Genome-wide admixture is common across the Heliconius radiation

Abstract

How frequent is gene flow between species? The pattern of evolution is typically portrayed as a phylogenetic tree, implying that speciation is a series of splits between lineages. Yet gene flow between good species is increasingly recognized as an important mechanism in the diversification of radiations, often spreading adaptive traits and leading to a complex pattern of phylogenetic incongruence. This process has thus far been studied in cases involving few species, or geographically restricted to spaces like islands, but not on the scale of a continental radiation. Previous studies have documented gene flow, adaptive introgression and hybrid speciation in a small subsection of the charismatic Neotropical butterflies Heliconius. Using genome-wide resequencing of 40 out of 45 species in the genus we demonstrate for the first time that admixture has played a role throughout the evolution of Heliconius and the sister genus Eueides. Modelling of phylogenetic networks based on 6848 orthologous autosomal genes (Maximum Pseudo-Likelihood Networks) or 5,483,419 high quality SNPs (Ancestral Recombination Graph) uncovers nine new cases of interspecific gene flow at up to half of the genome. However, f4 statistics of admixture show that the extent of the process has varied between subgenera. Evidence for introgression is found at all five loci controlling the colour and shape of the mimetic wing patterns, including in the putative hybrid species H. hecalesia, characterised by an unusual hindwing. Due to hybridization and incomplete coalescence during rapid speciation, individual gene trees show rampant discordance. Although reduced gene flow and faster coalescence are expected at the Z chromosome, we discover high levels of conflict between the 416 sex-linked loci. Despite this discordant pattern, both concatenation and multispecies coalescent approaches yield surprisingly consistent and fully supported genome-wide phylogenies. We conclude that the imposition of the bifurcating tree model without testing for interspecific gene flow may distort our perception of adaptive radiations and thus the ability to study trait evolution in a comparative framework.