Research map / Probabilistic and scientific modeling / Molecules and proteins
Unified Biomolecular Trajectory Generation via Pretrained Variational Bridge
Ziyang Yu, Wenbing Huang, Yang Liu
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 RL finetuning via adjoint matching to PVB (ablation) | PVB · PDBBind test (23 systems) | mean ligand RMSD: 6.904 -> 5.918 | Improved |
| add RL finetuning via adjoint matching to PVB (ablation) | PVB · PDBBind test (23 systems) | ligand RMSD <5A: 26.1% -> 43.5% | Improved |
| add RL finetuning via adjoint matching to PVB (ablation) | PVB · PDBBind test (23 systems) | mean pocket RMSD: 2.718 -> 1.965 | Improved |
| add RL finetuning to PVB versus AutoDock Vina docked pose | PVB · PDBBind test (23 systems) | mean ligand RMSD: 6.368 -> 5.918 | Improved |
| PVB without RL versus Vina initial pose | PVB · PDBBind test (23 systems) | mean ligand RMSD: 6.368 -> 6.904 | Hurt |
| add RL finetuning versus Vina on pocket RMSD | PVB · PDBBind test (23 systems) | mean pocket RMSD: 1.211 -> 1.965 | Hurt |
| replace UniSim with PVB (pretrained variational bridge) | PVB · ATLAS test (14 proteins) | VAL-CA: 0.106 -> 0.975 | Improved |
| replace UniSim with PVB | PVB · ATLAS test (14 proteins) | Decorr-TIC0: 0.786 -> 0.929 | Improved |
| replace UniSim with PVB | PVB · ATLAS test (14 proteins) | JSD MSM: 0.344 -> 0.333 | Improved |
| replace UniSim with PVB | PVB · ATLAS test (14 proteins) | RMSE-CONTACT: 0.129 -> 0.143 | Hurt |
| replace UniSim with PVB | PVB · mdCATH test (20 domains) | VAL-CA: 0.522 -> 0.963 | Improved |
| replace UniSim with PVB | PVB · mdCATH test (20 domains) | JSD TIC: 0.454 -> 0.376 | Improved |
Closest papers by meaning
- Generative Modeling of Molecular Dynamics Trajectories (NeurIPS 2024 · similarity 0.91)
- Action-Minimization Meets Generative Modeling: Efficient Transition Path Sampling with the Onsager-Machlup Functional (ICML 2025 · similarity 0.90)
- Atomic Trajectory Modeling with State Space Models for Biomolecular Dynamics (NeurIPS 2026 · similarity 0.89)
- Few-Step Cofolding with All-Atom Flow Maps (NeurIPS 2026 · similarity 0.88)
- Energy-Based Flow Matching for Generating 3D Molecular Structure (ICML 2025 · similarity 0.88)
- Swallowing the Bitter Pill: Simplified Scalable Conformer Generation (ICML 2024 · similarity 0.87)
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.
