PETModule: a motif module based approach for enhancer target gene prediction

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Identifying enhancer-target gene (ETG) pairs is essential for understanding gene transcriptional regulation.
• Experimental methods like Hi-C provide valuable ETG pair data, but are costly.
• Existing computational methods for ETG prediction are suboptimal and difficult to use.

Purpose of the Study:

• To develop a novel, accurate, and accessible computational approach for predicting enhancer-target gene (ETG) pairs.
• To evaluate the performance of the new method against existing computational approaches and experimental data.

Main Methods:

• Developed PETModule, a motif module-based computational approach for predicting ETG pairs.
• Validated PETModule predictions using Hi-C and ChIA-PET experimental data across human and mouse cell types.
• Compared PETModule's performance with two recent ETG prediction methods.

Main Results:

• A significant portion of PETModule's predicted ETG pairs were supported by experimental evidence (Hi-C/ChIA-PET).
• PETModule demonstrated superior recall compared to two recently developed methods.
• PETModule achieved similar or better F1 scores and Area Under the ROC Curve (AUC) than existing approaches.

Conclusions:

• PETModule is an effective and accurate tool for predicting enhancer-target gene pairs.
• The developed method offers an improvement over existing computational approaches for ETG identification.
• PETModule provides a valuable, freely available resource for researchers studying gene regulation.