Why complexity persists: Evolutionary dynamics of the amylase locus in primates

The amylase locus is among the most structurally variable regions of the human genome, frequently linked to starch digestion, metabolic traits, and dietary adaptation. Yet the causes of its recurrent duplication and exceptional variability remain unresolved. Why is this locus particularly prone to structural change? To address these questions, we analyzed 98 modern human genomes using long-read sequencing and optical mapping, alongside 53 high-quality primate assemblies. We identified 30 distinct amylase haplotypes in humans and documented more than 15 lineage-specific expansions and contractions across primates. Structural complexity appears to have been initiated by lineage-specific LTR insertions and subsequently shaped by non-allelic homologous recombination, with occasional contributions from microhomology-mediated break-induced replication. Independent duplications and salivary expression gains evolved repeatedly across primate lineages, but extensive within-species structural polymorphism is largely unique to humans. We further detected signatures of positive selection among primate paralogs, and dietary correlations with copy number suggest recurrent adaptive roles for amylase variation. The persistence of structural variation in this locus points to a unique combination of elevated mutational input, relaxed constraint, and ongoing selection, highlighting broader principles in the evolution of structurally unstable loci.