Multi-state Protein Design with DynamicMPNN
This addresses a key limitation in structural biology for designing functional multi-state proteins, though it appears incremental as an extension of existing inverse folding methods.
The paper tackles the problem of designing protein sequences that are compatible with multiple conformational states, which is crucial for biological processes like enzyme catalysis and membrane transport. DynamicMPNN, an inverse folding model trained on conformational ensembles, outperforms ProteinMPNN by up to 25% on decoy-normalized RMSD and 12% on sequence recovery in multi-state protein benchmarks.
Structural biology has long been dominated by the one sequence, one structure, one function paradigm, yet many critical biological processes - from enzyme catalysis to membrane transport - depend on proteins that adopt multiple conformational states. Existing multi-state design approaches rely on post-hoc aggregation of single-state predictions, achieving poor experimental success rates compared to single-state design. We introduce DynamicMPNN, an inverse folding model explicitly trained to generate sequences compatible with multiple conformations through joint learning across conformational ensembles. Trained on 46,033 conformational pairs covering 75% of CATH superfamilies and evaluated using Alphafold 3, DynamicMPNN outperforms ProteinMPNN by up to 25% on decoy-normalized RMSD and by 12% on sequence recovery across our challenging multi-state protein benchmark.