Multi-state Protein Design with DynamicMPNN

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Abrudan, Alex, Ojeda, Sebastian Pujalte, Joshi, Chaitanya K., Greenig, Matthew, Engelberger, Felipe, Khmelinskaia, Alena, Meiler, Jens, Vendruscolo, Michele, Knowles, Tuomas P. J.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911210592534528
author Abrudan, Alex
Ojeda, Sebastian Pujalte
Joshi, Chaitanya K.
Greenig, Matthew
Engelberger, Felipe
Khmelinskaia, Alena
Meiler, Jens
Vendruscolo, Michele
Knowles, Tuomas P. J.
author_facet Abrudan, Alex
Ojeda, Sebastian Pujalte
Joshi, Chaitanya K.
Greenig, Matthew
Engelberger, Felipe
Khmelinskaia, Alena
Meiler, Jens
Vendruscolo, Michele
Knowles, Tuomas P. J.
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2507_21938
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-state Protein Design with DynamicMPNN
Abrudan, Alex
Ojeda, Sebastian Pujalte
Joshi, Chaitanya K.
Greenig, Matthew
Engelberger, Felipe
Khmelinskaia, Alena
Meiler, Jens
Vendruscolo, Michele
Knowles, Tuomas P. J.
Machine Learning
Biomolecules
I.2.6; J.3
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.
title Multi-state Protein Design with DynamicMPNN
topic Machine Learning
Biomolecules
I.2.6; J.3
url https://arxiv.org/abs/2507.21938