Research map / Label-efficient and robust learning / Recognition and long-tail
Predicting Spatial Transcriptomics from Histology Images via High-Order Multi-Cell Interaction Modeling
Youhan Sun, Jiahua Rao, Kangrui Du, Jiancong Xie, Yuedong Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add many-body attention to pairwise attention (hierarchical coupling) (ablation) | MCToGene · HEST-1k LUNG | PCC: 0.608 -> 0.636 | Improved |
| add many-body attention to pairwise attention (hierarchical coupling) (ablation) | MCToGene · HEST-1k LYMPH | PCC: 0.302 -> 0.316 | Improved |
| add hierarchical coupling instead of simple stacking of Pair and MB (ablation) | MCToGene · HEST-1k LUNG | PCC: 0.617 -> 0.636 | Improved |
| use many-body attention only instead of pairwise only (ablation) | MCToGene · HEST-1k SKCM | PCC: 0.697 -> 0.678 | Hurt |
| use many-body attention only instead of pairwise only (ablation) | MCToGene · HEST-1k LUNG | PCC: 0.608 -> 0.624 | Improved |
| combine Pair and MB without coupling (ablation) | MCToGene · HEST-1k READ | PCC: 0.240 -> 0.240 | No clear effect |
| add many-body attention costs memory and time (ablation) | MCToGene · efficiency | memory (MB) and time (s/epoch): 6166 -> 8720 MB; 1.21 -> 2.01 s/epoch | Hurt |
| use MCToGene instead of TRIPLEX for efficiency | MCToGene · efficiency | memory (MB): 16368 -> 8720 | Improved |
| increase neighbors per spot K from 4 to 16 (ablation) | MCToGene · HEST-1k LUNG | PCC: 0.613 -> 0.636 | Improved |
| increase neighbors per spot K beyond 20 to 25 (ablation) | MCToGene · HEST-1k LUNG | PCC: 0.636 -> 0.606 | Hurt |
| use learned decoder aggregation instead of mean aggregation (ablation) | MCToGene · HEST-1k | PCC | Improved |
| use MCToGene instead of STFlow | MCToGene · HEST-1k | average PCC: 0.415 -> 0.435 | Improved |
Closest papers by meaning
- Scalable Generation of Spatial Transcriptomics from Histology Images via Whole-Slide Flow Matching (ICML 2025 · similarity 0.89)
- Diffusion Generative Modeling for Spatially Resolved Gene Expression Inference from Histology Images (ICLR 2025 · similarity 0.88)
- Towards Scalable Context-Aware Single-Cell Spatial Transcriptomics Prediction from Histology Images (NeurIPS 2026 · similarity 0.88)
- M2OST: Many-to-one Regression for Predicting Spatial Transcriptomics from Digital Pathology Images (AAAI 2025 · similarity 0.87)
- SToFM: a Multi-scale Foundation Model for Spatial Transcriptomics (ICML 2025 · similarity 0.86)
- MERGE: Multi-faceted Hierarchical Graph-based GNN for Gene Expression Prediction from Whole Slide Histopathology Images (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
