Molecule-edit templates for efficient and accurate retrosynthesis prediction

Mikołaj Sacha, Michał Sadowski, Piotr Kozakowski, Ruard van Workum, Stanislaw Kamil Jastrzebski

Keywords: retrosynthesis, reaction prediction, chemistry, life sciences, graph neural networks, reaction templates, generative models

TL;DR: METRO is a novel machine learning model that generates chemical reactions represented as minimal templates. This leads to fewer reaction templates needed and state-of-the art accuracy on benchmark datasets.

Abstract: Retrosynthesis involves determining a sequence of reactions to synthesize complex molecules from simpler precursors. As this poses a challenge in organic chemistry, machine learning has offered solutions, particularly for predicting possible reaction substrates for a given target molecule. These solutions mainly fall into template-based and template-free categories. The former is efficient but relies on a vast set of predefined reaction patterns, while the latter, though more flexible, can be computationally intensive and less interpretable. To address these issues, we introduce METRO (Molecule-Edit Templates for RetrOsynthesis), a machine-learning model that predicts reactions using minimal templates - simplified reaction patterns capturing only essential molecular changes - reducing computational overhead and achieving state-of-the-art results on standard benchmarks.

Submission Track: Original Research

Submission Number: 12