Capture of TAD reorganization endows variant patterns of gene transcription (Hi-C)

TAD (Topologically Associating Domain) reorganization occurs commonly in cell nucleus and contributes to gene activation and inhibition through separation or fusion of adjacent TADs. However, identification of functional genes impacted by TAD alteration and revealing of TAD-reorganization mechanism underlying gene transcription remain to be fully elucidated. Here, we firstly developed a novel approach termed Inter3D to specifically dig out genes regulated by TAD reorganization. By constructing the comprehensive TAD reorganization medicating epigenomic landscapes in tumor cells, we showed that TAD separation interrupted the intrachromosomal loop at MYL12B locus and abrogated its transcription, while TAD fusion generated the chromosomal interaction and triggered CYP27B1 activation. Our study provides a comprehensive insight into capture of TAD rearrangement-mediated gene loci and moves towards recognizing the functional role of TAD reorganization in gene transcription. Overall design: To address the questions about TAD reorganization, we developed a novel easy-to-use pipeline for Interactome identification with 3D multi-genomics (Inter3D) by performing integrative analysis of multi-omics datasets, including 3D genomics, epigenomics and transcriptomics. TAD boundaries are rich in various architectural proteins, among which we paid more attention to CTCF regulated ones. In details, Inter3D first identifies the CTCF-medicated TAD boundary alterations using Hi-C and CTCF ChIP-seq datasets of normal ARPE19 cells and cancer WERI-RB1 cells. Next, incorporating with the RNA-seq datasets of ARPE19 and WERI-RB1 cell lines, Inter3D identifies the candidate DEGs between variant samples regulated by CTCF binding alteration and TAD alteration. Finally, Inter3D identifies the candidate enhancers in promoter-distal peaks robust in ATAC-seq datasets of ARPE19 and WERI-RB1 cell lines, ruling out false positive target genes regulated by enhancer only rather than 3D structure.