CGLoop: a neural network framework for chromatin loop prediction

• DOI: 10.1186/s12864-025-11531-y
• Publication Date: 2025-04-05T06:12:35Z
• Authors: Junfeng Wang; Lili Wu; Jingjing Wei; Chaokun Yan; Huimin Luo; Junwei Luo; Fei Guo

Chromosomes of species exhibit a variety high-dimensional organizational features, and chromatin loops, which are fundamental structures in the three-dimensional (3D) structure genome. Chromatin loops visible speckled patterns on Hi-C contact matrix generated by chromosome conformation capture methods. The play an important role gene expression, predicting during whole genome interactions is crucial for deeper understanding 3D function. Here, we propose CGLoop, deep learning based neural network framework that detects matrix. CGLoop combines convolutional (CNN) with Convolutional Block Attention Module (CBAM) Bidirectional Gated Recurrent Unit (BiGRU) to features related comprehensively analyzing matrix, enabling prediction candidate loops. And employs density clustering method filter predicted model. Finally, compared CGloop other methods several cell line including GM12878, K562, IMR90, mESC. code available from https://github.com/wllwuliliwll/CGLoop . experimental results show that, high APA scores there enrichment multiple transcription factors binding proteins at anchors, outperforms terms accuracy validity prediction.