Sensitivity of the Assay. Five independent mixing studies were performed in duplicate using KRAS-mutant cell lines (SW480, LS174T, HCT116) paired with wild-type peripheral blood leukocyte DNA and KRAS-mutant formalin-fixed paraffin-embedded (FFPE) tissue samples paired with KRAS-WT FFPE samples as described in the methods section. For each mixing study the mutant melting peak height was measured as a percentage of the maximum mutant peak height observed in the undiluted KRAS-mutant sample specific to the particular Mutant:WT pairing. The left panel shows the average and SD (error bars) of the %mutant peak height across the 5 mixing studies using regular- (gray triangles) or COLD-PCR (black squares) conditions. Mutant peak heights were consistently above the limit of detection down to 1% mutant allele using COLD-PCR compared to 10% mutant allele with regular PCR. The right panel shows an example of one of the mixing studies (SW480 cell line; homozygous G12V) used to generate the sensitivity graph in the left panel. Mutant melting peaks are not clearly discernible for 1%, 2%, and 5% KRAS-mutant samples using regular PCR (top right panel), but are readily detected using COLD-PCR (bottom right panel).