Files

Abstract

Advances in genetic sequencing technology have given us an abundance of whole-genome sequencing data that can tell us much about evolutionary processes and population history. Coalescent hidden Markov models (CHMMs) are a powerful class of methods that use both genetic variation and genetic linkage information to untangle the complex demographic his- tories of natural populations. This thesis presents CHIMP (CHMM History-Inference ML Procedure), a novel CHMM implementation that can use both the height (TMRCA) and the total branch length (L) of the underlying genealogical tree as the latent variable in the HMM as the method moves sequentially along the genome. The primary application of CHIMP is in demographic inference problems, and we perform a suite of simulations to benchmark the performance of CHIMP among other state-of-the-art CHMMs. We also demonstrate the use of CHIMP to perform demographic inference in structured populations. Finally, we introduce CHIMP-PD, an extension that is used to decode the posterior probability of the CHMM, and explore its use in uncovering patterns of adaptive variation. This work ultimately demonstrates that CHIMP provides a flexible, efficient alternative to other methods, particularly when analyzing unphased and pseudohaploid data.

Details

Actions

PDF

from
to
Export
Download Full History