution, ecology, and conservation, and many of the necessary samples and close ties with collaborative groups were readily available.
Using previously collected genotype data for 44,000 single nucleotide polymorphisms (SNP) on the Affymetrix dog SNP array, I analyzed genetic variation in 123 North American wolves. This involved using several analytical tools to survey population structure. To study selection, I used traditional measures of population differentiation (e.g. FST) as well as newer methods developed for human studies also applicable to wolves (e.g. XP-EHH) to identify statistical “outlier” regions putatively under selection among different ecotypes. I also used software to test whether specific SNP allele frequencies co-vary with environmental variables above a neutral expectation, which would imply that these SNPs may be linked to genes that have undergone positive selection in response to the environment.
Using these SNP data and individual-level measurements of 12 environmental variables, I identified six ecotypes: West Forest, Boreal Forest, Arctic, High Arctic, British Columbia, and Atlantic Forest. I found consistent signals of selection on genes related to morphology, coat coloration, metabolism, vision and hearing. My findings showed that local adaptation can occur despite gene flow in a highly mobile species and can be detected through a genomic scan with a moderate number of SNPs. These patterns of local adaptation revealed by SNP genotyping likely reflect high fidelity to natal habitats of dispersing wolves, strong ecological divergence among habitats, and moderate levels of linkage in the wolf genome.
As a continuation of the above study, I designed a targeted capture of 1040 genes, including all exons and flanking regions, as well as 5000 1 kilobase non-genic neutral regions and re-sequenced these regions in 107 wolves. Using tests that infer selection, I identified potentially functional variants related to local adaptation. By integrating these data with results from my first chapter, I demonstrated that combining data from genome wide SNP genotyping arrays with large-scale re-sequencing and environmental data provide a powerful approach to discern candidate functional variants in natural populations.
The above results were published in two first-author papers in early 2016 in Molecular Ecology.
Selection on black coat color in wolves
I am especially interested in the evolution of coat coloration in mammals, and especially in wild canids. In many species, natural pigment variation is controlled by the Agouti- Mc1r pathway and many melanistic phenotypes are caused by a mutation in that pathway. Melanism in the gray wolf is caused by a mutation in the K locus. A previous study suggested that the melanistic KB allele (a dominantly-inherited 3bp deletion) was introduced into the genome of North American wolves from the domestic dog via interbreeding, and then underwent positive selection.
I designed a custom capture array to re-sequence five megabases surrounding the K locus core deletion in a larger sample of North American wolves from multiple areas to assess patterns of nucleotide and haplotype diversity, population-specific decay in linkage disequilibrium, and hierarchical patterns of genetic divergence among populations. I am currently preparing a manuscript for submission in the next few months.
Through a collaboration with Dr. John Novembre and his group at the University of Chicago, we are utilizing data from a multi-generation Yellowstone pedigree to calculate wolf-specific mutation and recombination rates. These sequences will provide a definitive estimate of mutation rate in wild wolves, a value that is the essential parameter in determining the evolutionary rate of genes and how selection alters the genome. The mutation and recombination rates will also enable us to construct locus-specific population models to understand selection at the K locus for black coat color.
Demographic history and selection during dog domestication
Despite the domestic dog being our best friend, we know relatively little about the who, what, or when of domestication. Through a multi-institutional collaboration headed by Dr. Adam Freedman (Harvard), Dr. John Novembre (University of Chicago), and Dr. Robert Wayne (UCLA), we used whole-genome sequencing to clarify the demographic history of dog domestication and incorporate this demographic history into selection scans to uncover the genes that make dogs so different from their wild wolf ancestor.
We found that dogs and wolves likely diverged 11-16kya, before the advent of agriculture, through a dynamic process of bottlenecks in both lineages, with high levels of admixture. The wolves from which dogs were domesticated were more diverse than present day wolves, and no extant wolf population we sampled was more closely related to dogs. We also scanned the genomes of dogs to find regions that showed evidence of selection, and used our previously constructed demographic model to estimate a false discovery rate. Of the top 100 regions showing positive selection, many included genes involved in behavior, brain function, and lipid metabolism. These genetic changes were likely advantageous for wolves living alongside hunter-gatherer human populations.
These results were published in two papers, led by Adam Freedman, in PLoS Genetics in 2014 and 2016.
High altitude adaptation and evolution of metabolic networks in deer mice
My proposed postdoctoral research in the Cheviron lab at the University of Montana will focus on high altitude adaptation and evolution of metabolic pathways. I will use deer mice as a model species to pursue questions about how animals adapt to high altitude, a question that has long been of interest to biologists. The survival of mice depends on their performance at different altitudes, and genetic patterns indicate natural selection among populations living at different altitudes. Differentiation of the hemoglobin molecule (which animals use to absorb oxygen) in red blood cells between low- and high-altitude populations of deer mice is a well-studied case of genomic and physiological adaptation. The process of how oxygen is subsequently transported through the animal’s tissues is one example of a metabolic network. My research will examine the metabolic networks that control multiple physical traits involved in how deer mice adapt to high altitude to describe genetic patterns that are important to adaptation. My work is funded by an NSF Postdoctoral Research Fellowship in Biology.