Many species hybridize with other species, and this can lead to wholesale transfer of variation important in adaptive evolution. In some cases, this sudden acquisition of recombined adaptive traits can lead to the origins of new species, or "hybrid speciation." These processes were long recognized as important in plants, but increasingly are being discovered in suitable groups of animals. Our group, working with Heliconius, ithomiines, and other insects, has been at the forefront of this recent research. Around 35% of Heliconius are known to hybridize and backcross with at least one other species. Although many species hybridize, the fraction of individuals that hybridize is always low, or the species would disintegrate completely. Up to about 0.1% of specimens in sympatric species pairs are hybrids. In contrast, for species like H. himera and H. erato that are in parapatric contact, the rates may be up to 5% in populations from the narrow hybrid zones where the species overlap.
Using ABBA-BABA tests first applied to the study of human and Neanderthal genomes, we have found examples of the transfer of mimicry genes among sister lineages, such as between Heliconius melpomene and H. timareta. There are also longer range transfers between non-sister lineages, as between Heliconius melpomene and H. elevatus. Detailed studies of H.cydno and H. timareta have revealed that a substantial fraction of regions from across the genome may be admixed with H. melpomene (30-40% or more).
In the future, we aim to characterize the nature of transferred genomic regions across the genome, to determine how much of this is controlled by selection.
It is rather unlikely that hybridization "adaptive" in itself: most F1 and backcross hybrids suffer from hybrid sterility, other genetic incompatibilities, and sexual selection against hybrids. Hybrids are often poorly fitted to the ecological niche of either parent; for example, mimicry or host plant adaptation may be compromised. Because it is largely counter-adaptive, the rate of interspecific hybridization is probably selected to be minimized reinforcement). However, once this rate is reduced this low, the impact on the parental populations are minimized, and the hybridization rates become effectively neutral, leading to lack of further reinforcement. A mating system that avoids all interspecific mating might have negative consequences for sexual selection within species, and thereby could lead to a balanced, although low rate of hybridization. This remains to be investigated.