All detectable ancient whole-genome duplications involve hybridization

Abstract

Whole-genome duplication (WGD), or polyploidy, is a major evolutionary force throughout the tree of life. Ancient WGDs are commonly inferred from the distribution of synonymous substitutions per synonymous site (Ks) between duplicated genes. Here, we argue that the standard interpretation of these distributions is incomplete. Using haplotype-phased genome assemblies spanning canonical autopolyploid and allopolyploid systems, we find that Ks accumulates primarily between gene copies on non-recombining chromosomes. Thus, Ks-based WGD detection depends not simply on WGD, but on whether duplicated copies evolve independently. This distinction reframes many inferences of ancient WGD as signatures of hybridization and the evolution of meiotic isolation and alters how we interpret the origin, detectability, and evolutionary consequences of polyploidy across the tree of life. Significance Whole-genome duplication (WGD) is prevalent across the tree of life, and its signatures appear in nearly every sequenced genome. Here we show that the predominant gene-based approach for detecting ancient WGD events is biased towards allopolyploid events arising through hybridization. The current approach is poorly suited to detecting WGD that occurs within a single species, also known as autopolyploidy. These findings suggest we likely underestimate the frequency of ancient polyploids. Furthermore, most signals of ancient WGD may therefore record not only genome doubling, but also the merger of previously divergent evolutionary lineages. Overall, our results suggest that the paleopolyploidy record inferred from current methods is an incomplete account of WGD and reveals deep hybridization preserved in eukaryotic genomes.