977-CoCo-ST: for Spatial Transcriptomics Domain Detection

Researchers have developed CoCo-ST, a novel computational framework designed to identify both prominent and subtle biological structures within spatial transcriptomics data. By utilizing a graph contrastive learning approach that compares target tissues against background samples, the tool successfully isolates unique local features often missed by traditional algorithms. The study demonstrates that CoCo-ST excels at detecting early-stage precancerous regions and characterizing complex cell-cell communication patterns during tumor evolution. It effectively handles batch-effect correction, allowing for the seamless integration of multiple samples across different experimental conditions. Furthermore, the technology proves highly scalable, maintaining its accuracy across diverse platforms ranging from spot-level Visium data to high-resolution subcellular and single-cell imaging. Ultimately, CoCo-ST provides a robust method for visualizing tissue heterogeneity and tracing the progression of diseases like lung cancer. References: • Aminu M, Zhu B, Vokes N, et al. CoCo-ST detects global and local biological structures in spatial transcriptomics datasets[J]. Nature cell biology, 2025: 1-13.