Research map / Probabilistic and scientific modeling / Molecules and proteins
RNA-FM: Flow-Matching Generative Model for Genome-wide RNA-Seq Prediction
Yaxuan Song, Jianan Fan, Tianyi Wang, Qiuyue Hu, Hang Chang, Heng Huang, Weidong Cai
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 pathway-aware patchify (ablation) | RNA-FM · TCGA-LUAD (T50) | PCC: 0.831 -> 0.865 | Improved |
| add pathway-aware patchify (ablation) | RNA-FM · TCGA-LUAD (T1000) | PCC: 0.679 -> 0.729 | Improved |
| add pathway-aware patchify (ablation) | RNA-FM · TCGA-LUAD (T1000) | RMSE: 0.097 -> 0.074 | Improved |
| add inter-pathway graph attention (ablation) | RNA-FM · TCGA-LUAD (T50) | PCC: 0.852 -> 0.865 | Improved |
| add inter-pathway graph attention (ablation) | RNA-FM · TCGA-LUAD (T1000) | PCC: 0.716 -> 0.729 | Improved |
| use learning rate 1e-4 instead of 2e-4 (ablation) | RNA-FM · TCGA-LUAD (T1000) | PCC: 0.701 -> 0.729 | Improved |
| use learning rate 1e-4 instead of 2e-4 (ablation) | RNA-FM · TCGA-LUAD (T100) | PCC: 0.801 -> 0.834 | Improved |
| use linear instead of logistic interpolant (ablation) | RNA-FM · TCGA-LUAD (T1000) | PCC: 0.709 -> 0.729 | Improved |
| use linear instead of logistic interpolant (ablation) | RNA-FM · TCGA-LUAD (T50) | RMSE: 0.058 -> 0.050 | Improved |
| add classifier-free guidance (s=2) (ablation) | RNA-FM · TCGA-LUAD (T1000) | PCC: 0.697 -> 0.729 | Improved |
| add classifier-free guidance (s=2) (ablation) | RNA-FM · TCGA-LUAD (T1000) | RMSE: 0.098 -> 0.074 | Improved |
| raise guidance scale from 2 to 3 (ablation) | RNA-FM · TCGA-LUAD (T1000) | PCC: 0.729 -> 0.718 | Hurt |
Later papers that compared with this method
1 later accepted paper reports RNA-FM as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- RMSAGen: Integrating Multiple Sequence Alignment for Function RNA Design (AAAI 2026 · better on 6 of 6 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GeneFlow: Translation of Single-cell Gene Expression to Histopathological Images via Rectified Flow (NeurIPS 2025 · similarity 0.86)
- Diffusion Generative Modeling for Spatially Resolved Gene Expression Inference from Histology Images (ICLR 2025 · similarity 0.86)
- Flow Matching for Count Data (NeurIPS 2026 · similarity 0.85)
- Modeling Dense Multimodal Interactions Between Biological Pathways and Histology for Survival Prediction (CVPR 2024 · similarity 0.85)
- Scalable Generation of Spatial Transcriptomics from Histology Images via Whole-Slide Flow Matching (ICML 2025 · similarity 0.84)
- BEACON: Benchmark for Comprehensive RNA Tasks and Language Models (NeurIPS 2024 · similarity 0.83)
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.
