Fig. 2: Record-matching accuracies for sets of SNPs that possess a specific characteristic.
From: Toward minimal SNP sets for record-matching with CODIS STR profiles
Rows correspond to different statistics used for SNP selection. A Random selection. B MAF. C Pop-MAF (minimal MAF across 1000 Genomes Project super-populations). D Distance between SNP and CODIS STR. E D’avg. Columns correspond to the four matching scenarios. Each plot shows record-matching accuracies for SNP sets of four sizes. Points represent the median, and error bars represent the range of record-matching accuracies across 100 replicates (10 partitions of the individuals, 10 SNP sets). Colors and point shapes correspond to different conditions. Closed symbols correspond to conditions that achieve comparable record-matching accuracies to those obtained with the full set of 192,672 SNPs. The full numerical values are shown in Table S5.
