A new method to accurately identify single nucleotide variants using small FFPE breast samples.

Most tissue collections of neoplasms are composed of formalin-fixed and paraffin-embedded (FFPE) excised tumor samples, and routine diagnosis in oncology relies on histopathological analysis of those samples. Genomic sequencing is becoming increasingly important in the clinical management as well as the basic science of cancer. Unfortunately, genomic sequencing of FFPE samples is difficult due to the small amounts of DNA available particularly from early cancers, as well as degradation of that DNA. We developed a new bioinformatic algorithm to robustly identify somatic mutations using small amounts of DNA extracted from archival FFPE samples of breast ductal carcinoma in situ, a preinvasive form of breast cancer. We optimized this strategy using 28 pairs of technical replicates, in which the same DNA sample was sequenced twice independently. After optimization, the mean similarity between replicates was 88.3%, range 66.7-100%, and we were able to detect an average of 19.9 (range 1-61) single nucleotide variants in each sample. We found that the accuracy of identifying SNVs severely declined when there was less than 40ng of DNA available. High depth resequencing also showed that insertion-deletion (indel) variants are an unreliable subset of mutations, using current methods. This new algorithm was empirically optimized and validated. It provides a significant improvement in detecting somatic single nucleotide variants in FFPE samples that can be used to accurately profile the genomes of neoplasms.