Looking for mutations in PacBio cancer data: an alignment-free method

To determine the TP53 mutations present in a patient cohort (~1500 patients), breast tumor TP53 mRNA was sequenced with PacBio technology. However, none of the existing tools, e.g. VarScan, GATK, has proven to be suitable for this type of data. Indeed, in addition to the high sequencing error rate generated by PacBio (~15%), the tumor biopsies are contaminated by healthy tissue. This makes it difficult to differentiate real mutations as they are lost within the high background noise. To circumvent this problem, we have developed a method for detecting mutations using De Bruijn graphs: MICADo for Mutations In Cancer Data.