Fig. 4

Decoded and missing genotypes in data for both imputation scenarios.The minor and major alleles are denoted m and M. For simplicity, the simulated decoded genotypes from pooling are represented in GT format. We remind adaptive GL are provided later in the experiment for running imputation on data informed with the pooling outcomes. Half-decoded (GT = M/. or ./m) and not decoded (GT = ./.) genotypes are considered as missing data. The relative genotypes proportions are scaled in [0, 1] within each bin. a The markers only in the LD data set are fully assayed, all other markers have been deleted. b True heterozygous genotypes (dark blue) are never fully decoded, whereas the rare variants are almost all fully decoded or at least one of the two alleles is determined