Abstract:AI-assisted synthesis planning has emerged as a valuable tool in accelerating synthetic chemistry for the discovery of new drugs and materials. The template-free approach, which showcases superior generalization capabilities, is seen as the mainstream direction in this field. However, it remains unclear whether such an end-to-end approach can achieve problem-solving performance on par with experienced chemists without fully revealing insights into the chemical mechanisms involved. Moreover, there is a lack of unified and chemically inspired frameworks for improving multitask reaction predictions in this area. In this study, we have addressed these challenges by investigating the impact of fine-grained reaction-type labels on multiple downstream tasks and propose a novel framework named SynCluster. This framework incorporates unsupervised clustering cues into the baseline models and identifies plausible chemical subspaces which is compatible with multitask extensions and can serve as model-independent indicators to effectively enhance the performance of multiple downstream tasks. In retrosynthesis prediction, SynCluster achieves significant improvements of 4.1 and 11.0% in top-1 and top-10 prediction accuracy, respectively, compared to the baseline Molecular Transformer, and achieves a notable enhancement of 13.9% in top-10 accuracy when combined with Retroformer. By incorporating simplified molecular-input line-entry system augmentation, our framework achieves higher top-10 accuracy compared to state-of-the-art sequence-based retrosynthesis models and improves over the baseline on the diversity and validity of reactants. SynCluster also achieves 94.9% top-10 accuracy in forward synthesis prediction and 51.5% top-10 Maxfrag accuracy in reagent prediction. Overall, SynCluster provides a fresh perspective with chemical interpretability and reinforcement of domain knowledge in the synthesis design. It offers a promising solution for improving the accuracy and efficiency of AI-assisted synthesis planning and bridges the gap between template-free approaches and the problem-solving abilities of experienced chemists.

Improving Few- and Zero-Shot Reaction Template Prediction Using Modern Hopfield Networks

Modern Hopfield Networks for Few- and Zero-Shot Reaction Template Prediction

Retrosynthesis Prediction with Local Template Retrieval

Prediction of Organic Reaction Outcomes Using Machine Learning

Site-Specific Template Generative Approach for Retrosynthetic Planning

Deep Retrosynthetic Reaction Prediction using Local Reactivity and Global Attention

RetroPrime: A Diverse, Plausible and Transformer-based Method for Single-Step Retrosynthesis Predictions

Molecule-Edit Templates for Efficient and Accurate Retrosynthesis Prediction

RetroComposer: Composing Templates for Template-Based Retrosynthesis Prediction

SynCluster: Reaction Type Clustering and Recommendation Framework for Synthesis Planning

Hierarchical Framework for Retrosynthesis Prediction with Enhanced Reaction Center Localization

A generalized-template-based graph neural network for accurate organic reactivity prediction

Hard-threshold-Neural-Network based Prediction of Organic Synthetic Outcomes

A Graph to Graphs Framework for Retrosynthesis Prediction

Permutation invariant graph-to-sequence model for template-free retrosynthesis and reaction prediction

RetroPrime: A Chemistry-Inspired and Transformer-based Method for Retrosynthesis Predictions

Hard-threshold Neural Network-Based Prediction of Organic Synthetic Outcomes

Learning to Plan Chemical Syntheses

Dissecting Errors in Machine Learning for Retrosynthesis: A Granular Metric Framework and Transformer-Based Model for More Informative Predictions

Predicting Organic Reaction Outcomes with Weisfeiler-Lehman Network

Predicting retrosynthetic pathways using a combined linguistic model and hyper-graph exploration strategy