The red-browed finch Neochmia temporalis occupies diverse habitats, from monsoonal tropics to temperate forests. To study the genomics of local climate adaptation, we performed whole genome sequencing of 60 N. temporalis, to 2x per specimen. We justify our use of WGS by demonstrating that it is a powerful population genomics tool, with the potential to also reveal functional targets of selection.
Entire mitochondrial genomes were reconstructed from WGS data for all specimens. Phylogenetic analysis distinguished the northern subspecies, N. t. minor, from its southern counterpart, N. t. temporalis. Within N. t. temporalis, mitochondrial divergence was low (0.27%), mostly comprising singletons, and exhibited a star-like phylogeny, suggesting a recent rapid expansion in temperate habitats. Mitochondrial divergence between the two subspecies was high (2%). We found no evidence for mitochondrial selection.
To facilitate genome wide scans for selection, reads from each subspecies were pooled separately and population-level SNPs called. A parallel analysis of two other widespread finches, Stizoptera bichenovii and Lonchura castaneothorax, was also performed. Concordant islands of differentiation between the three species may represent genomic regions underlying local climate adaptation.
We also attempted individual-level SNP analysis, arguing that due to biases in read generation, some genomic regions are likely to have sufficient coverage to call variants in most specimens. Around 500 nuclear SNPs were obtained in this proof-of-principle analysis. Clustering of these SNPs identified N. t. minor as per the mitochondrial analysis, verifying the utility of this approach.
Intriguingly, the individual-level SNPs also resolved a second subspecies, N. t. loftyi, endemic to South Australia. Discordance with the mitochondrial results suggests a role for selection, rather than neutral processes, in the differentiation of N. t. loftyi and N. t. temporalis. The divide between N. t. loftyi and N. t. temporalis also represents the first direct molecular evidence of the Murravian biogeographical barrier.