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

Area of Science:

• Genomics
• Systems Biology
• Computational Biology

Background:

• Distal enhancers play a critical role in regulating gene expression for various biological processes.
• Current enhancer-target gene prediction methods often require extensive genomic features, limiting their applicability across diverse cell types due to data availability constraints.

Purpose of the Study:

• To develop an efficient and broadly applicable computational tool for predicting enhancer-gene (EG) interactions.
• To overcome the limitations of existing methods by utilizing a reduced feature set.

Main Methods:

• Developed EAGLE (Enhancer and Gene Learning Ensemble), a machine learning ensemble method.
• EAGLE utilizes only six features derived from genomic and gene expression data.
• Validated performance through cross-validation and enrichment analyses (transcription factors, epigenetic modifications, eQTLs).

Main Results:

• EAGLE demonstrated superior performance compared to existing enhancer-target gene prediction methods.
• Enrichment analyses confirmed EAGLE's ability to accurately distinguish interacting from non-interacting EG pairs.
• Identified over 7.6 million (mouse) and 7.4 million (human) EG interactions across numerous tissue and cell types.

Conclusions:

• EAGLE provides a robust and efficient approach for predicting enhancer-gene interactions with minimal data requirements.
• The identified interactions offer valuable insights into gene regulation and are accessible via the EnhancerAtlas database.
• The EAGLE tool and predicted datasets are publicly available for further research.