INTRODUCTION
For centuries, species have been delineated based on their physical appearance. Early naturalists differentiated between species based on functional traits such as jaw and limb shape, and beak size, or otherwise on traits that function in social selection, such as fur pattern and plumage coloration (Birkhead, 2008; Leroi, 2014). Such phenotypic characters, representing clear visual differences to the human eye, have been thought to aid the species themselves in recognising their own kind and finding potential mates. It has long been of interest to identify which traits mediate reproductive isolation between related species (Mayr, 1942), and more recently, which genes underpin those traits; in other words, the genes that mediate speciation (Nosil & Schluter, 2011). Hybrid zones between related species provide an unparalleled opportunity to identify genes associated with phenotypic characters (Delmore et al., 2016; Brelsford et al., 2017), including those that play an important role in mate choice and species recognition (Harrison & Larson, 2016). In some such cases, it is just a small set of genes that maintain phenotypic differences between hybridising taxa, especially those functioning in coloration and pattern differences (Poelstra et al., 2014; Toews et al., 2016). And evidence is starting to mount that the extent of divergence in plumage coloration between populations affects assortative mating (Haas et al., 2009; Scordato et al., 2017; Billerman et al., 2019).
Genes involved in coloration and patterning pathways have received a lot of attention (Mundy, 2005; Hoekstra, 2006), with a number of recent studies identifying genes associated with the melanin pigmentation pathway as a result of research across hybrid zones (e.g., (Delmore et al., 2016; Toews et al., 2016; Walsh et al., 2016; Campagna et al., 2017). Carotenoid-based coloration, incorporating the reds, oranges and yellows, has received far less attention (Toews et al., 2017), because carotenoid uptake is dependent on diet, meaning that identifying the genes involved in carotenoid processing is especially challenging (Campagna et al., 2017). Nevertheless, much progress has been made recently with the scavenger receptor B1 (SCARB1 ) identified as a mediator of carotenoid-colour expression in birds (Toomey et al., 2017), while a region of the Z chromosome has been identified as a potential regulatory region for follistatin and is associated with sex-linked red colour polymorphism in Gouldian finch Erithrura gouldiae (Kim et al., 2019). Previous work has also identified candidate genes to explain differences between red and yellow feather coloration through lab crosses of yellow canaries with red siskins (Lopes et al., 2016) and red and orange billed individuals of zebra finches (Mundy et al., 2016). Both studies specifically identified the gene CYP2J19 as the one that encodes the ketolases that catalyze the conversion of yellow dietary carotenoids to red C4-ketocarotenoids. Both studies were focused on laboratory-based crosses involving artificially selected forms of species developed to express certain traits (“red factor” canaries and yellowbeak zebra finches respectively). A field-based study in long-tailed finch Poephila acuticauda provided further support for the role of CYP2J19 in carotenoid-based bill colour across a natural hybrid zone (Hooper et al., 2019), while CYP2J19 gene expression has also been associated with red coloration in weaverbirds (Twyman et al., 2018b). But evidence that it explains variation between red and yellow feather coloration in wild populations of birds is still equivocal.
Red-fronted (Pogoniulus pusillus, Dumont, 1816) and yellow-fronted (P. chrysoconus, Temminck, 1832) tinkerbirds provide such an opportunity to investigate the genes involved in carotenoid processing because they hybridise extensively at a contact zone in southern Africa (Nwankwo et al., 2019). They differ phenotypically most noticeably in their forecrown coloration; hence their vernacular names (Fig. 1). Their other plumage differences are also in carotenoid-based colours, but those are more subtle differences, to the human eye at least, between shades of yellow and white that characterize the colour of their underparts, rump, wing bar, and supercilium (Hockey et al., 2005). The two species have been shown to hybridise extensively with asymmetric introgression of red forecrown plumage into the P. chrysoconus genomic background, in spite of phylogenetic analysis on mitochondrial DNA revealing the two species have been diverging for over four million years (Nwankwo et al., 2019). Despite the deep divergence between the species, the extent of introgression indicates that hybrids are viable and fertile, though some reduced hybrid fitness might maintain a tension zone and inhibit the two species collapsing into one (e.g., Kearns et al., 2018). Thus, if at least one species continues to differentiate the two phenotypes and mates assortatively, the plumage trait and the genes underlying it may play an important role in maintaining species divergence.
Here we investigate the genomic regions associated with the red and yellow feathers in red-fronted and yellow-fronted tinkerbirds that could play an important role in reproductive isolation. We sampled 79 individuals across the hybrid zone in southern Africa, collecting forecrown feathers for spectrophotometric analysis (Nwankwo et al., 2019). We performed a genome-wide association study (GWAS) to identify single nucleotide polymorphisms (SNPs) associated with plumage coloration. The study included assembly of a short-read genome of an individual Pogoniulus pusillus and double-digest restriction association DNA sequencing (ddRAD). Our aim was to identify the genes associated with carotenoid plumage coloration across a natural contact zone, and specifically which regions of the genome explain differences between red and yellow plumage. We further tested whether those genes, or SNPs at other loci, might be associated with the shade of yellow to white or extent of yellow versus black plumage in other feather patches, while also testing for any differences in plumage attributable to sex. We then compared alleles at loci associated with forecrown colour with the hybrid index across autosomes and the Z chromosome, and interspecific heterozygosity of individuals, to elucidate the patterns of interbreeding across genotypes that explain asymmetric introgression between species that have been diverging for over 4 million years (Nwankwo et al., 2019).