Activity: Talk or presentation typesPresentation


The interplay between mutation, genetic drift, directional, and balancing selection in shaping populations’ diversity is highly convoluted and difficult to disentangle. This requires sophisticated phylogenetic models that have a high degree of flexibility and can handle multi-individual data. For these purposes, our group has developed a polymorphism-aware phylogenetic set of models called PoMos. These models are based on the Moran model and have recently been proven effective in inferring species trees as well as mutational effects, fixation biases and GC-bias rates in great apes and grasshoppers. To make these models more accessible, we implemented them in the open-source Bayesian inference framework RevBayes. The advantage of the framework is its implementation in a graphical model environment and the possibility to compute coverage frequencies for the validation analysis. In this study, we further developed PoMos to study neutral, directional and, for the first time, balancing selection. The key advantage of our novel approach for studying the balancing selection is that PoMos allow for ancestral polymorphisms that can be maintained, and parameters that can measure frequency-dependent selection. We have tested our new method on a set of simulated data with a popular evolutionary framework Slim and a custom Moran model simulator implemented in RevBayes. Furthermore, we investigated real sequences of African human populations to understand the evolutionary history of genomic regions that are known to be under balancing selection driven by malarial parasites.
Period11 Mar 2023
Event titleProbabilistic Modeling in Genomics
Event typeConference
LocationCold Spring Harbor, United States, New YorkShow on map
Degree of RecognitionInternational