Identification of high and low divergence in the genomes of swamp and river buffalo

At an estimated time of 28-35k years BP, river (Bubalus bubalis bubalis) and swamp (Bubalus bubalis carabanensis) buffalo diverged from a common ancestor, the Indian wild buffalo (Bubalus arnee). These two subspecies of domestic water buffalo have been evolving separately since then, being subjected to distinct events of selection, migration, bottleneck and genetic drift. As a result, they accumulated marked morphological and behavioral differences. However, in spite of the genome rearrengment resulting in karyotype differences, river (2n = 50) and swamp (2n = 48) buffalo share high genome sequence synteny with each other and with domestic cattle (Bos taurus and Bos indicus). Therefore, as the reference water buffalo genome assembly is still under progress, oligonucleotide probes targeting common single-nucleotide polymorphism (SNP) markers across river and swamp buffalo can be mapped against the more mature bovine genome assembly to assist the identification of divergent loci between the two subspecies. Here, using a panel of 50,384 SNPs mapped to the UMDv3.1 bovine assembly, we sought 1Mb-long chromosomal segments sheltering multiple SNPs with very high or very low allele frequency (AF) differences between 185 river (distributed in 15 breeds) and 153 swamp (distributed in 13 breeds) animals. AF differences were scored using the FLK statistics, which uses a parallel drift model to estimate deviations from ancestral AF accounting for drift in the genome background. Divergence scores were computed as the average FLK across markers in the same window, and expressed as standard deviates. Only windows with at least five markers were analyzed. Considering a threshold of +/- four standard units, we found a total of 15 and three signals of higher (indicating directional selection) and lower (indicating balancing selection) than expected divergence, respectively. Interestingly, at least five of the high divergence signals sheltered genes involved in erythrocyte physiology and formation of hemoglobin, including HBB. The low divergence signals pointed to immune genes, mainly members of the MHC-II complex. This preliminary analysis suggests that river and swamp buffalo experienced different selective pressures for erythrocyte phenotypes after divergence, as opposed to an apparent convergent selection for maintainence of standing ancestral variants in the MHC-II complex.