DeepChIA-PET: Accurately predicting ChIA-PET from Hi-C and ChIP-seq with deep dilated networks

Abstract

Chromatin interaction analysis by paired-end tag sequencing (ChIA-PET) can capture genome-wide chromatin interactions mediated by a specific DNA-associated protein. The ChIA-PET experiments have been applied to explore the key roles of different protein factors in chromatin folding and transcription regulation. However, compared with widely available Hi-C and ChIP-seq data, there are not many ChIA-PET datasets available in the literature. A computational method for accurately predicting ChIA-PET interactions from Hi-C and ChIP-seq data is needed that can save the efforts of performing wet-lab experiments. Here we present DeepChIA-PET, a supervised deep learning approach that can accurately predict ChIA-PET interactions by learning the latent relationships between ChIA-PET and two widely used data types: Hi-C and ChIP-seq. We trained our deep models with CTCF-mediated ChIA-PET of GM12878 as ground truth, and the deep network contains 40 dilated residual convolutional blocks. We first showed that DeepChIA-PET with only Hi-C as input significantly outperforms Peakachu, another computational method for predicting ChIA-PET from Hi-C but using random forests. We next proved that adding ChIP-seq as one extra input does improve the classification performance of DeepChIA-PET, but Hi-C plays a more prominent role in DeepChIA-PET than ChIP-seq. Our evaluation results indicate that our learned models can accurately predict not only CTCF-mediated ChIA-ET in GM12878 and HeLa but also non-CTCF ChIA-PET interactions, including RNA polymerase II (RNAPII) ChIA-PET of GM12878, RAD21 ChIA-PET of GM12878, and RAD21 ChIA-PET of K562. In total, DeepChIA-PET is an accurate tool for predicting the ChIA-PET interactions mediated by various chromatin-associated proteins from different cell types.

Predicting chromatin conformation contact maps

“…Similarly, Epiphany [20] predicts Hi-C contacts from 1D epigenomic data, including ChIP-seq measurements of chromatin accessibility, CTCF binding, and several types of histone modifications. Alternatively, DeepChIA-PET [21] predicts ChIA-PET contacts from a combination of Hi-C and ChIP-seq data. Despite being able to generalize to biosamples that they were not trained on, these methods are limited to making predictions for a single type of contact map and require that a fixed set of epigenomic experiments have been performed.…”

Learning Micro-C from Hi-C with diffusion models

“…Chromatin contact matrices can be predicted from one-dimensional (1D) factors, such as DNA sequence [24,25], epigenomic signals [26][27][28], and their hybrid [29]. Using 1D marks to predict 2D interactions usually results in generating massive false positive loops [30]. Based on the ChIP-seq matrix from the ENCODE project website, gathering various epigenomic marks for not-well-researched cell types as input is still a challenge.…”

EpiMCI: Predicting Multi-Way Chromatin Interactions from Epigenomic Signals

“…Pierro et al proposed an ensemble method, which used epigenetic marks as the input and followed an energy landscape model for chromatin organization to generate 3D chromosome conformations [ 28 ]. Liu et al presented DeepChIA-PET, a supervised deep learning approach to predict ChIA-PET interactions using Hi-C and epigenomic signals [ 29 ]. Yang et al designed Epiphany, a neural network to predict cell-type-specific Hi-C contact maps from widely available epigenomic tracks [ 30 ].…”