Genomic Divergence across Ecological Gradients in a Central African Songbird (Andropadus virens)

Rainforests are heralded for their exceptionally high biological diversity, yet the relevant evolutionary mechanisms for the generation and maintenance of this diversity have long been debates. Long-term studies of the little greenbul (Andropadus virens), which is distributed broadly across ecologically diverse habitats in sub-Saharan Africa, have found morphological and behavioral divergence across rainforest-savanna transition zones (ecotones), mountains, and islands, suggesting divergent natural selection may contribute to adaptive phenotypic divergence despite high levels of gene flow. With the rapid advances in genomics and environmental modeling, it is now possible to examine these mechanisms in greater depth. Here we used RADseq to investigate patterns of differentiation among A. virens populations across environmental gradients, and to assess the extent to which natural selection may drive patterns of differentiation across habitats in Central Africa. We find compelling evidence that individuals form discreet genetic clusters corresponding to habitat types. Pairwise F_ST between populations in different habitats is significantly higher than within habitats, and this differentiation is greater than what is expected as the result of geographic distance alone. Moreover, we identified 140 SNPs that showed extreme differentiation among populations through a genome-wide selection scan. These outliers were significantly enriched in exonic and coding regions, suggesting these SNPs are functionally important. Finally, environmental association analysis suggested geographic distance alone cannot account for SNP variations across the range of A. virens, and that several environmental factors, including temperature and elevation, may serve important roles in driving the pattern of genomic diversification in A. virens.