My Research

I am an evolutionary computational biologist, and I develop new methods to understand and estimate genetic admixture, relatedness, and population divergences. My larger interests are in the field of population genetics, in generating, analyzing, and distributing large-scale genomic data applied to conservation, biological control, agriculture, and human evolution. I also collaboratively work on genome/transcriptome projects, with particular interest in assembly, annotation, association, model-based analyses of demography, and molecular phylogenetics.


Inferring differential migration and genomic islands of speciation














Differentially introgressing loci are often implicated to be under some form of natural selection, and hypothesized to contain adaptively evolving regions, called "genomic islands". Previously, the inference of these islands has been contentious, with most researchers using summary statistics like Fst in genome-scans. In collaboration with Vitor Sousa, and Jody Hey, I am developing a parallel program for the inference of differential migration (across loci), under the Isolation with Migration (IM) model. We have now analyzed genome-wide patterns of differential migration in several species, including Anopheles gambiae subtypes, Heliconius melpomene wing-pattern morphs, and great apes (P.t. troglodytes versus P.t.verus) Watch this space for updates on this new software! This work is funded by an NIH-R01 grant to Jody Hey.


Parallel MCMC and Inference of Ancient Demography under the Isolation with Migration (IM) Model

I developed a parallel framework for ancestral demography inference under a Bayesian Markov Chain Monte Carlo (MCMC) framework. IMa2p has now been released, which distributes chains across multiple processors to efficiently compute divergence times, migration rates, and effective population sizes under the IM model. IMa2p can handle larger haplotypic data-sets (more individuals, more loci, more populations), and is efficient, with nearly linear improvements in computation time with number of processors used. As part of an ongoing collaboration with Sarah Tishkoff at the University of Pennsylvania, I am also currently analyzing the ancestral population demography of African Hunter-Gatherers. See our recent publication here, and download the source code, and instructions on how to install and run IMa2p at my Git page here. This work was funded by an NIH-R01 grant to Jody Hey.


Conservation Genetics of the Blanding's Turtle (Emys blandingii) in the Midwestern United States

We investigated the population genetics of the imperilled semi-box Blanding's Turtle by sampling 212 turtles from 18
Emys blandingii
populations to the West of the Mississippi. Genetic analyses using microsatellite markers revealed the presence of considerable genetic structure among all sampled locales, and the presence of ancestral gene flow in this region north and east from an ancient refugium in the central Great Plains, concordant with post-glacial recolonization time-scales. For more details, see the publication in Conservation Genetics. This work was supported by several grants to my PhD adviser, Fred Janzen, and a Departmental grant and a James Cornette Fellowship to me.



Population Genomics of the convergent ladybird beetle (Hippodemia convergens) across the Americas
Hippodemia convergens

The predatory convergent ladybird beetle, Hippodemia convergens has been utilized as an insect parasitoid and populations of beetles have been artificially introduced into locations across the Americas. This study focuses on analysis of genetic admixture and patterns of migration of these beetles (either naturally or artificially) supplanted from a potential source population in California, and is currently being undertaken in collaboration with John Obrycki at the University of Kentucky, and several undergraduate students in the Janzen lab. See our recent publication in Biological Control here. Furthermore, we are interested in studying the effects of (a) pervasive inbreeding, (b) cessation of admixture from source populations, and (c) resource competition with other hetero-specific beetles - Harmonia axyridis, and Coccinella septempunctata.This work was funded by a USDA-NIFA grant to John Obrycki, and an Entomological Society of America travel grant to me.


Fast Multinomial Clustering of Population Genetic data to delineate genetic admixture and population structure

In collaboration with my co-adviser, Karin Dorman, and Wei-Chen Chen (former graduate student, currently a post-doc at Oakridge National Laboratories), I developed a likelihood based method for understanding pop
ulation structure and detecting genetic admixture. The method, MULTICLUST (Sethuraman 2013, Sethuraman et al. in prep), features a robust framework for (a) detecting admixture proportions, (b) testing hypotheses of different models of population structure, (c) accelerated runs of an EM algorithm for several genetic data-types, including SNPs, microsatellites, and allozymes, and (d) visualizing admixture plots on the landscape through pie plots in R.


Maximum Likelihood Estimation of Pairwise Genetic Relatedness in Admixed Populations

The presence of population genetic structure biases traditional estimates of
pairwise genetic relatedness by failing to capture signals owing to localized allele frequency distributions. I developed a maximum likelihood based method to estimate pairwise genetic relatedness that accounts for the presence of population structure by utilizing admixture proportions to capture existing diversity in subpopulations. I compare the bias in estimation of pairwise relatedness using the new method with 10 other methods that have been previously developed for the purpose, including the methods of Queller and Goodnight (1989), Lynch and Ritland (1999), and Anderson and Weir (2007). A manuscript on the same is currently under preparation, in collaboration with my PhD co-adviser, Karin Dorman.



Comments