Abstract 3274: Pan cancer analysis of fusion genes in TCGA using ChimeRScope, an alignment free algorithm

Abstract

Abstract Fusion transcripts, frequent in cancer formed from the concatenation of two unrelated genes results primarily from the structural rearrangements of the genome. Fusion transcripts that are unique to a cancer type can be exploited to understand the underlying mechanisms for malignancy and can serve as diagnostic or prognostic markers. We used ChimeRScope, a novel alignment-free algorithm developed recently by our group. This program is specifically designed to identify fusions in cancer transcriptomes that contain frequent chromosomal aberrations and structural variations. We explored the fusion landscape of 31 cancers from TCGA with a focus to identify transcriptionally induced fusions along with recurrent fusions across cancer types. Level 1 paired-end RNA sequencing data from The Cancer Genome Atlas (TCGA) GDC data portal was downloaded and unmapped reads extracted from both tumor and tumor adjacent normal tissue. Unmapped fastq files were then analyzed with the ChimeRScope pipeline and fusions identified in tumor adjacent normal tissue was removed from further analysis. Fusions identified by ChimeRScope Examiner module were further screened to remove those with low confidence using custom python scripts. We classified fusions based on the distance of break point from the participating exon boundaries as E-E (both break points near exon-intron junctions), E-M (one break point near junction), and M-M (both break points away from the junction). Further, samples with somatic whole-genome sequence data were downloaded from TCGA database and analyzed for structural variants using BreakDancer to identify fusions supported by genomic events. A total of 8,409 primary tumor samples and 730 normal samples spanning 31 cancer types from TCGA were analyzed in this study. Of the identified fusions across different cancer types, Breast Invasive Carcinoma had the maximum number of fusions. Thymoma, Melanoma, Breast Invasive Carcinoma, Testicular Germ Cell Tumor, and Acute Myeloid Leukemia had the highest number of recurrent fusions. Uterine carcinoma, Lung Squamous cell carcinoma, and Sarcoma had the maximum number of samples with at least one fusion. Approximately 30% of the fusions identified in each tumor was in-frame. We identified a number of fusions that are recurrent across cancer types involving kinases, oncogenes or tumor suppressors. This is a comprehensive analysis of fusions across all of the major class of cancers from TCGA. Data from this study indicate the presence of a vast pool of unexplored fusion events that need to be evaluated for functionality. Citation Format: Neetha Nanoth Vellichirammal, Jasjit Banwait, Abrar Albahrani, You Li, Chittibabu Guda. Pan cancer analysis of fusion genes in TCGA using ChimeRScope, an alignment free algorithm [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3274.