Reference Regulatory Element-Guided Gene Expression Analysis for Mechanistic Inference of Gene Regulatory Networks

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• The depth-breadth gap in regulatory genomics limits mechanistic Gene Regulatory Network (GRN) analysis.
• Deep multi-omics data offers regulatory detail but lacks scalability.
• Broad gene expression datasets often lack the necessary regulatory structure.

Purpose of the Study:

• To develop an interpretable and scalable framework for GRN analysis using gene expression data.
• To infer transcription factor (TF)-regulatory element (RE)-gene programs.
• To bridge the gap between deep multi-omics and broad expression datasets.

Main Methods:

• Developed Regulatory Elements Guided Analysis (REGA), a computational framework.
• Utilized reference Regulatory Element (RE) catalogs to guide analysis.
• Applied REGA to diverse datasets including ChIP-seq, knockdown, Hi-C, eQTLs, snRNA-seq, spatial transcriptomics, and Perturb-seq.

Main Results:

• REGA prioritized functional REs and improved RE-gene and TF-gene inference compared to baselines.
• Successfully recovered coherent regulatory modules across benchmarks.
• Identified disease-associated modules and TF activities in PsychENCODE data.
• Linked regulatory dysregulation to genetic risk and detected neuronal-glial programs.
• Integrated cell-intrinsic programs with intercellular communication in spatial transcriptomics.
• Mapped perturbation responses to trait-associated regulatory architectures in Perturb-seq.

Conclusions:

• REGA provides a scalable and interpretable method for GRN analysis across various expression datasets.
• The framework effectively bridges the depth-breadth gap in regulatory genomics.
• REGA enhances the understanding of gene regulation in complex biological systems and diseases.