Toward a new version of medium-density Buffalo SNPchip array

The water buffalo (Bubalus bubalis) is a large bovid used, since ancient times, for human needs. Today it is bred in many countries across the globe. There are two distinct species of water buffalo: river (from western Asia to Europe) and swamp (more common in eastern Asia). In Italy, river buffalo have been present since the Roman period and today Italian Mediterranean is bred for milk production, used to produce ‘mozzarella di Bufala campana’. To study buffalo genomic variability, a first SNP array (Axiom Buffalo Genotyping – ThermoFisher Scientific) containing 90k SNP was developed by the International Buffalo Genome Consortium, covering the water buffalo (river and swamp) genome-wide diversity. However, to improve milk quantity and quality in the Italian Mediterranean breed, also through Genomic Selection, a new and breed-specific version of the Axiom Buffalo array was developed. The first step was to validate the already existing SNPs Axiom Buffalo array in the Mediterranean breed. All the probes were mapped to the most recent Mediterranean buffalo reference sequence, removing the ones with low mapping quality or without mapping. After that, a quality control of the probes was performed using around 700 Mediterranean animals. Probes with low call rate, monomorphic or with not optimal clustering were removed. The remaining probes were evaluated: SNPs previously identified as important (e.g. present in IMAGE array, already associated with milk characteristics, etc.) were prioritised. In almost all the cases only one probe was kept. Finally, around 85k probes (and 76k SNPs) over 123k available sports were retained. The second step consisted in filling the new array gaps using WGS Mediterranean data. The sequences were aligned and variant called, using state-of-the-art bioinformatic pipelines. SNPs associated with functional genes were added (for example, casein gene cluster), together with SNPs in the Y chromosome (which was not previously included), and in the regions with lower density, compared with the expected. The final array density reached around 35 SNPs/Mbp. In conclusion, the new version of the buffalo array has 25% more SNPs than the first released, with a very high usage rate in the Mediterranean breed.