Abstract 2034: Patient-specific enhancer-gene networks for hundreds of primary cancer samples

Epigenomic alterations at regulatory elements have been implicated in cancer tumorigenesis. As a hallmark for epigenetic changes, chromatin accessibility profiles are generated for cancer patients. However, not all accessible regions are active regulatory elements. We developed a computational model that uses machine learning to integrate ATAC-seq and RNA-seq data and is trained using 3D genomic data. Our model is able to classify accessible regions into active regulatory elements with links to target genes and poised elements with no links to target genes. We used multi-omics data including histone modification profiles from ENCODE to validate our results. Application of our model on chromatin accessibility and transcriptomic data from 400 cancer patients of 22 cancer types revealed novel cancer-type and -subtype-specific enhancer-gene connections for known cancer genes. In one example, we identified two enhancers that regulate the expression of ESR1 in only ER+ breast cancer samples but not in ER- samples. These enhancers explain the high expression of ESR1 in 93% of ER+samples. Functional validation using CRISPRi confirms that inhibition of these enhancers decreases the expression of ESR1 in ER+ samples. Overall, our model predicts the connections of active enhancers to their target genes in a patient-type specific manner revealing key regulatory loci that modulate the expression of key genes. Citation Format: Duo Xu, Andre Neil Forbes, Sandra Cohen, Ann Palladino, Tatiana Karadimitriou, Ekta Khurana. Patient-specific enhancer-gene networks for hundreds of primary cancer samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2034.