Next generation sequencing is increasingly used as a diagnostic tool in the clinical setting. This is driven by the vast increase in molecular targeted therapy, which requires detailed information on what genetic variants are present in patient samples. In the hospital setting, most cancer diagnostics are based on Formalin Fixed Paraffin Embedded (FFPE) samples. The FFPE routine is very beneficial for logistical purposes and for some histopathological analyses, but creates problems for molecular diagnostics based on DNA. These problems derive from sample immersion informalin, which results in DNA fragmentation, interstrand DNA crosslinking and sequence artefacts due to hydrolytic deamination. Distinguishing such artefacts from true somatic variants can be challenging, thus affecting both research and clinical analyses.

In order to identify FFPE-artefacts from true variants in next generation sequencing data from FFPE samples, I developed the novelprogram FUSAC (FFPE tissue UMI based Sequence Artefact Classifier) for the facility Clinical Genomics in Uppsala. FUSAC utilizes UniqueMolecular Identifiers (UMI's) to identify and group sequencing reads based on their molecule of origin. By using UMI's to collapse duplicate paired reads into consensus reads, FFPE-artefacts are classified through comparative analysis of the positive and negative strand sequences. My findings indicate that FUSAC can succesfully classify UMI-tagged next generation sequencing reads with FFPE-artefacts, from sequencing reads with true variants. FUSAC thus presents a novel approach in bioinformatic pipelines for studying FFPE-artefacts.