Using the 90K Axiom Buffalo Array for Genome Wide Association Studies

The buffalo population in Brazil is around 1,332,284 heads. Although it represents just 0.7% of the world population, it is the biggest population in Americas. It has a potential growth since there is a demand for mozzarella production in the country. The aim of the abstract was to sum up the genomic studies that are being developed at the Animal Science Department, Unesp-Jaboticabal-SP, Brazil. Information from 452 buffaloes (Bubalus bubalis) (57 sires and 395 dams) genotyped using the 90k Axiom ® Buffalo Genotyping (Affymetrix) and 384 buffaloes (dams) genotyped using the Illumina BovineHD BeadChip (777k) were used in the studies of linkage disequilibrium, genome-wide association and genomic selection. For the Illumina 777k cattle chip, it was possible to verify that ~88% of the SNPs were genotyped in buffaloes, but only ~2% (16,580) of the SNPs were segregating. It means that the majority of them were fixed and don’t contribute to the understanding the genetic dynamics of the species. For Affymetrix chip, almost all the SNPs were genotyped and ~66% (61,145) of the SNPs have good quality to be used in genomics selection and association studies. Linkage disequilibrium studies were done using the Affymetrix chip. A r² value was 0.30 ± 0.33 (with an average distance of 0.043 ± 0.027 Mb).  It means that the SNP information is a suitable tool to obtain genomic breeding values. The LD decreased with increasing physical distance between SNP pairs. The effective size was 86 estimated in the last 10 generations and it is close to the minimum limits recommended by FAO. So, there is a necessity to monitor inbreeding and mating decisions to maintain the genetic diversity. These studies were done using Illumina panel and similar results were found. Genome-wide association studies were done for six production traits and four reproductive traits using the Affymetrix chip. The associations were done with the deregressed breeding values. Major genes were prospect for each trait as well as genes with pleiotropic effect. These studies were also done using the Illumina panel, but different regions were found, mainly because the genome coverage is poorer. Genomic selection studies were done using the Illumina chip. Models with the matrix A and a model with the matrix H were compared for six production traits. The heritabilities estimates were similar, but the accuracies of the breeding values were higher for the models with the matrix H. The lack of a buffalo genome makes the genetic progress and understanding of the species a bit slower. The idea of the group is to genotype more animals in order to increase the accuracy of the estimates predicted and repeat the analyses done as well as develop others. We look for collaborators who may contribute with the current buffalo genomic project.