Effect of misidentification and multiple populations on the genomic relationship matrix and genomic evaluations

Mislabeled genotyped animals can impact genomic analysis, as they give rise to false relationships within the population. Diagonal elements of the genomic relationship matrix (G) may be a useful indicator of mislabeled animals if G is scaled using current allele frequencies. Simulated data were used to find theoretical diagonal elements of G and field data were used to evaluate the utility of diagonal elements to separate a small and large population of animals in a genotyped chicken dataset. The effect of mislabeled animals on the accuracy of genomic predictions was also evaluated. When the diagonal elements are centered close to 1.00, mislabeled animals may have incorrectly scaled and abnormally large diagonal elements. Use of diagonal elements of G can identify animals from secondary populations; populations must be of unequal size or have different allele frequencies. Presence of mislabeled animals negatively affected genomic evaluations through loss of accuracy.