A common source of false positive mutation calls is contamination of the tumor DNA with DNA from other individuals. Germline SNPs in the contaminating DNA appear as somatic mutations. We have previously demonstrated that such contamination can yield a large number of false positives and developed a tool, ContEst46 to estimate the contamination level, fcont, in sequencing data. Low-level contamination of DNA is a common phenomenon and even 2% contamination can give rise to 166 false positive calls per megabase and 10/Mb when excluding known SNP sites46. To protect against this type of false positives and enable analysis of contaminated samples, we replace the reference model with a variant model Mfcontm. This guarantees that variants are called only when they are highly unlikely to be explained by contamination.
