Research map / Probabilistic and scientific modeling / Molecules and proteins
Multi-state Protein Design with DynamicMPNN
Alex Abrudan, Sebastian Pujalte Ojeda, Chaitanya K. Joshi, Matthew Greenig, Felipe Engelberger, Alena Khmelinskaia, Jens Meiler, Michele Vendruscolo, Tuomas Knowles
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 |
|---|---|---|---|
| train on multi-state conformation pairs jointly (Combined Training) instead of single-state data only (ablation) | DynamicMPNN · Multi-state benchmark (n=96) | decoy-normalized RMSD: 0.479 -> 0.200 | Improved |
| train on multi-state pairs instead of single-state only (ablation) | DynamicMPNN · Multi-state benchmark (n=96) | sequence recovery (%): 27.1 -> 41.0 | Improved |
| train on multi-state pairs instead of single-state only (ablation) | DynamicMPNN · Multi-state benchmark (n=96) | AF3 pLDDT: 62.59 -> 79.28 | Improved |
| use DynamicMPNN instead of ProteinMPNN MSD | inverse folding · Multi-state benchmark subset (n=61) | decoy-normalized RMSD: 0.275 -> 0.206 | Improved |
| use DynamicMPNN instead of ProteinMPNN MSD | inverse folding · Multi-state benchmark (n=96) | sequence recovery (%): 38.0 -> 42.7 | Improved |
| use DynamicMPNN instead of ProteinMPNN MSD | inverse folding · Multi-state benchmark subset (n=61) | decoy-normalized pLDDT: 1.416 vs 1.401 | No clear effect |
| add Deep Symmetric Set cross-attention between conformations (ablation) | DynamicMPNN · Multi-state benchmark (n=96) | sequence recovery (%): 41.0 -> 38.8 | Hurt |
| add DSS (ablation) | DynamicMPNN · Multi-state benchmark (n=96) | AF3 RMSD (A): 4.37 -> 4.38 | No clear effect |
| add single-state pretraining then multi-state finetuning vs combined training (ablation) | DynamicMPNN · Multi-state benchmark (n=96) | sequence recovery (%): 41.0 -> 42.7 | Improved |
Closest papers by meaning
- Sidechain conditioning and modeling for full-atom protein sequence design with FAMPNN (ICML 2025 · similarity 0.89)
- Learning to engineer protein flexibility (ICLR 2025 · similarity 0.88)
- Bridge-IF: Learning Inverse Protein Folding with Markov Bridges (NeurIPS 2024 · similarity 0.88)
- ProtInvTree: Deliberate Protein Inverse Folding with Reward-guided Tree Search (NeurIPS 2025 · similarity 0.87)
- Simultaneous Modeling of Protein Conformation and Dynamics via Autoregression (NeurIPS 2025 · similarity 0.87)
- Fast Uncovering of Protein Sequence Diversity from Structure (ICLR 2025 · 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.
