Figure 3
Variation of SNP-wise imputation accuracy* and linkage disequilibrium along bovine chromosome 1. Top: SNP-wise correlation between imputed and observed genotypes (CORR) is plotted against the genomic coordinates (in Mb) for SNPs located on chromosome 1, which was divided in windows of about 50 subsequent markers; windows with the lowest (a) and highest (b) average imputation accuracies are highlighted. Middle: Heatmap representing the extent of linkage disequilibrium (r2) in window A (51 markers located between 44.71 and 44.91 Mb; averages for accuracy, MAF and r2 were 0.390, 0.195 and 0.103, respectively). Bottom: Heatmap representing the extent of r2 in window B (48 markers located between 69.40 and 69.49 Mb; averages for accuracy, MAF and r2 were 1.000, 0.270 and 0.321, respectively). *In order to exemplify the amount of variation verified for SNP-wise imputation accuracy on a single chromosome, the results obtained from Analysis 9 (Figure 1) are presented (i.e. using the 11a7k chip and FImpute considering pedigree information to impute genotypes of young sires).