Comparative genomics of two Empidonax flycatchers reveal candidate genes for bird song production.

Abstract

Whole-genome level comparisons of sister taxa that vary in phenotype against a background of high genomic similarity can be used to identify the genomic regions that might underlie their phenotypic differences. In wild birds, this exploratory approach has detected markers associated with plumage coloration, beak and wing morphology, and complex behavioral traits like migration. Here, we use genomic comparisons of two closely related suboscine flycatchers (Empidonax difficilis and E. occidentalis) and their hybrids to search for candidate genes underlying their variation in innate vocal signals. We sequenced the genomes of 20 flycatchers that sang one of two species-specific parental song types and 14 putative hybrid individuals with intermediate song types. In the resulting genomic comparisons, we found six areas of high differentiation that may be associated with variation in non-learned songs. These narrow regions of genomic differentiation contain a total of 67 described genes, three of which have been previously associated with forms of language impairment and dyslexia in humans, and 18 of which are known to be differentially expressed in the song nuclei regions of the avian brain compared to adjacent parts of the avian brain. This 'natural experiment' therefore may help identify loci associated with song differences that merit further study across bird lineages with both learned and innate vocalizations.