Abstract 1574: Computational Approach for Discovery of Regulatory Noncoding Variants in Cancer
Yu Liu,Michael Edmonson,Xiaotu Ma,Michael Rusch,Yongjin Li,Benshang Li,Shuhong Shen,A. Thomas Look,Jinghui Zhang
DOI: https://doi.org/10.1158/1538-7445.am2017-1574
IF: 11.2
2017-01-01
Cancer Research
Abstract:Abstract Noncoding mutations in the cancer genome have historically been poorly understood. Recent studies have shown however that mutations in intergenic regions may play an important role in cancer by introducing de novo regulatory elements and thus cis-regulating target gene expression. While mono-allelic expression of the dysregulated gene is an important signature of regulatory noncoding mutations, recognition of this feature has not been implemented by any methods evaluating noncoding mutations to date. To perform genome-wide detection of noncoding mutations with a potential regulatory function, we developed a novel computational approach implementing this allelic specific expression signature (ASE) as a core component, using WGS and RNAseq data as input. We first searched for genes showing ASE across the tumor genome by comparing the minor allele frequency of heterozygous markers between WGS and RNAseq data. We excluded genes known to be in an imprinted locus, as well as sites matching a precomputed set of 5,576,857 problematic polymorphism markers (estimated by analyzing WGS data from 625 normal genomes). Next, the somatic mutations were filtered for regulatory candidates that could introduce a transcription factor (TF) binding motif absent from the reference sequence. We included 614 TF binding motifs from the HOCOMOCO database and performed motif analysis with the FIMO package. Finally, the somatic mutations were assigned to dysregulated genes in the same topologically associating domains in human defined by public Hi-C data and prioritized by open chromatin state defined by DNaseI hypersensitivity regions (ENCODE project). We applied our approach to 22 T-ALL tumor samples collected at Shanghai Children’s Medical Center in a pilot study consisting of 13 diagnosis and 9 relapse samples (9 of these were trios). WGS was performed on both tumor and normal samples and RNAseq was performed for the tumor sample only. We predicted 213 regulatory noncoding mutations (157 unique variants) in total from this cohort, potentially targeting 119 genes. Among the dysregulated targets we found TAL1 and LMO2, which are known TFs playing a key role in T-ALL. TAL1 mutations were identified in a total of 6 tumors (3 trios). Besides the known oncogenic insertion at ~7kb upstream of TAL1 (1 trio), we also discovered a novel mutation in the first intron of TAL1 (2 trios) that may potentially drive ectopic TAL1 expression. Novel LMO2 mutations were found in 6 tumors (3 trios) at 3 different sites. This approach also identified novel targets such as GPSM1 and PTPRN2, shedding light on heretofore unknown players in T-ALL targeted by noncoding mutations. By integrating WGS and RNAseq data, our novel computational framework can identify somatic noncoding mutations in cancer that may introduce de novo regulatory elements, thus driving target gene expression. Wider application of this approach will expand our understanding of the role of noncoding mutations in the cancer genome. Citation Format: Yu Liu, Michael Edmonson, Xiaotu Ma, Michael Rusch, Yongjin Li, Benshang Li, Shuhong Shen, A. Thomas Look, Jinghui Zhang. Computational approach for discovery of regulatory noncoding variants in cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1574. doi:10.1158/1538-7445.AM2017-1574