When a reference set was used, the data were imputed with reference to data from the Genome of the Netherlands (GoNL) project (Boomsma et al. 2014, Genome of the Netherlands Consortium 2014).
In each study, we performed an association analysis of up to 2 543 888 SNPs, including genotypes that were imputed with reference to the International HapMap Project CEU panel release 22.
Imputation procedures generally involve a set of target samples in which genotypes will be imputed, a reference panel of phased haplotypes from which genotypes are copied, and an algorithm for the copying procedure.
We compared our method for regional analysis against standard SNP-based analyses with (i) only the genotyped SNPs or with (ii) the full set of SNPs after imputing against reference panels from phase 2 of the HapMap (HapMap2).
A two-step genotype imputation, comprising a phasing step to resolve the haplotypes of the genotyped markers (using MACH software) and an imputation step in which unmapped SNPs are imputed to a reference panel (using Minimac software), was applied to the GWAS genotyped dataset after QC.
For several confounders, missing values were imputed to the reference value to allow for comparison of nested models.
Generally, only a subset of all polymorphisms of a person are genotyped, after which the other polymorphisms are imputed with a reference panel.
DNA samples were genotyped on the Infinium Omni Express bead microarray and imputed up to reference panels from the 1000 Genomes Project.
Of 639 214 SNP, 39 679 SNP were retained for 397 validation animals, while genotypes for the remaining 599 535 SNP were imputed using 400 reference animals.
When a breed was imputed using a reference population that comprised individuals of the other pure breed (external-breed scenario), the mean relationship decreased to approximately one fifth of that obtained with the within-breed scenario.
