HiCNN2 is an improved version of our previously developed tool HiCNN for enhancing resolution of Hi-C data and uses three architectures to learn the mapping between low-resolution and high-resolution Hi-C contact matrices. HiCNN2-1 uses one convolutional neural network (ConvNet1); HiCNN2-2 consists of an ensemble of two different ConvNets (ConvNet1 and ConvNet2); HiCNN2-3 uses an ensemble of three different ConvNets (ConvNet1, ConvNet2, and ConvNet3).


Source code (Python) for HiCNN2 can be downloaded here (168M).

How to generate input files for HiCNN2 prediction from .hic files can be found here.


  1. Tong Liu and Zheng Wang. HiCNN2: Enhancing the Resolution of Hi-C Data Using an Ensemble of Convolutional Neural Networks. Genes, 2019, 10(11):862.

  2. Tong Liu and Zheng Wang. HiCNN: A very deep convolutional neural network to better enhance the resolution of Hi-C data. Bioinformatics, 2019, 35(21):4222-4228.


For any questions or suggestions, please contact:
Dr. Zheng Wang
Department of Computer Science
University of Miami