Journal of Chemical Information and Modeling · 2025 · 10 citations · 41 references
PharmacyMedicinal ChemistryComputational ChemistryDrug DesignDrug DiscoveryChemical Space ExplorationCombinatorial SynthesisComputational Drug DiscoveryDe Novo Drug DesignHigh-throughput SynthesisRetrosynthetic AnalysisAccuracyChemical Synthesis PlanningComputational Synthesis PlanningComputer-assisted Synthesis PlanningMultistep SynthesisPredictionTransformer ModelsTransformer-based Prediction Models
Traditional computer-aided synthesis planning (CASP) methods rely on iterative single-step predictions, leading to exponential search space growth that limits efficiency and scalability. We introduce a series of transformer-based models that leverage a mixture of experts approach to directly generate multistep synthetic routes as a single string, conditionally predicting each transformation based on all preceding ones. Our DMS Explorer XL model, which requires only target compounds as input, outperforms state-of-the-art methods on the PaRoutes dataset with 1.9x and 3.1x improvements in Top-1 accuracy on the n 1 and n 5 test sets, respectively. Providing additional information, such as the desired number of steps and starting materials, enables both a reduction in model size and an increase in accuracy, highlighting the benefits of incorporating more constraints into the prediction process. The top-performing DMS-Flex (Duo) model scores 25–50% higher on Top-1 and Top-10 accuracies for both n 1 and n 5 sets. Additionally, our models successfully predict routes for the FDA-approved drugs not included in the training data, demonstrating strong generalization capabilities. While the limited diversity of the training set may affect performance on less common reaction types, our multistep-first approach presents a promising direction toward fully automated retrosynthetic planning.
41
Decoupled Weight Decay Regularization
Ilya Loshchilov, Frank Hutter · arXiv (Cornell University) · 2017 · 9.1K citations · Full text
Gaussian Error Linear Units (GELUs)
Dan Hendrycks, Kevin Gimpel · arXiv (Cornell University) · 2016 · 3.2K citations · Full text
Planning chemical syntheses with deep neural networks and symbolic AI
Marwin Segler, Mark P. Waller · Nature · 2018 · 1.9K citations · Full text
Computer-Assisted Design of Complex Organic Syntheses
E. J. Corey, W. Todd Wipke · Science · 1969 · 718 citations