Abstract 5900: Integrating prior knowledge of transcription factor occupied elements with transcriptome-wide association analysis identifies 153 breast cancer susceptibility genes

Abstract

Listen

Translate article

Abstract Transcriptome-wide association studies (TWAS) have been successful in uncovering large numbers of putative disease susceptibility genes. TWAS approaches such as S-PrediXcan, FUSION and TIGAR, show comparable overall performance for susceptibility gene discovery; however, integrating prior disease-specific regulatory information is lacking. To improve accuracy of gene expression prediction by integrating information on susceptible transcription factor (sTF) occupancy in target cells and tissues, we developed an analytical framework, called sTF-TWAS. We evaluated six sets of putative regulatory variants selected from sTF-occupied sites, and performed generalized Berk-Jones (GBJ) analysis to combine sets and quantify overall associations. We demonstrated that sTF-TWAS outperforms S-PrediXcan in both simulation and real data analyses. Using summary statistics data from the Breast Cancer Association Consortium (BCAC), we also incorporated alternative splicing (sp) data in TF-TWAS and identified 76 and 88 significant genes for breast cancer risk at a Bonferroni-corrected P <0.05 from sTF-TWAS and sp-sTF-TWAS, respectively. Of 153 putative susceptibility genes from both sTF-TWAS and sp-sTF-TWAS, 14 genes in 12 loci had novel associations with breast cancer risk and 90 genes were previously unreported by breast cancer GWAS. We also provide strong evidence for 15 previously unreported genes, including 11 essential for breast cell proliferation and five suggested cancer predisposition or driver genes: RAD50, IGHMBP2, ATXN3, KANSL1, and IL6ST. Overall, we demonstrated that our approach has significant advantages over regular TWAS and our findings provide additional insights into the genetic susceptibility of breast cancer. Citation Format: Xingyi Guo, Jingni He, Alicia Beeghly-Fadiel, Zhishan Chen, Chen Cao, Xiao-ou Shu, Wei Zheng, Quan Long. Integrating prior knowledge of transcription factor occupied elements with transcriptome-wide association analysis identifies 153 breast cancer susceptibility genes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5900.