Problem
Research questions and friction points this paper is trying to address.
Drug Design
Conditional Generation
Structure-based Design
Property Prediction
Molecular Optimization
One Sequence, Many Decodings: CAGenMol-2 Recasts Drug Design as Masked Molecular Inference
🤖 AI Summary
This study addresses the reliance on task-specific independent models in drug design by proposing CAGenMol-2, a unified masked diffusion model. By encoding molecules, properties, and protein pockets into a single sequence, the model unifies predictive and generative tasks through a masking strategy. Key innovations include the first single-checkpoint multi-task inference mechanism and a gradient-free adaptive fragment optimization algorithm (AdaFO). Integrated with sequence packing and iterative local search techniques, CAGenMol-2 achieves a 70.8% success rate on the CrossDocked2020 benchmark, substantially outperforming existing methods while preserving favorable drug-likeness.
📝 Abstract
Drug design couples property evaluation, conditional generation, structure-based design, and local optimization, yet machine learning systems typically address these capabilities with separate task-specific models. We introduce CAGenMol-2, a masked diffusion molecular language model that represents molecules, continuous scalar properties, and 3D protein pockets within a single wrapped sequence. Within this pretrained interface, downstream operations are selected by which sequence regions are observed or masked at inference, allowing one checkpoint to perform property prediction, property- and pocket-conditioned generation, and partial-constraint design without task-specific architectures or backbone fine-tuning. We further propose Adaptive Fragment Optimization (AdaFO), a gradient-free mask-and-refill search that turns the masked decoder into an iterative local molecular optimizer. On CrossDocked2020, AdaFO increases Success Rate from 30.2\% to 70.8\%, the best reported under this protocol, while largely preserving drug-likeness and diversity. Finally, scaffold-preserving directional editing and CRBN/VHL case studies demonstrate its use in compound design workflows spanning local molecular editing, structure-based prioritization, and downstream simulation-based screening.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Masked Diffusion Model
Unified Molecular Language Model
Adaptive Fragment Optimization
Structure-Based Drug Design
Gradient-Free Search
🔎 Similar Papers
No similar papers found.
