The abstract is read at the publisher; the summary is JClub's.
A GPU implementation, gamma_smc_cu, estimates pairwise coalescence times for 63,190 haplotype pairs in 15.91 seconds with accuracy comparable to the CPU reference.
- Why it matters: Pairwise coalescence variation can reveal selection-linked genealogical differences, but existing computation is too slow for genome-wide, population-scale scans.
- What they did: The method was applied to the 1000 Genomes Project to generate gene lists with low coalescence time across 26 populations and a catalog of 143 regions, then evaluated using genomic covariates and benchmarks versus iHS and XP-EHH.
- The result: The scan highlights regions with selection-like coalescence patterns, including case studies at GRK2 and TREML1/TREM2, and the authors show that both positive and purifying selection can yield similar inferred local signals.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.