Goat genomic selection: Impact of the integration of genomic information in the genetic evaluations of the Spanish Florida goats

Genomic evaluations have been proposed as a mean to improve the reliabilities of the estimation of breeding values. This hypothesis has been tested using 50,649 lactation records from 19,067 Florida goats, daughters of 4397 dams and 500 sires. A sample of 538 dams and 87 sires from the formerly described population were genotyped with the Illumina 55 K Goat Bead-Chip (53,347 SNP). Genetic parameters were estimated using the animal model and REML methodology comparing two approaches: (i) a classical approach with the BLUP methodology; and (ii) a single step (ss) approach with the single step genomic BLUP (ssGBLUP) methodology. The BLUPF90 software was used to obtain the genomic relationship matrix (G) and to estimate genetic parameters and breeding values. The results showed a correlation between A (pedigree relationship matrix) and G matrix of 0.826. When the full studied population was considered, no significant differences were observed between the estimations of the genetic parameters obtained with A and combined A and G matrices. The correlation between the EBV and GEBV was 0.989. An increment of 1.06% in the average reliability of the estimations was observed in the ssGBLUP vs. the traditional BLUP evaluation. When only the EBVs of the animals genotyped were compared, the correlation between the estimates obtained with both approaches decreased to 0.952, but with an increment of 5.86% of the average reliability of the GEBVs.