Joint learning of multiple gene networks from single-cell gene expression data.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Gene network inference is crucial for understanding cellular functions.
• Single-cell RNA sequencing (scRNA-seq) enables gene network analysis at the single-cell level.
• Traditional methods struggle with scRNA-seq data's sparsity and heterogeneity.

Purpose of the Study:

• To develop a novel method for inferring gene networks from scRNA-seq data.
• To address challenges posed by cellular heterogeneity and data sparsity (dropout events).
• To identify both common and unique gene network structures across different cell subgroups.

Main Methods:

• Introduction of a joint Gaussian copula graphical model (JGCGM).
• The model jointly estimates multiple gene networks for multiple cell subgroups.
• JGCGM is designed to handle non-Gaussian data with missing values.

Main Results:

• JGCGM outperforms existing state-of-the-art network inference models on synthetic data.
• Application to real scRNA-seq data reveals gene networks for different cell subgroups.
• Identified hub genes within the inferred networks exhibit significant biological relevance.

Conclusions:

• The proposed JGCGM is a suitable and effective method for gene network inference from scRNA-seq data.
• The model successfully captures complex network structures, including common and unique patterns.
• This approach facilitates deeper understanding of cellular organization and function through gene network analysis.