Reimagining Target-Aware Molecular Generation through Retrieval-Enhanced Aligned Diffusion

2025/06/17 by Dong Xu, Zhangfan Yang, Xu, Dong +11 · 1 voice · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · Medicine · #Biomolecules (q-bio.BM) #Chemical Synthesis and Analysis #FOS: Biological sciences #FOS: Computer and information sciences #Machine Learning (cs.LG) #Molecular Biology Techniques and Applications #Monoclonal and Polyclonal Antibodies Research #cs.LG #q-bio.BM

paper · pdf · doi:10.48550/arxiv.2506.14488

openalex publication_date 2025/06/17 · arxiv published 2025/06/17 · arxiv updated 2025/06/17 · openalex created_date 2025/10/18 · openalex updated_date 2026/07/28

Abstract

Breakthroughs in high-accuracy protein structure prediction, such as AlphaFold, have established receptor-based molecule design as a critical driver for rapid early-phase drug discovery. However, most approaches still struggle to balance pocket-specific geometric fit with strict valence and synthetic constraints. To resolve this trade-off, a Retrieval-Enhanced Aligned Diffusion termed READ is introduced, which is the first to merge molecular Retrieval-Augmented Generation with an SE(3)-equivariant diffusion model. Specifically, a contrastively pre-trained encoder aligns atom-level representations during training, then retrieves graph embeddings of pocket-matched scaffolds to guide each reverse-diffusion step at inference. This single mechanism can inject real-world chemical priors exactly where needed, producing valid, diverse, and shape-complementary ligands. Experimental results demonstrate that READ can achieve very competitive performance in CBGBench, surpassing state-of-the-art generative models and even native ligands. That suggests retrieval and diffusion can be co-optimized for faster, more reliable structure-based drug design.

Citations

• Equivariant Neural Diffusion for Molecule Generation
• UniMoMo: Unified Generative Modeling of 3D Molecules for De Novo Binder Design
• Molecule Generation with Fragment Retrieval Augmentation
• Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think
• Aligning Target-Aware Molecule Diffusion Models with Exact Energy Optimization
• CBGBench: Fill in the Blank of Protein-Molecule Complex Binding Graph
• Structure-based drug design by denoising voxel grids
• From Local to Global: A Graph RAG Approach to Query-Focused Summarization
• MolCRAFT: Structure-Based Drug Design in Continuous Parameter Space
• DecompDiff: Diffusion Models with Decomposed Priors for Structure-Based Drug Design
• Dynamic Retrieval Augmented Generation of Ontologies using Artificial Intelligence (DRAGON-AI)
• Functional-Group-Based Diffusion for Pocket-Specific Molecule Generation and Elaboration
• Learning Subpocket Prototypes for Generalizable Structure-based Drug Design
• Generation of 3D Molecules in Pockets via Language Model
• Geometric Latent Diffusion Models for 3D Molecule Generation
• 3D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Prediction
• DiffBP: Generative Diffusion of 3D Molecules for Target Protein Binding
• Structure-based Drug Design with Equivariant Diffusion Models
• Retrieval-based Controllable Molecule Generation
• GEM-2: Next Generation Molecular Property Prediction Network by Modeling Full-range Many-body Interactions
• Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pockets
• Generating 3D Molecules for Target Protein Binding
• Equivariant Diffusion for Molecule Generation in 3D
• A structural biology community assessment of AlphaFold2 applications
• Chemical-Reaction-Aware Molecule Representation Learning
• Highly accurate protein structure prediction with AlphaFold
• Score-based Generative Modeling in Latent Space
• Improved Denoising Diffusion Probabilistic Models
• Score-Based Generative Modeling through Stochastic Differential Equations
• Denoising Diffusion Implicit Models
• Denoising Diffusion Probabilistic Models
• Improved Techniques for Training Score-Based Generative Models
• AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading

Cited by

• Latent Retrieval Augmented Generation of Cross-Domain Protein Binders

Discussions

• Dong Xu, Zhangfan Yang, Ka-chun Wong, Zexuan Zhu, Jiangqiang Li, Junkai Ji: Reimagining Target-Aware Molecular Generation through Retrieval-Enhanced Aligned Diffusion https://arxiv.org/abs/2506.14488 [bsky, 0 points, 0 comments]

Related

• SiDGen: Structure-informed Diffusion for Generative modeling of Ligands for Proteins
• Structure-based Drug Design with Equivariant Diffusion Models
• PILOT: Equivariant diffusion for pocket conditioned de novo ligand generation with multi-objective guidance via importance sampling
• Binding-Adaptive Diffusion Models for Structure-Based Drug Design
• BInD: Bond and Interaction-generating Diffusion Model for Multi-objective Structure-based Drug Design