I'm an evolutionary biologist who studies how populations move across landscapes and adapt to varying environments. To do that I build computational tools and statistical pipelines that combine ecological and genetic data both from my own fieldwork and public datasets. Some of my recent work has identified cryptic species of migratory birds, mapped the migratory routes of declining North American songbirds, and documented exponential population growth in hummingbird populations.
I've also developed interactive simulation webapps to help students learn the basics of population genetics, and worked with my collaborator Ethan Linck to study how bioinformatic processing affects inference of population structure from genetic data.
I'm currently a postdoctoral researcher in the Kern Lab at the University of Oregon Institute of Ecology and Evolution, where I'm developing new simulation tools and machine learning methods to improve our ability to identify the genetic basis of trait variation in species that are structured over space, including mosquitoes and humans. I did my PhD in the Klicka Lab at the University of Washington Department of Biology and the Burke Museum of Natural History, where my dissertation was focused on the impacts of seasonal migration on genetic diversity in birds.
Scroll down for more info on my research, links to code, and pictures of animals.
Standard models in population genetics are based on randomly mating populations,
but real organisms mate with nearby individuals and disperse a limited distance
from their parents. In the Kern Lab one of my primary goals has been to
develop new simulation and inference tools that let us model genetic variation in
continuous space. This is one of the oldest challenges in population genetics,
which was *almost* solved by Wright and Malecot in the
early twentieth century. Unfortunately their models of isolation by distance are inherently
flawed because they fail to account for an important ecological process -- density-dependent
population growth -- which prevents individuals from clustering (see Felsenstein's excellent
1975 Am Nat article "A Pain in the Torus: Some Difficulties with Models of Isolation
By Distance" for the full details).
Working with Andy Kern and Peter Ralph, we have developed new forward-time simulators
that allow us to generate chromosome-scale alignments of tens of thousands of individuals
evolving in continuous space, and we are using these tools to investigate how
inference from genetic data is shaped by spatial processes. The figure at the top of
this section shows an early result of our simulator -- mapping how dispersal distance
determines the geographic spread of ancestry over time.
This research involves a lot of esoteric modeling and programming, but it has important implications for human health and society. When researchers attempt to connect genomes and traits -- for example, by trying to identify genes responsible for insecticide resistance in mosquitoes or heart disease in humans -- the standard models in use today assume that populations are exposed to similar environments. But when populations are structured over space, this is unlikely to be true. The figure below is a preview of a new study we're working on to show how current methods of correcting genome-wide association studies (GWAS) for population structure can fail when breeding structure and the environment covary over space (preprint coming soon!). We need better methods to model the interaction of geographic population structure and spatial variation in the environment if we are going to understand how genotypes connect to the complex phenotypes that matter to efforts like public health and vector control. That's why we're now developing deep learning methods to co-estimate phenotype associations and population structure, which will be my focus over the next year.
Around 30% of bird species migrate seasonally between different
habitats, and similar migratory behaviors are found in (among others)
butterflies, insects, fish, mammals, snakes, flatworms, fruit flies, and
people. My dissertation research is focused on understanding how this process
impacts speciation and the capacity for local adaptation in birds.
Historically most migratory species were thought to be genetically homogenous,
because spatial mixing of genotypes between years should spread genetic
variation widely across the range. My research on vireos (pdf), buntings (pdf), and hummingbirds (working on
it) has found that a strong correlation between geographic and genetic
distance (i.e. "isolation by distance") is instead found even in small bodied
species that migrate without family groups, suggesting effective gene flow
across the range is relatively low despite the species' large annual
movements. Introgression among strongly divergent lineages is frequently
observed at range boundaries, but at least in the species I have studied
rarely spreads to the center of the range. This combination of IBD and
persistent hybrid clines suggests that selection plays an important role in
maintaining population differentiation in migratory birds.
How does the interaction of gene flow, drift, and selection shape variation across the genome of migratory species? How much information about past demographics or selective regimes can be reliably inferred from genomic data? I'm currently working on a set of simulation studies and a whole-genome sequence analysis in order to address these questions in the Rufous/Allen's Hummingbird Species Complex.
"All species ranges are the result of successful past range
expansions" - Keitt et al. 2001, Am Nat.
I'm interested in how species ranges change over time, and how human modification of the landscape has shaped their evolution over the last hundred years. Recently I've analyzed two cases (both currently in review): a drop in elevation ranges in Puerto Rican Anolis lizards likely caused by forest regrowth on former agricultural lands during industrialization, and the dramatic northern range expansion of the Anna's Hummingbird caused by introduced plants and hummingbird feeders.
For recent updates see my github
driftR: an interactive population genetic simulation website that allows students to explore the impacts of genetic drift, selection, migration, mutation, and population sizes on a variety of summary statistics.https://cjbattey.shinyapps.io/driftR/
adaptR: simulate selective sweeps and other processes with varying selection over time.https://cjbattey.shinyapps.io/adaptR/
structurePlotter: plot output of genotype clustering algorithms with fancy color selection and a permutation algorithm to deal with label switching.https://cjbattey.shinyapps.io/structurePlotter/
A nonrandom subset of recent pictures. Find more on my tumblr
Postdoctoral Researcher, Kern Lab
University of Oregon Institute of Ecology and Evolution
301 Pacific Hall
Eugene, OR 97402