TY - GEN
T1 - Genomic prediction of methane emission using microbiome data and genomic markers in Holstein cows
AU - Saedi, Naghmeh
AU - Ye, Xiaoxing
AU - Cai, Zexi
AU - Lund, Mogens Sandø
AU - Karaman, Emre
PY - 2024
Y1 - 2024
N2 - Ruminant animals produce methane as a by-product from microbial digestion of feed in the rumen and cattle are the largest contributor to methane emissions among ruminants. Host genetics affect methane emissions of cows directly, and indirectly through affecting the rumen microbiome composition. In this study, we aimed to investigate if modeling direct and indirect effects of host (cow) genetics improves genomic prediction of breeding values for methane. A total of 667 Holstein cows, with genotypes, phenotypes, and rumen microbiome information were used. The rumen microbiome data included operational taxonomic units (OTUs) or zero-radius OTUs (ZOTUs). Genomic (GBLUP) and pedigree-based (PBLUP) best linear unbiased prediction methods were used for benchmarking a method (GMBLUP) combining results of a set of two linear models: (i) one for modeling rumen microbiome profiles (OTUs or ZOTUs) as a linear function of single-nucleotide polymorphism (SNP) genotypes, and (ii) one for modeling methane phenotypes as a linear function of SNP genotypes. The heritability of methane emission was estimated to be 0.14 and 0.27, from PBLUP and GBLUP respectively. It was 0.27 and 0.29 from GMBLUP models using OTUs and ZOTUs, respectively. The microbiome composition explained 0.020 and 0.039 of the phenotypic variances using GMBLUP(O) and GMBLUP(Z), respectively. Similar levels of prediction accuracy were observed for GBLUP (0.21), GMBLUP(O) (0.20), and GMBLUP(Z) (0.19), contrasting with the notably lower prediction accuracy of 0.12 achieved by the PBLUP model. In conclusion, similar prediction accuracies obtained from GBLUP and GMBLUP indicated that OUTs (or ZOTUs) are responsible for only a small fraction of the variation in methane emissions among individuals, and most of this variation is explained by the direct genetic effects of cows.
AB - Ruminant animals produce methane as a by-product from microbial digestion of feed in the rumen and cattle are the largest contributor to methane emissions among ruminants. Host genetics affect methane emissions of cows directly, and indirectly through affecting the rumen microbiome composition. In this study, we aimed to investigate if modeling direct and indirect effects of host (cow) genetics improves genomic prediction of breeding values for methane. A total of 667 Holstein cows, with genotypes, phenotypes, and rumen microbiome information were used. The rumen microbiome data included operational taxonomic units (OTUs) or zero-radius OTUs (ZOTUs). Genomic (GBLUP) and pedigree-based (PBLUP) best linear unbiased prediction methods were used for benchmarking a method (GMBLUP) combining results of a set of two linear models: (i) one for modeling rumen microbiome profiles (OTUs or ZOTUs) as a linear function of single-nucleotide polymorphism (SNP) genotypes, and (ii) one for modeling methane phenotypes as a linear function of SNP genotypes. The heritability of methane emission was estimated to be 0.14 and 0.27, from PBLUP and GBLUP respectively. It was 0.27 and 0.29 from GMBLUP models using OTUs and ZOTUs, respectively. The microbiome composition explained 0.020 and 0.039 of the phenotypic variances using GMBLUP(O) and GMBLUP(Z), respectively. Similar levels of prediction accuracy were observed for GBLUP (0.21), GMBLUP(O) (0.20), and GMBLUP(Z) (0.19), contrasting with the notably lower prediction accuracy of 0.12 achieved by the PBLUP model. In conclusion, similar prediction accuracies obtained from GBLUP and GMBLUP indicated that OUTs (or ZOTUs) are responsible for only a small fraction of the variation in methane emissions among individuals, and most of this variation is explained by the direct genetic effects of cows.
M3 - Article in proceedings
BT - Genomic prediction of methane emission using microbiome data and genomic markers in Holstein cows
ER -