Protein Inverse Folding From Structure Feedback

TL;DR

This paper presents a novel method using Direct Preference Optimization to improve protein inverse folding models by leveraging structure feedback, significantly enhancing their ability to generate sequences that fold into desired structures.

Contribution

The paper introduces a DPO-based fine-tuning approach for inverse folding models that utilizes structure feedback to improve sequence design accuracy.

Findings

TM-Score increased from 0.77 to 0.81 after DPO fine-tuning

Significant improvement in sequence recovery performance

Iterative DPO application yields 79.5% TM-Score gain on challenging structures

Abstract

The inverse folding problem, aiming to design amino acid sequences that fold into desired three-dimensional structures, is pivotal for various biotechnological applications. Here, we introduce a novel approach leveraging Direct Preference Optimization (DPO) to fine-tune an inverse folding model using feedback from a protein folding model. Given a target protein structure, we begin by sampling candidate sequences from the inverse-folding model, then predict the three-dimensional structure of each sequence with the folding model to generate pairwise structural-preference labels. These labels are used to fine-tune the inverse-folding model under the DPO objective. Our results on the CATH 4.2 test set demonstrate that DPO fine-tuning not only improves sequence recovery of baseline models but also leads to a significant improvement in average TM-Score from 0.77 to 0.81, indicating enhanced structure similarity. Furthermore, iterative application of our DPO-based method on challenging protein structures yields substantial gains, with an average TM-Score increase of 79.5\% with regard to the baseline model. This work establishes a promising direction for enhancing protein sequence design ability from structure feedback by effectively utilizing preference optimization.