Genomic imputation and evaluation using high-density Holstein genotypes

Area of Science:

• Animal Genetics
• Bioinformatics
• Quantitative Genetics

Background:

• Genomic evaluations are crucial for dairy cattle breeding.
• High-density (HD) genotyping offers potential for improved accuracy.
• Imputation methods are essential for utilizing diverse marker densities.

Purpose of the Study:

• To evaluate the accuracy of genomic evaluations using different marker densities in Holstein cattle.
• To assess the impact of imputation methods on genotype accuracy.
• To compare genomic evaluations with traditional parent averages.

Main Methods:

• Genomic evaluations computed for 161,341 Holsteins using up to 311,725 markers from the Illumina BovineHD Genotyping BeadChip.
• Genotypes from various densities (HD, 50K, 8K, 6K, 3K, 0K) were utilized and imputed using findhap and FImpute programs.
• Imputation accuracy was tested on simulated data, and genomic reliability was compared against traditional parent averages using deregressed evaluations.

Main Results:

• Imputation accuracy was high, with findhap and FImpute achieving >90% correct genotypes across marker densities.
• Genomic reliability averaged 61.1% with 311,725 markers and 60.7% with 45,187 markers, compared to 29.6% from parent average.
• A modest 0.4 percentage point average increase in reliability was observed with HD markers compared to 50K markers.

Conclusions:

• High-density genotyping provides a slight but consistent improvement in genomic evaluation reliability for Holsteins.
• Imputation strategies significantly impact genotype accuracy, with multi-step imputation showing benefits for lower-density chips.
• Further increases in HD genotypes beyond 1,074 did not substantially enhance Holstein genomic evaluation reliability in this study.