Leveraging prior knowledge to infer gene regulatory networks from single-cell RNA-sequencing data

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) is vital for inferring gene regulatory networks (GRNs).
• scRNA-seq data's inherent noise and sparsity challenge accurate GRN inference.
• Prior knowledge integration offers a promising solution to enhance GRN reliability.

Purpose of the Study:

• To review methods for integrating prior knowledge into GRN inference from scRNA-seq data.
• To categorize prior knowledge types, representation methods, and GRN inference algorithms.
• To propose a standardized benchmarking framework for evaluating GRN inference algorithms.

Main Methods:

• Categorization of prior knowledge types (e.g., experimental data, databases).
• Discussion of methods for representing prior knowledge, especially using graph structures.
• Classification of GRN inference algorithms based on prior knowledge incorporation capabilities.

Main Results:

• Identification of various prior knowledge integration strategies.
• Assessment of algorithm performance in different contexts.
• Proposal of a standardized benchmarking framework for fair algorithm evaluation.

Conclusions:

• Prior knowledge integration significantly enhances GRN inference accuracy from scRNA-seq data.
• A standardized framework is needed for robust algorithm comparison and validation.
• This review guides researchers and developers in advancing GRN inference methodologies.