Next-Level Genomic Selection: Mitigating Inbreeding

Abstract

Analysis of 88,068 autosomal or X-linked SNPs in the ANAFIBJ Holstein genomic database including males and females showed declining SNP heterozygosity over time. In 1990, average SNP heterozygosity was 0.3620. During the pre-genomics period (1990-2010), the annual decline was -0.0003, reaching 0.3558 in 2010. However, in the genomics period (2010-2024), the average SNP heterozygosity declined to 0.3191 in 2024, with an annual decline of -0.0027, over 7 times higher than before. So far, this trend has been highly linear (R²=0.987), which would extrapolate (in the absence of other sources of genetic variation than selection and inbreeding) to result in a complete loss of genetic variation in ~130 years.

We developed measures to estimate genomic expected future inbreeding (Gefi) based on Runs-Of-Homozygosity (ROH). A comparison with CDCB genomic future inbreeding (GFI) based on the genomic relationship matrix (GRM) was done using 38,280 genotyped proven males covering 70 years (1951-2020), which resulted in a Pearson correlation of 0.959. The CDCB GRM G was computed as G = ZZ’/Σ2p(1 – p) using p=0.5. Gefi estimates had a mean of 6.9% with a standard deviation of 2.6%. Minimum Gefi was 0.1% and maximum was 15.3%. GFI estimates had a mean of 7.2% with a standard deviation of 2.6%. Minimum and maximum GFI were -3.1% and 13.5%. The correlation was quite high even though Gefi is an identity-by-descent (IBD) measure in the probability space of [0,1], whereas GFI measures identity-by-state (IBS) in the correlation space of [-1,1]. Comparison of the inbreeding depression across traits showed that the depression is largest on yield traits followed by contents, somatic cell score (SCS) and fertility at around 20% of the depression on yield traits.

ANAFIBJ aims to reduce the increase in future inbreeding by giving a premium to male and female animals which are less related to the recent population, while penalizing those that are more related.