Unified Guidance for Geometry-Conditioned Molecular Generation

2025-01-05

AI Summary

Existing molecular generation models are highly task-specific and exhibit weak geometric control, limiting their applicability across diverse drug design scenarios—including scaffold-based design, fragment linking, and ligand conformational sampling. Method: We propose UniGuide, the first plug-and-play unified geometric guidance framework for unconditional diffusion models, enabling flexible, task-agnostic geometric constraints—such as binding pocket shape, fragment connection points, or ligand conformation—without model retraining. Its core innovation lies in constructing a task-invariant guidance vector, integrated via latent-space conditional projection and geometry-aware guidance throughout the diffusion process. Contribution/Results: On multiple benchmarks, UniGuide matches or surpasses state-of-the-art task-specific models, significantly improving the 3D geometric validity and target-binding compatibility of generated molecules while preserving model generality and deployment simplicity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Effectively designing molecular geometries is essential to advancing pharmaceutical innovations, a domain, which has experienced great attention through the success of generative models and, in particular, diffusion models. However, current molecular diffusion models are tailored towards a specific downstream task and lack adaptability. We introduce UniGuide, a framework for controlled geometric guidance of unconditional diffusion models that allows flexible conditioning during inference without the requirement of extra training or networks. We show how applications such as structure-based, fragment-based, and ligand-based drug design are formulated in the UniGuide framework and demonstrate on-par or superior performance compared to specialised models. Offering a more versatile approach, UniGuide has the potential to streamline the development of molecular generative models, allowing them to be readily used in diverse application scenarios.

Problem

• Molecular Design
• Flexibility Limitation
• Drug Design

Innovation

• UniGuide
• Molecular Design
• Adaptive Modeling