Evolutionary History of Red Locus Haplotypes

To infer the evolutionary history of the haplotypes at the Red locus we constructed gene genealogies and compared the level of evolutionary divergence between the black and red haplotypes with those of orthologous sequences in other species. First, blood samples of 14 species in the family Estrildidae including red-faced (E. pealii, E. psittacae) or blue-faced (E. tricolor and E. trichroa) Erythrura species were obtained from an aviculturist. Blood samples were collected and stored using Whatman FTA cards (GE Healthcare, Little Chalfont, UK) and DNA was extracted following the manufacturer’s instructions. Sequences of the fragments amplifiable using the M1–M3 primers (Supplementary Table 2) were obtained using Sanger sequencing. For zebra finch, the sequences from the reference genome assembly were used.
We tested alternative nucleotide substitution models using the R package, phangorn v2.3.1^{78 (link)}, which suggested GTR+Γ+I to be the best model for the sequence of the Red locus. Maximum-likelihood trees were then constructed using this model and tested with 1000 bootstraps using phangorn and visualised using ggtree v1.4.20^{79 (link)}. To examine if the molecular clock hypothesis applied to the Red locus, we used Tajima’s nonparametric relative rate test implemented in pegas v0.10^{80 (link)}.