TAGMol: Target-Aware Gradient-guided Molecule Generation

TL;DR

TAGMol presents a gradient-guided diffusion framework for structure-based drug design that decouples molecular generation from property prediction to enable target-aware multi-objective optimization. It introduces an SE(3) invariant GNN-based property guide trained on noisy 3D complexes and uses gradient-guided sampling to steer diffusion toward binding affinity and pharmacological properties such as QED and SA. Empirically, TAGMol achieves a 22% improvement in AutoDock Vina scores and higher hit rates while preserving molecular geometry and diversity relative to baselines like TargetDiff and DecompDiff. This approach reduces reliance on post-hoc optimization and provides a principled pathway for designing ligands that are both potent and drug-like in 3D space.

Abstract

3D generative models have shown significant promise in structure-based drug design (SBDD), particularly in discovering ligands tailored to specific target binding sites. Existing algorithms often focus primarily on ligand-target binding, characterized by binding affinity. Moreover, models trained solely on target-ligand distribution may fall short in addressing the broader objectives of drug discovery, such as the development of novel ligands with desired properties like drug-likeness, and synthesizability, underscoring the multifaceted nature of the drug design process. To overcome these challenges, we decouple the problem into molecular generation and property prediction. The latter synergistically guides the diffusion sampling process, facilitating guided diffusion and resulting in the creation of meaningful molecules with the desired properties. We call this guided molecular generation process as TAGMol. Through experiments on benchmark datasets, TAGMol demonstrates superior performance compared to state-of-the-art baselines, achieving a 22% improvement in average Vina Score and yielding favorable outcomes in essential auxiliary properties. This establishes TAGMol as a comprehensive framework for drug generation.

Figures (5)

• Figure 1: Overview of TAGMol. (a) Training a property-oriented guide using existing data. (b) Utilizing the trained guide and diffusion model to steer the diffusion sampling process towards the optimal regions of the property of interest.
• Figure 2: Comparison of distance distributions between all-atom distances of reference molecules in the test set (Reference) and distances in model-generated molecules. The Jensen-Shannon divergence (JSD) between these two distributions is calculated and reported.
• Figure 3: Visualization of reference molecules (left), alongside molecules generated by our backbone, TargetDiff (middle), and TAGMol (right), for two targets: 5D7N and 5MGL.
• Figure 4: Distribution of QED, SA, and Vina Score properties in the training split, with the mean values represented by dotted lines
• Figure 5: Distribution of molecular properties in molecules generated by the backbone model without any guidance (No Opt), when guided by two properties while excluding the one indicated on the x-axis (Excluded Property on x-axis), and when guided by all three properties (All Properties). ($\downarrow$) denote properties where lower values are preferred, while ($\uparrow$) indicate properties where higher values are desirable.