Comparative Analysis of Various Targeted Enrichment Methodologies to Identify Actionable Mutations in Lung Cancer.

e13110 Background: The recent expansion of knowledge about genetic changes associated with various diseases including cancers, using whole genome or exome sequencing has led to the development of focused or targeted gene panels. Such panels can be used to characterize both germ-line and somatic changes with a high degree of specificity and sensitivity. Two different approaches-hybridization capture and PCR-have been widely used to generate target specific libraries covering relevant loci. The limited amount of sample available, particularly in tissue biopsies and FFPE specimens, can restrict the use of both approaches for investigating a large number of gene targets. In addition, false positives may occur during the early steps of PCR or during whole genome amplification introducing potential errors. Methods: We used probes specific for the coding regions of 215 genes from Nimblegen (capture approach), NuGEN (AMP approach) and Illumina (PCR approach) to identify clinically actionable mutations and gene fusions in lung cancer. Various control DNAs were sequenced at 10,000X coverage and the data was analyzed at different depths. DNA isolated from 32 FFPE lung tumor samples of varying quality were interrogated by these three approaches. Results: A minimum coverage of 500X was required for identifying genomic aberrations with a high degree of confidence. The majority of identified variants (90%) were called by all three approaches. In these experiments, samples older than 5 years showed a high duplication rate and required greater coverage to call variants. We found on average 3 pathogenic mutations to be present in genes such as TP53, ATM and ATR in the 32 analyzed samples. Conclusions: All three methods can be used to profile DNA isolated from FFPE samples. It is expected 10% of the variants will be unique to each method. Some of this is likely due to coverage of target region and the ability of the bio-informatics tools to detect variants. To identify any false positives, uniquely identified variants from each approach are being confirmed by pyrosequencing. These datasets are also being further analyzed by various pathway analysis tools to define possible pathogenic mechanisms in lung cancer.