Genome-wide detection of enhancer-hijacking events from Hi-C data in re-arranged genomes

Although recent efforts have shown that large structural variations (SVs) can disrupt the 3D genome organization in cancer cells and lead to the formation of neo-TADs (topologically associating domains, TADs) and enhancer-hijacking events, no computational tools have been developed to detect such events genome-wide. Here we develop a computational pipeline to identify the chromatin interactions induced by SVs such as inter-chromosomal translocations, large deletions, and inversions. Our framework works by reconstructing local Hi-C maps surrounding the breakpoints, normalizing copy number variation (CNV) and allele effects, and detecting local optimal signals. We also built in modules for 3D structure inference and interactive visualization of complex SV assemblies. By applying our pipeline to Hi-C data from 54 cancer cell lines and primary tumors, we identified tens of recurrent enhancer-hijacking genes in different cancer types, such as MYC in brain tumor, ETV1 in prostate adenocarcinoma and RAB36 in leukemia. To further validate our algorithm, we performed gene editing by CRISPR/Cas9 to delete hijacked enhancers and showed that their deletions led to the reduction of target gene expression in another chromosome. In summary, our tool paves the way for identifying new tumorigenic mechanisms and thus new diagnostic and therapeutic targets.