ChromaFold predicts the 3D contact map from single-cell chromatin accessibility
Vianne R. Gao, Rui Yang, Arnav Das, Renhe Luo, Hanzhi Luo, Dylan McNally, Ioannis Karagiannidis, Martín A. Rivas, Zhong-Min Wang, Darko Barišić, Alireza Karbalayghareh, Wilfred Wong, Yingqian A. Zhan, Christopher R. Chin, William Stafford Noble, Jeff Bilmes, Effie Apostolou, Michael G. Kharas, Wendy Béguelin, Aaron D. Viny, Danwei Huangfu, Alexander Y. Rudensky, Ari Melnick, Christina S. Leslie
Abstract
Identifying cell-type-specific 3D chromatin interactions between regulatory elements can help decipher gene regulation and interpret disease-associated non-coding variants. However, achieving this resolution with current 3D genomics technologies is often infeasible given limited input cell numbers. We therefore present ChromaFold, a deep learning model that predicts 3D contact maps, including regulatory interactions, from single-cell ATAC sequencing (scATAC-seq) data alone. ChromaFold uses pseudobulk chromatin accessibility, co-accessibility across metacells, and a CTCF motif track as inputs and employs a lightweight architecture to train on standard GPUs. Trained on paired scATAC-seq and Hi-C data in human samples, ChromaFold accurately predicts the 3D contact map and peak-level interactions across diverse human and mouse test cell types. Compared to leading contact map prediction models that use ATAC-seq and CTCF ChIP-seq, ChromaFold achieves state-of-the-art performance using only scATAC-seq. Finally, fine-tuning ChromaFold on paired scATAC-seq and Hi-C in a complex tissue enables deconvolution of chromatin interactions across cell subpopulations. Obtaining a high-resolution contact map using current 3D genomics technologies can be challenging with small input cell numbers. Here, the authors develop ChromaFold, a deep learning model that predicts cell-type-specific 3D contact maps from single-cell chromatin accessibility data alone.