#32 · Genomics

Genome-scale tests of selection theory and diagnostics for dated and Bayesian phylogenies

The common thread is testing whether inference frameworks in population and phylogenetic genetics hold up against data. The nearly neutral theory study links πN/πS and dN/dS to diversity and Ne across 150 mammals using 6000 orthologs, providing empirical validation at divergence and population-genomic scales. The seasonal fluctuating selection study is simulation-based, characterizing its genomic footprints in diversity, SFS, and haplotype patterns relative to sweeps and heterozygote advantage. DiagnoDating proposes outlier detection, posterior predictive checks, and residual analysis to assess dated phylogenies in microbial population genetics. The bgphy package implements Bayesian Population Monte Carlo for mixed Gaussian trait-evolution models with posterior predictive evaluation. What is new is the emphasis on checkable assumptions: DiagnoDating supplies diagnostics for dated trees, and the selection study distinguishes seasonal fluctuating selection from established forms by its footprint.