Skip to main content

Table 3 Accuracy comparisons between the two-step piecemeal and the classic one-step imputation on the Simmental datasets

From: Whole genome SNP genotype piecemeal imputation

  5-Fold cross validation   Independent testing
BaseProgram Imputation a c c 1 a c c π + #Clusters #TClusters a c c 1 a c c π +
  6 K →50 K 69.35 % 70.81 % 1.46 % 100 100 60.68 % 61.39 % 0.71 %
Beagle 6 K →660 K 72.37 % 74.92 % 2.55 % 800 800 66.00 % 67.76 % 1.76 %
  50 K →660 K 86.61 % 88.89 % 2.28 % 1000 1000 72.83 % 74.11 % 1.29 %
  6 K →50 K 75.95 % 76.70 % 0.75 % 55 55 61.87 % 62.16 % 0.29 %
FImpute 6 K →660 K 79.11 % 80.11 % 1.00 % 1000 1000 68.43 % 68.95 % 0.52 %
  50 K →660 K 90.31 % 90.74 % 0.43 % 1000 1999 77.11 % 77.33 % 0.22 %
  1. Results are on the Simmental datasets for markers on chromosome 14. Columns 3–7 contain the 5-fold cross validation results on the 82 animals, with the selected markers and their associated target marker clusters. Independent testing results on the 367 animals are in columns 8–10, using the selected markers and their associated target marker clusters from the cross validation. 1In the independent testing from 50K to 660K, 8 markers of the Affymetrix 660K chip were filtered out due to their genotype disagreeing with the alternating alleles specified by sequencing, and consequently only 999 target marker clusters were used. The columns labelled with + show the improvements, in bold, of the piecemeal imputation over the one-step imputation