Validation Of A Procedure For Simultaneous Variant Calling And Differential Expression In Cancer Study

Proceedings: AACR 106th Annual Meeting 2015; April 18-22, 2015; Philadelphia, PAWe have designed an integrated RNA-seq pipeline that determines both sequence variants and the differential expression of the transcripts from the same alignment in cancer samples. Comparison with other widely used protocols for RNA-seq data analysis showed that our method was more sensitive for determining subtle phenotypes in studies with large sample size, which frequently occur in compound screening approaches for cancer therapy. This whole transcriptome sequencing technology provides a cost-effective approach to gain understanding into the RNA relative biological events by deciphering RNA sequences. Benefiting from the fast developing technology, the cost of deep RNA-seq is no longer a barrier for current applications. Furthermore, an RNA-seq dataset provides more insight into disease processes than DNA-seq, since single nucleotide variant (SNV) detection in addition to gene and exon level expression and new isoform detection can be garnered from the same information. Based on the recent public data for mixed cancer cell lines and pooled normal human samples in MicroArray Quality Control (MAQC) phase III, also referred to as Sequencing Quality Control (SEQC), we evaluated our method with previous published results by means of different alignment strategies, sequencing depths and replicates numbers. Receiver operating characteristic (ROC) analysis showed our gene expression results agreed with the ones from the SEQC teams. In most cases, our solution has similar performance with other widely used strategies, while our STAR 2 pass based strategy has better true positive and true negative rates in the experiments with large sample size and subtle difference between groups. At the same time, the Genome Analysis Toolkit (GATK) 3 in our pipeline gained a reasonable balance between accuracy and performance. In summary, our new RNA-seq analysis pipeline has good performance and balance between precision and speed. Furthermore, this work also implicitly indicates that our gene expression method generates more accurate results in a large sample size experiment among similar groups.Citation Format: Gang Feng, Jamie Osman, Sean Leighton, Lynn Carmichael, Yuchen Bai, John Begemann, Richard Mazzarella. Validation of a procedure for simultaneous variant calling and differential expression in cancer study. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4869. doi:10.1158/1538-7445.AM2015-4869