CeSpGRN: inferring cell-specific gene regulatory networks from single-cell multi-omics and spatial data

Area of Science:

• Genomics
• Computational Biology
• Systems Biology

Background:

• Single-cell sequencing reveals cell-to-cell variations driven by gene regulatory networks (GRNs).
• Existing methods often infer population-level GRNs, failing to capture cell-specific regulatory dynamics.
• Reconstructing dynamic and cell-specific GRNs is crucial for understanding cellular heterogeneity.

Purpose of the Study:

• To develop a computational method for inferring cell-specific gene regulatory networks (GRNs).
• To enable the reconstruction of GRNs that capture regulatory rewiring at the single-cell level.
• To leverage multi-omic and spatial data for enhanced GRN inference.

Main Methods:

• CeSpGRN (Cell Specific Gene Regulatory Network inference) utilizes a kernel-weighted Gaussian Copula Graphical Model.
• Incorporates single-cell resolution ATAC-seq data, unlike population-level approaches.
• Integrates multi-omic or spatial transcriptomic data into the objective function for improved inference.

Main Results:

• CeSpGRN demonstrates superior performance in reconstructing cell-specific GRNs compared to baseline methods.
• The method accurately detects regulatory interactions that differ between individual cells.
• Uncovered rewiring of regulatory interactions during biological processes using real-world datasets.

Conclusions:

• CeSpGRN provides a powerful tool for inferring cell-specific GRNs from various single-cell data types.
• The method advances the understanding of gene regulation dynamics at a single-cell resolution.
• Enables the discovery of novel regulatory mechanisms driving cellular heterogeneity and biological processes.