LineGRN: A Line Graph Neural Network for Gene Regulatory Network Inference.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Gene regulatory networks (GRNs) are crucial for understanding cellular processes.
• Single-cell RNA sequencing (scRNA-seq) enables GRN inference at high resolution.
• Existing methods struggle with capturing gene pair associations and network topology.

Purpose of the Study:

• To develop a novel computational framework, LineGRN, for inferring GRNs from scRNA-seq data.
• To address limitations in existing methods regarding association patterns and network topology.
• To enhance information propagation within inferred GRNs.

Main Methods:

• Proposed LineGRN, a line graph neural network framework.
• Modeled neighborhood relationships between gene pairs to preserve interaction signals.
• Utilized line graph transformation to create a high-degree-node-dominated local topology.

Main Results:

• LineGRN significantly outperformed seven state-of-the-art methods on real datasets.
• The method demonstrated low sensitivity to parameter variations and noise.
• Case studies validated LineGRN's ability to uncover potential transcription factor-target associations.

Conclusions:

• LineGRN offers a superior approach for inferring GRNs from scRNA-seq data.
• The framework enhances topological properties for efficient information propagation.
• LineGRN provides a robust and accurate tool for regulatory network analysis.