CENTRE: A gradient boosting algorithm for Cell-type-specific ENhancer-Target pREdiction

Motivation Identifying target promoters of active enhancers is a crucial step for realizing gene regulation and deciphering phenotypes and diseases. Up to now, several computational methods were developed to predict enhancer gene interactions but they require either many epigenomic and transcriptomic experimental assays to generate cell-type-specific predictions or a single experiment applied to a large cohort of cell types to extract correlations between activities of regulatory elements. Thus, inferring cell-type-specific enhancer gene interactions in unstudied or poorly annotated cell types becomes a laborious and costly task. Results Here, we aim to infer cell-type-specific enhancer target interactions, using minimal experimental input. We introduce CENTRE, a machine learning framework that predicts enhancer target interactions in a cell-type-specific manner, using only gene expression and ChIP-seq data for three histone modifications for the cell type of interest. CENTRE exploits the wealth of available datasets and extracts cell-type agnostic statistics to complement the cell-type specific information. CENTRE is thoroughly tested across many datasets and cell types and achieves equivalent or superior performance than existing algorithms that require massive experimental data. Availability CENTRE’s open source code is available at GitHub via https://github.com/slrvv/CENTRE