Breeding and Genetics: Genomic Methods and GWAS

摘要

Single-step genomic BLUP (ssGBLUP) is a method that combines allsources of information in a single analysis to compute genomic EBV(GEBV). For single-step genome-wide association studies (ssGWAS),GEBV are backsolved to SNP effects, and those effects are converted toproportion of explained additive genetic variance. Thus far, no formalframework for hypothesis test is currently present in ssGWAS fromthe BLUPF90 software suite. Our objective was to implement P-valuesfor ssGWAS and to apply the method to a large dairy cattle population.P-values were obtained based on the prediction error (co)variance forSNP, which uses the inverse of the coefficient matrix for genotypedanimals and formulas to compute SNP effects. Six steps are needed forthe calculation of P-values: 1) factorize and invert the LHS of ssGBLUP;2) solve MME using sparse Cholesky factor; 3) extract the LHS~(−1) forgenotyped animals; 4) backsolve GEBV to SNP effects; 5) obtain theprediction error covariance for SNP effects; 6) calculate P-values usingthe cumulative standard normal function of SNP effect divided by standarddeviation of SNP effect. The US Holstein data used in this studyconsisted of almost 800k udder depth records for 500k cows. Pedigreeinformation was available for 1.3M animals, of which 8,802 sires weregenotyped. The model contained the same effects as the official modelused for linear type trait evaluation in the US; however, in a single-traitsetup. Computation of P required 20Gb of memory and no inflation wasobserved. The SNP passing the Bonferroni threshold of 6.1 in the −log10scale were the same as those that explained the highest proportion ofadditive genetic variance. The exact P-value for ssGWAS is a verygeneral and efficient strategy for QTL detection and test. It can be usedin complex data sets such as the ones used in animal breeding, whereonly a proportion of pedigreed animals are genotyped. The BLUPF90software suite is now equipped with the P-value calculation tool.