MolChord: Structure-Sequence Alignment for Protein-Guided Drug Design

TL;DR

MolChord introduces a novel structure-sequence alignment method for protein-guided drug design, integrating text, structural, and property data with advanced models to improve alignment accuracy and drug property targeting.

Contribution

The paper presents MolChord, a new framework combining text-based and structural representations with property-aware optimization for improved structure-based drug design.

Findings

Achieves state-of-the-art performance on CrossDocked2020

Effectively aligns protein and molecule structures with textual descriptions

Guides molecules toward desired pharmacological properties

Abstract

Structure-based drug design (SBDD), which maps target proteins to candidate molecular ligands, is a fundamental task in drug discovery. Effectively aligning protein structural representations with molecular representations, and ensuring alignment between generated drugs and their pharmacological properties, remains a critical challenge. To address these challenges, we propose MolChord, which integrates two key techniques: (1) to align protein and molecule structures with their textual descriptions and sequential representations (e.g., FASTA for proteins and SMILES for molecules), we leverage NatureLM, an autoregressive model unifying text, small molecules, and proteins, as the molecule generator, alongside a diffusion-based structure encoder; and (2) to guide molecules toward desired properties, we curate a property-aware dataset by integrating preference data and refine the alignment process using Direct Preference Optimization (DPO). Experimental results on CrossDocked2020 demonstrate that our approach achieves state-of-the-art performance on key evaluation metrics, highlighting its potential as a practical tool for SBDD.