Research map / Large language models / Agents, code and math
Proteo-R1: Reasoning Foundation Models for De Novo Protein Design
Fang Wu, Weihao Xuan, Heli Qi, Hanqun Cao, Heng-Jui Chang, Zeqi Zhou, Li Erran Li, Haokai Zhao, Jian Ma, Zijian Carl Ma, Yu-Chi Cheng, Kuan Pang and 17 more
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 oracle (ground-truth) hotspot anchors (ablation) | Proteo-R1 · Multi-CDR redesign (SAbDab test) | RMSD CDR-H3: 3.81 -> 3.34 | Improved |
| add oracle hotspot anchors (ablation) | Proteo-R1 · Multi-CDR redesign (SAbDab test) | Loop-RMSD: 4.51 -> 3.93 | Improved |
| add oracle hotspot anchors (ablation) | Proteo-R1 · Multi-CDR redesign (SAbDab test) | IMP: 56.58 -> 62.25 | Improved |
| add oracle hotspot anchors (ablation) | Proteo-R1 · Multi-CDR redesign (SAbDab test) | JSD_bb: 0.2661 -> 0.2043 | Improved |
| add oracle hotspot anchors (ablation) | Proteo-R1 · Multi-CDR redesign (SAbDab test) | RMSD CDR-H1: 1.33 -> 1.43 | Hurt |
| add oracle hotspot anchors (ablation) | Proteo-R1 · Multi-CDR redesign (SAbDab test) | Clash_out: 0.14% -> 0.23% | Hurt |
| add reasoning-guided anchors to UniMoMo generator | UniMoMo (all), 100 generations · CDR-H3 design, 100 generations (UniMoMo backend) | IMP: 65.00% -> 67.79% | Improved |
| add reasoning-guided anchors to UniMoMo generator | UniMoMo (all), 100 generations · CDR-H3 design, 100 generations (UniMoMo backend) | delta G: 8.46 -> 7.35 | Improved |
| add reasoning-guided anchors to UniMoMo generator | UniMoMo (all), 100 generations · CDR-H3 design, 100 generations (UniMoMo backend) | RMSD: 1.04 -> 0.83 | Improved |
| add reasoning-guided anchors to UniMoMo generator | UniMoMo (all), 100 generations · CDR-H3 design, 100 generations (UniMoMo backend) | AAR: 52.34% -> 48.94% | Hurt |
| add oracle anchors to Proteo-R1 (UniMoMo) (ablation) | Proteo-R1 (UniMoMo) · CDR-H3 design, 100 generations (UniMoMo backend) | IMP: 67.79% -> 74.5% | Improved |
| increase number of generated samples from 1 to 100 | UniMoMo (all) · CDR-H3 design, 100 generations (UniMoMo backend) | AAR: 21.44% -> 52.34% | Improved |
Closest papers by meaning
- ProSpero: Active Learning for Robust Protein Design Beyond Wild-Type Neighborhoods (NeurIPS 2025 · similarity 0.88)
- Protein Design with Dynamic Protein Vocabulary (NeurIPS 2025 · similarity 0.88)
- Protein Design with Agent Rosetta: A Case Study for Specialized Scientific Agents (ICML 2026 · similarity 0.87)
- Proteus: Exploring Protein Structure Generation for Enhanced Designability and Efficiency (ICML 2024 · similarity 0.87)
- BindGPT: A Scalable Framework for 3D Molecular Design via Language Modeling and Reinforcement Learning (AAAI 2025 · similarity 0.86)
- ProtCLIP: Function-Informed Protein Multi-Modal Learning (AAAI 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.
