OrgaCCC: Orthogonal graph autoencoders for constructing cell-cell communication networks on spatial transcriptomics...

Area of Science:

• Cellular Biology
• Bioinformatics
• Genomics

Background:

• Cell-cell communication (CCC) is vital for multicellular organism function, tissue homeostasis, and adaptation.
• Inferring CCC mechanisms from spatial transcriptomics (ST) data is challenging due to limitations in current computational methods relying on incomplete gene interaction lists.

Purpose of the Study:

• To develop an advanced computational method, OrgaCCC, for accurate cell-cell communication inference from spatial transcriptomics data.
• To overcome the limitations of existing methods by leveraging comprehensive biological information.

Main Methods:

• Proposed OrgaCCC, an orthogonal graph autoencoders approach utilizing deep generative models.
• Integrated gene expression profiles, spatial locations, and ligand-receptor relationships.
• Employed orthogonally coupled variational graph autoencoders for cell/spot and gene feature extraction and combined them via feature similarity maximization.

Main Results:

• OrgaCCC demonstrated superior performance in CCC inference compared to state-of-the-art methods across five ST datasets.
• Achieved higher accuracy and reliability at cell-type, cell/spot, and ligand-receptor levels.
• Effectively captured complex intercellular communication patterns within spatial contexts.

Conclusions:

• OrgaCCC offers a robust and accurate approach for inferring cell-cell communication from spatial transcriptomics data.
• The method provides valuable insights into biological processes by improving the understanding of intercellular signaling pathways.
• OrgaCCC represents a significant advancement in computational biology for analyzing complex biological systems.