RetroComposer: Composing Templates for Template-Based Retrosynthesis Prediction

Abstract

The main target of retrosynthesis is to recursively decompose desired molecules into available building blocks. Existing template-based retrosynthesis methods follow a template selection stereotype and suffer from limited training templates, which prevents them from discovering novel reactions. To overcome this limitation, we propose an innovative retrosynthesis prediction framework that can compose novel templates beyond training templates. As far as we know, this is the first method that uses machine learning to compose reaction templates for retrosynthesis prediction. Besides, we propose an effective reactant candidate scoring model that can capture atom-level transformations, which helps our method outperform previous methods on the USPTO-50K dataset. Experimental results show that our method can produce novel templates for 15 USPTO-50K test reactions that are not covered by training templates. We have released our source implementation.

15 pages; Accepted by the journal of Biomolecules

References in corpus (5)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• State-of-the-Art Augmented NLP Transformer models for direct and single-step retrosynthesis
• Predicting Organic Reaction Outcomes with Weisfeiler-Lehman Network
• Learning Graph Models for Retrosynthesis Prediction
• Retrosynthesis Prediction with Conditional Graph Logic Network

Cited by in corpus (2)

• Re-evaluating Retrosynthesis Algorithms with Syntheseus
• Node-Aligned Graph-to-Graph (NAG2G): Elevating Template-Free Deep Learning Approaches in Single-Step Retrosynthesis