scDCA: deciphering the dominant cell communication assembly of downstream functional events from single-cell RNA-seq...

Area of Science:

• Computational biology
• Single-cell genomics
• Cancer research

Background:

• Cell-cell communications (CCCs) are crucial for biological processes, but computational methods to quantify their impact on receiver cells are limited.
• Understanding complex CCCs is vital for deciphering cancer progression mechanisms and identifying therapeutic targets.

Purpose of the Study:

• To develop a novel deep learning-based method, scDCA, for quantifying the contribution of cell type combinations to specific functional processes in receiver cells.
• To identify dominant cell communication assemblies (DCAs) influencing gene expression and cell states using single-cell RNA-seq data.

Main Methods:

• Proposed scDCA, a deep learning model utilizing a multi-view graph convolution network to reconstruct the CCCs landscape at single-cell resolution.
• Employed an attention mechanism within scDCA to interpret and identify DCAs impacting specific functional events.
• Applied scDCA to single-cell RNA-seq data from advanced renal cell carcinoma and immunotherapy-treated patients.

Main Results:

• Successfully identified DCAs affecting crucial gene expression in immune cells within renal cell carcinoma samples.
• Revealed DCAs responsible for variations in 14 functional states of malignant cells.
• Explored alterations in CCCs under clinical intervention, comparing DCAs related to cytotoxic factors in patients with and without immunotherapy.

Conclusions:

• scDCA provides a valuable tool for deciphering cell type combinations with dominant impacts on receiver cell functions.
• The findings highlight the significance of scDCA for understanding cancer biology and advancing precise cancer treatment strategies.
• The developed method and associated data are publicly available for further research.