Yun S. Song

This study compares the accuracy of selection coefficient estimates from time-series genetic data with and without accounting for population size history, finding that ignoring drift does not significantly reduce accuracy.

This paper uses deep learning to infer population genetic parameters and detect selection signals in Drosophila melanogaster genomes, finding more soft sweeps and balancing selection near centromeres and revised demographic bottlenecks.

The paper develops methods to compute the sample frequency spectrum under selection and variable population size, and applies them to assess the effects of selection on demographic parameter estimation and detect negative selection in growing populations.

diCal-IBD detects identity-by-descent tracts between unrelated individuals with higher recall and precision than existing methods.

This paper introduces an efficient method for inferring population size histories and locus-specific mutation rates from large genomic datasets, validated on simulated and real data.