Skip to main content

Table 1 Total variant calls made by each caller, pre-filtering and post-filtering out variants not suitable for validation by Sanger sequencing a

From: A simple consensus approach improves somatic mutation prediction accuracy

  MJS MJ MS JS M J S Total
All SNVs predicted, n (% of all variants)b 1,483 (16.1%) 83 (0.9%) 462 (5.0%) 298 (3.2%) 1,756 (19.0%) 2,387 (25.9%) 2,757 (29.9%) 9,226
SNVs suitable for validation (% of all SNVs suitable for validation)c 1385 (54.4%) 16 (0.6%) 370 (14.5%) 57 (2.2%) 279 (11.0%) 80 (3.1%) 360 (14.1%) 2,547
Average number (range) of filtered SNVs per sample 51.3 0.6 13.7 2.1 10.3 3.0 13.3 94.3
(1 to 246) (0 to 6) (5 to 42) (0 to 11) (3 to 33) (0 to 23) (5 to 25) (38 to 321)
Number (%) of SNVs that could not be used for validation 98 (6.6%) 67 (80.7%) 92 (19.9%) 241 (80.9%) 1477 (84.1%) 2307 (96.6%) 2397 (86.9%) 6679 (86.9%)
  1. Abbreviations: J. JointSNVMix2; JS, JointSNVMix2 + SomaticSniper; M, MuTect, MJ, MuTect + JointSNVMix2; MJS, MuTect + JointSNVMix2; + SomaticSniper; MS, MuTect + SomaticSniper; NRAF, non-reference allele frequency; S, SomaticSniper; SNV, single nucleotide variant.
  2. aCall sets are identified as described in Figure 1.
  3. bPredicted variants were filtered on the basis of read depth and NRAF, as described in the text.
  4. c'Suitable for validation’ means evidence for SNV met criteria for validation by Sanger sequencing (tumor NRAF ≥0.2, germline NRAF ≤0.05 and ≥8 reads in both samples.