Figure 4.

Background confidence intervals eliminate most variants. A) Observed substitution VAFs for a Tier III probe, illustrating the varying mutation presence at each nucleotide locus. Error is standard deviation across 20 individuals. B) Subsets of the IDH2 probe region from two different groups of 10 individuals illustrates the degree of similarity and differences that are commonly observed between samples. Error is standard deviation across the individuals. C) Total number of variants deemed significantly above background when only triplet context was used to generate expected background substitution rates (Generic), or when position-specific substitution rates generated from a different experiment (External) or the same experiment (Internal) are used. D) The total numbers of variants called as significantly above background for 5 individuals (labeled as sample number) at confidence intervals from 0.9 - 0.999999999999999. where the number of trailing 9’s is indicated by x-axis value. E) The total numbers of variants called as significantly above background for two leukemic patients using sample 4 from MRD1 and sample 3 from MRD2 (See Figure 6), which are both points that had followed treatment, and at which leukemic burden was low.