Exploring the role of local sequence context on patterns of germline mutation

Abstract

Germline mutations occur at a non-uniform rate across the genome. While the influence of neighboring nucleotides is well known and accounted for in mutation models, the effective range of these flanking influences and the extent of their higher-order interactions remain insufficiently understood. Here, we leverage millions of singleton variants from the 1000 Genomes Project as a proxy for \textit{de novo} variants to systematically map the contextual influence of local nucleotide sequence context on germline mutation rates. We show that while nucleotides immediately adjacent to the mutation site exert the strongest effect, the influence of nucleotides at flanking positions does not decay monotonically with increased distance from the focal position. Notably, significant associations between flanking nucleotides and mutation status persist up to hundreds of base pairs from the focal site. Furthermore, higher-order interactions among flanking positions significantly modulate mutation rates, though their effect size diminishes as the order of interaction increases. Additionally, these mutational signatures are highly consistent across diverse ancestral groups. Together, these findings provide a comprehensive map of how local genomic architecture shapes the mutational landscape.