The adapted Activity-By-Contact model for enhancer-gene assignment and its application to single-cell data

Area of Science:

• Genomics
• Epigenomics
• Computational Biology

Background:

• Identifying genomic regulatory regions and their target genes is crucial for understanding cellular epigenomic landscapes.
• The Activity-By-Contact (ABC) model predicts enhancer-gene interactions but has limitations in accounting for all target genes and requires multiple assays.
• Existing methods lack integration with transcription factor (TF) binding data and efficient single-cell analysis capabilities.

Purpose of the Study:

• To generalize the ABC model for improved accuracy in predicting enhancer-gene interactions.
• To integrate transcription factor binding information for a more comprehensive regulatory analysis.
• To develop a computational pipeline for efficient single-cell epigenomic data analysis.

Main Methods:

• Adapted the ABC model to incorporate enhancer contact frequency and all gene transcription start sites.
• Integrated transcription factor binding affinities into the generalized ABC model.
• Developed the STARE computational pipeline for processing single-cell ATAC-seq data.

Main Results:

• The generalized ABC model demonstrated higher accuracy in predicting enhancer-gene interactions, comparable to the original model using only one assay.
• The STARE pipeline successfully analyzed single-cell ATAC-seq data from the human heart.
• Cell type-specific regulatory interactions were characterized, and gene expression was predicted based on TF affinities.

Conclusions:

• The enhanced ABC model and STARE pipeline provide a powerful tool for dissecting cell type-specific regulatory networks.
• This approach facilitates a deeper understanding of gene regulation in complex biological systems.
• The STARE software is publicly available for broader research application.