LR Hunting: A Random Forest Based Cell-Cell Interaction Discovery Method for Single-Cell Gene Expression Data.

Area of Science:

• Computational Biology
• Genomics
• Immunology

Background:

• Cell-cell interactions (CCIs) and communication (CCC) are fundamental to biological systems.
• Single-cell RNA sequencing (scRNA-seq) enables the study of CCIs/CCCs by identifying ligand-receptor (LR) gene interactions.
• Existing methods often focus on individual gene pairs, limiting comprehensive analysis.

Purpose of the Study:

• To develop a novel computational approach, LR hunting, for comprehensive analysis of cell-cell interactions.
• To improve the analysis of ligand-receptor interactions within complex biological systems using scRNA-seq data.
• To overcome limitations of methods analyzing single gene pairs by enabling simultaneous analysis of all LR pairs.

Main Methods:

• Implemented a random forests (RFs)-based data imputation technique to link data across different cell types.
• Utilized an aggregated imputed minimal depth index (IMDI) for robust data imputation by repeating computations.
• Employed unsupervised RFs to simultaneously identify significant LR interactions among all possible LR pairs.

Main Results:

• The LR hunting approach successfully identified meaningful cell-cell interactions.
• Demonstrated the method's efficacy on a mouse CITE-seq dataset.
• Validated the approach using a triple-negative breast cancer scRNA-seq dataset.

Conclusions:

• LR hunting provides a robust and comprehensive method for deciphering cell-cell interactions from scRNA-seq data.
• The approach enhances the understanding of biological systems by enabling simultaneous analysis of numerous ligand-receptor interactions.
• This novel method facilitates the discovery of biologically relevant CCIs in various biological contexts.