Cell type-specific prediction of 3D chromatin organization enables high-throughput in silico genetic screening

Experimental methods for measuring 3D chromatin organization, such as Hi-C, are costly and have technical limitations, restricting their broad application particularly in high-throughput genetic perturbations. We present C.Origami, a deep neural network model that performs de novoprediction of cell type–specific chromatin organization using as inputs the DNA sequence and two cell type–specific genomic features — chromatin accessibility and CTCF binding. C.Origami predicts chromatin organization within a 2 mega-base window and enables in silico experiments to examine the impact of genetic perturbations on chromatin interactions in pathologies such as cancer. We assess how individual DNA elements contribute to the organization of 3D chromatin organization and identify a compendium of cell type–specific trans-regulators across multiple cell types. We demonstrate that cell type–specific in silico genetic perturbation and screening, enabled by C.Origami, can be used to systematically discover chromatin regulatory mechanisms in both normal and disease-related biological systems. Overall design: ATAC-seq is performed on CUTLL1 cell line to evaluate chromatin accessibility