EAGLE: An algorithm that utilizes a small number of genomic features to predict tissue/cell type-specific enhancer-gene interactions

• DOI: 10.1371/journal.pcbi.1007436
• Publication Date: 2019-10-30T17:28:34Z
• Authors: Tianshun Gao; Jiang Qian
• Keywords: Identification

Long-range regulation by distal enhancers is crucial for many biological processes. The existing methods enhancer-target gene prediction often require genomic features. This makes them difficult to be applied cell types, in which the relevant datasets are not always available. Here, we design a tool EAGLE, an enhancer and learning ensemble method identification of Enhancer-Gene (EG) interactions. Unlike tools, EAGLE used only six features derived from expression datasets. Cross-validation revealed that outperformed other methods. Enrichment analyses on special transcriptional factors, epigenetic modifications, eQTLs demonstrated could distinguish interacting pairs non- ones. Finally, was mouse human genomes identified 7,680,203 7,437,255 EG interactions involving 31,375 43,724 genes, 138,547 177,062 across 89 110 tissue/cell types human, respectively. obtained accessible through interactive database enhanceratlas.org. available at https://github.com/EvansGao/EAGLE predicted http://www.enhanceratlas.org/.