Abstract 818: Refining and assessing quality for next generation sequencing of formalin-fixed, paraffin-embedded tissues.

Abstract

Abstract With increasing acceptance of the use of formalin-fixed, paraffin-embedded (FFPE) for next generation sequencing (NGS), researchers and clinicians need to choose whether to analyze FFPE or cryopreserved (CRYO) tissue. In many clinical situations normal tissue would be available from blood, FFPE normal and tumor tissues from pathology archives, while fresh or CRYO tumor would require an invasive biopsy. Options include analysis of either CRYO or FFPE tumor coupled with either CRYO or FFPE normal tissue. Alternately, analyses could incude both CRYO and FFPE normal tissue to detect the impact of FFPE on NGS results, or could defer analysis of normal tissue with identification of novel mutations by comparisons of variants in tumor with SNP databases. Little data are available directly comparing results or tradeoffs resulting from these choices, prompting this study comparing NGS analysis of matched samples of normal CRYO and FFPE and tumor CRYO and FFPE. Appropriate kits from Qiagen were used to extract DNA from sections of CRYO and FFPE tissues of bladder cancers obtained from our institution's biorepository with informed consent. Libraries were constructed using Agilent library kits and targeted with SureSelect Kinome reagents. Libraries were sequenced by 100 bp paired end Illumina sequencing on a HiSeq. Our analysis pipeline used BWA for alignment and conducted subsequent analyses in GeneSpring NGS 12.1, including inspecting tile quality, removing duplicates, filtering on base quality, post alignment read (alignment score 95 and mapping quality 40 or above), and read status (mate near required). Comparing a representative specimen of tumor CRYO with tumor FFPE, targeted sequencing library preparations yielded 19.9 and 12.8 million reads for CRYO and FFPE respectively. For this analysis filtering by tile quality was not required. Uniquely mapped reads to human genome were 92% of CRYO and 86% of FFPE. Out of the total reads that aligned, 73.5% of CRYO and 74.3% of FFPE reads aligned without error. Post all filters, 39% of CRYO and 44% of FFPE reads were available for variant calling. Comparing variant calls from reference sequence to novel variants, there were no differences comparing normal CRYO vs normal FFPE, or tumor CRYO vs tumor FFPE. Comparing tumor with normal samples there were 30 variants for CRYO samples, 25 for FFPE samples, and 23 when CRYO normal was compared with FFPE tumor. In comparing normal vs tumor, 22 variants calls were the same for all comparisons of tissue types. Results will be presented for a more detailed analyses of multiple samples Overall we found high concordance between results from the CRYO and FFPE samples. Our results support use of FFPE for NGS of clinical samples and highlights potential contribution of including sequencing a normal cryropreserved specimen to ensure analysis quality. Citation Format: Umar Farooq, Michael J. Gooch, Linda Burian, Prashant Singh, Shirin Karimi, Lorrie A. Perpetua, Mary M. Sanders, Richard B. Everson. Refining and assessing quality for next generation sequencing of formalin-fixed, paraffin-embedded tissues. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 818. doi:10.1158/1538-7445.AM2013-818