Fig. 2

Identification of mislabeled specimens. The variant allele frequencies are extracted from T/N paired samples according to unique mutation events defined by chromosome, position, reference allele and variant allele, subsequently used to generate a scatter plot between two samples. a A significant portion of shared mutations from correct T/N paired samples are distributed closer to the diagonal of the scatter plot with a large majority of heterozygous and homozygous variants located at regions of 50 and 100% allele frequencies. b Two additional groups of points located at the top and right axis as indicated by red ovals in the scatter plot can be easily depicted when T/N mismatched samples were used to generate this figure, based on the concept that these variants are less likely to change all their variant frequencies from 50 to 100% through mutation events