Exploring chromatin conformation and gene co-expression through graph embedding

• DOI: 10.1093/bioinformatics/btaa803
• Publication Date: 2020-10-17T19:10:16Z
• Authors: Marco Varrone; Luca Nanni; Giovanni Ciriello; Stefano Ceri
• Keywords: ENCODE; Chromosome conformation capture; ChIA-PET; Graph Embedding; Gene regulatory network

Abstract Motivation The relationship between gene co-expression and chromatin conformation is of great biological interest. Thanks to high-throughput chromosome capture technologies (Hi-C), researchers are gaining insights on the tri-dimensional organization genome. Given high complexity Hi-C data difficult definition networks, development proper computational tools investigate such rapidly interest researchers. One most fascinating questions in this context how topology correlates with which physical interaction patterns predictive relationships. Results To address these questions, we developed a framework for prediction networks from data. We first define network where each associated its profile; then, apply two graph embedding techniques extract low-dimensional vector representation network; finally, train classifier pairs predict if they co-expressed. Both outperform previous methods based manually designed topological features, highlighting need more advanced strategies encode information. also establish that recent technique, random walks, superior. Overall, our results demonstrate regulation share non-linear embeddings relevant information, could be used downstream analysis. Availability implementation source code analysis available at: https://github.com/marcovarrone/gene-expression-chromatin. Supplementary information at Bioinformatics online.