Improving Protein Sequence Design through Designability Preference Optimization
By: Fanglei Xue , Andrew Kubaney , Zhichun Guo and more
Potential Business Impact:
Designs proteins that reliably fold into shapes.
Protein sequence design methods have demonstrated strong performance in sequence generation for de novo protein design. However, as the training objective was sequence recovery, it does not guarantee designability--the likelihood that a designed sequence folds into the desired structure. To bridge this gap, we redefine the training objective by steering sequence generation toward high designability. To do this, we integrate Direct Preference Optimization (DPO), using AlphaFold pLDDT scores as the preference signal, which significantly improves the in silico design success rate. To further refine sequence generation at a finer, residue-level granularity, we introduce Residue-level Designability Preference Optimization (ResiDPO), which applies residue-level structural rewards and decouples optimization across residues. This enables direct improvement in designability while preserving regions that already perform well. Using a curated dataset with residue-level annotations, we fine-tune LigandMPNN with ResiDPO to obtain EnhancedMPNN, which achieves a nearly 3-fold increase in in silico design success rate (from 6.56% to 17.57%) on a challenging enzyme design benchmark.
Similar Papers
Protein Inverse Folding From Structure Feedback
Machine Learning (CS)
Designs proteins that fold into specific shapes.
De Novo Molecular Design Enabled by Direct Preference Optimization and Curriculum Learning
Machine Learning (CS)
Finds new medicines faster and cheaper.
AnnoDPO: Protein Functional Annotation Learning with Direct Preference Optimization
Biomolecules
Helps computers understand what proteins do better.