Jun Zhang

St2cell reconstructs in situ single-cell spatial transcriptomics by integrating high-resolution histological images with deep learning and convex quadratic programming.

The paper introduces MLA-GNN, a graph neural network that discovers clinically interpretable pathway-level biomarkers by integrating multi-level graph features and a saliency mechanism, achieving state-of-the-art results on transcriptomic and proteomic data.