Including Imprinting Effects in Genomic Best Linear Unbiased Prediction Method for Genomic Evaluation

Abstract Text: This study presented a genomic best linear unbiased prediction (GBLUP) method, including additive and imprinting effects (GBLUP-I), in which an imprinting relationship matrix was constructed from paternal and maternal alleles. The proposed method was implemented to investigate the extent of additive and imprinting variances, and to assess the accuracy of genomic predictions in simulation data. GBLUP-I showed a capacity for capturing imprinting variances. Assuming that imprinting variance accounted for 7.3%, 14.6%, and 23.0% of the total genetic variance, the accuracies in GBLUP-I exceeded those in the GBLUP by 3.9%, 4.7%, and 9.9% when paternal and maternal alleles were known and 2.1%, 2.3%, and 6.4% when paternal and maternal alleles were imputed. An improvement in the performance of GBLUP-I was observed when both paternal and maternal alleles were silenced by imprinting and genetic variances between males and females were almost the same.

Keywords:

imprinting

genomic best linear unbiased prediction