Inference of selection acting on coding sequences in the human lineage

Selection forces have shaped the current human DNA sequence through evolution. These forces have been consistently changing, in magnitude and direction, across the different evolutionary episodes of the extant human lineage. Understanding and quantifying these forces helps us understand the genomic basis of human diseases, among other benefits. Tools that quantify these forces include two approximations: parallel mutations (where identical by state mutations are not identical by descent) and multiple substitutions (where a mutated site cannot re-mutate) are not modelled. In this work, we use multiple sequence alignments of primates to quantify the selection forces acting on the human lineage. Moreover, we present novel approaches to incorporate parallel mutations and multiple substitutions in the model. Finally, we use the modelled parallel mutations to estimate effective population sizes for historical episodes of evolution.