Don't Retrain, Just Reuse: Recovering Dual-Target Molecules from Single-Target Diffusion Models
2026-05-25 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors study how to design single molecules that can effectively interact with two biological targets at once, which is harder than targeting just one. Instead of changing the molecule-generating model or its process, they keep the original model fixed and search its input space to find molecules that work well for both targets. They develop a method called REUSE that uses a smart search combining exploration and selection to find balanced, high-quality molecules for both targets. Their experiments show REUSE outperforms previous methods in finding molecules with strong dual-target activity without sacrificing quality.
polypharmacologymolecular generative modelsdiffusion modelsdual-target drug designmulti-objective optimizationevolutionary algorithmsdrug-likenessmolecular affinityinput space searchchemical diversity
Authors
Qingyuan Zeng, Pengxiang Cai, Zixin Guan, Ziyang Chen, Anglin Liu, Lang Qin, Xinyao Lai, Jintai Chen
Abstract
Designing a single molecule that modulates two targets is a promising strategy for polypharmacology, but it remains substantially harder than standard single-target generation because one candidate must satisfy two binding requirements while preserving drug-likeness and synthesizability. Existing dual-target generative methods typically introduce dual-target capability by either retraining the generator or intervening in the diffusion process during sampling. The former can be costly and difficult to stabilize when dual-target supervision is sparse, while the latter may be sensitive to denoising-time target balancing and competing update directions. These limitations motivate a generator-preserving alternative that keeps the pretrained prior intact: can dual-target candidates instead be recovered from the input space of a frozen single-target diffusion model, without modifying its parameters or denoising dynamics? We formulate this task as a constrained multi-objective optimization problem and propose REUSE, a hierarchical evolutionary input-space search framework that combines pair-conditioned exploration with structured multi-stage selection to enforce dual-target affinity, chemical quality, and diversity. Experiments show that, compared with methods that modify the diffusion process, REUSE consistently improves dual-target affinity and balance, achieving a 20.9-percentage-point gain in Dual High Affinity over the strongest prior baseline while maintaining competitive molecular quality.