|
Abstract:
|
Population bottlenecks can cause dramatic shifts in allele frequencies, due to drift and, in some cases, forces such as selection. Discerning between drift and selection is essential to understanding functional consequences of the variants. In this study, we model bottleneck effects with a modified Moran process. This corresponds to a discrete Markov chain with two transition matrices for the bottleneck and subsequent growth. Constructing a closed-form equation, we calculate the probability of observing the shift in allele frequency under the null hypothesis of genetic drift alone. When this probability is significantly small, we incorporate and estimate a selection coefficient. We apply this approach to next generation sequencing in 69 cell populations isolated from subjects at the National Institutes of Health. We show we can adjust our model to the experimental procedure, including one extreme bottleneck, followed by five minor bottlenecks. We therefore present a flexible, probability-based approach to directly modeling the biological process of population bottleneck and growth and identifying variants with a selection advantage.
|
