CellGAT: A GAT-Based Method for Constructing a Cell Communication Network Integrating Multiomics Information.

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Intercellular communication coordinates multicellular organism development via cell-to-cell signaling.
• Existing computational methods for inferring cell communication often overlook protein-protein interactions (PPIs), potentially limiting accuracy.

Purpose of the Study:

• To develop a novel computational framework, CellGAT, for accurate inference of intercellular communication networks.
• To integrate diverse biological data, including gene expression and PPIs, for robust cell communication analysis.

Main Methods:

• CellGAT integrates gene expression, PPIs, protein complex, and pathway information.
• It employs node embedding and graph attention networks on single-cell RNA sequencing (scRNA-seq) data.
• A built-in cell clustering algorithm is included for comprehensive analysis.

Main Results:

• CellGAT accurately predicts cell-cell communication (CCC) by leveraging PPI evidence.
• The framework effectively analyzes the impact of CCC on downstream pathways and neighboring cells.
• CellGAT demonstrates utility in analyzing effects related to drug interventions.

Conclusions:

• CellGAT provides a powerful, integrated approach for inferring cell communication networks.
• The framework's ability to incorporate PPIs improves prediction accuracy, especially without prior cell type information.
• CellGAT offers valuable insights into cellular coordination and potential therapeutic strategies.