scGTN: Deep Siamese Graph Transformer Network for Single-cell RNA Sequencing Clustering

2026-06-17 · Source: Artificial Intelligence · Field: Science & Research — Life Sciences & Biology, Artificial Intelligence & Machine Learning, Research Methodology & Innovation · Depth: Expert, quick

Summary

The scGTN (Deep Siamese Graph Transformer Network) is a novel framework designed for single-cell RNA sequencing (scRNA-seq) data clustering, addressing the challenges of sparsity, noise, and complex intercellular structural information inherent in such data. This method explicitly integrates gene expression profiles with intercellular structural dependencies. scGTN formulates scRNA-seq data as a graph, then constructs two augmented graph views to capture complementary intercellular information. A Siamese graph transformer network is employed to incorporate shortest-path information and node-wise distances, enriching the understanding of cellular relationships. Finally, an optimal transport strategy guides the cell clustering in a self-supervised manner. Extensive experiments on multiple benchmark scRNA-seq datasets demonstrate scGTN's consistent outperformance compared to existing methods. The code is publicly available.

Key takeaway

For research scientists developing single-cell RNA sequencing analysis pipelines, scGTN offers a robust approach to overcome data sparsity and noise. You should consider integrating graph-based transformer networks and optimal transport strategies to explicitly model intercellular structural dependencies. This method can significantly enhance cell type identification accuracy, providing a more nuanced understanding of cellular heterogeneity compared to traditional clustering techniques. Evaluate scGTN's open-source implementation for your specific benchmark datasets.

Key insights

scGTN integrates gene expression and intercellular structure via a Siamese graph transformer for robust scRNA-seq clustering.

Principles

• Explicitly integrate structural dependencies.
• Use dual graph views for complementary data.
• Guide clustering with optimal transport.

Method

Formulate scRNA-seq data as a graph. Construct two augmented graph views. Employ a Siamese graph transformer network to capture shortest-path and node-wise distances. Guide clustering via optimal transport.

In practice

• Apply graph transformers to scRNA-seq.
• Utilize self-supervised optimal transport.
• Integrate gene expression with cell structure.

Topics

• Single-cell RNA Sequencing
• Graph Transformer Networks
• Siamese Neural Networks
• Optimal Transport
• Cell Clustering
• Gene Expression Profiling

Code references

• W-RMSL/scGTN

Best for: AI Scientist, Research Scientist

Related on AIssential

• Identifying spatial single-cell-level interactions with graph transformer — A self-supervised graph transformer identifies spatial single-cell interactions without known ligand-receptor pairs.
• scLong: a billion-parameter foundation model for capturing long-range gene context in single-cell transcriptomics — scLong is a billion-parameter foundation model for scRNA-seq, capturing long-range gene dependencies and integrating external gene knowledge.
• Dissecting epigenetic heterogeneity in single-cell DNA methylomes with a unified framework — scMethCraft accurately models single-cell DNA methylation data by addressing its unique characteristics with a hybrid neural network.
• EssentialGIN: a new approach for gene essentiality prediction based on graph isomorphism neural networks — Graph isomorphism networks, by integrating biological data and preserving network topology, significantly improve essential gene prediction accuracy.
• OCOO-T : A Simple and Scalable Virtual Cell Model for Transcriptional Perturbation Response Prediction — OCOO-T uses a simple Transformer and flow matching to directly denoise gene expression for perturbation prediction.
• scHelix: Asymmetric Dual-Stream Integration via Explicit Gene-Level Disentanglement — scHelix integrates scRNA-seq data by disentangling genes into invariant Anchors and sensitive Variants to preserve biological fidelity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.