RetroTRAE: retrosynthetic translation of atomic environments with Transformer

16 August 2021, Version 2
This content is a preprint and has not undergone peer review at the time of posting.

Abstract

Herein we present a new retrosynthesis prediction method, viz. RetroTRAE, which uses fragment-based tokenization combined with the Transformer architecture. RetroTRAE mimics chemical reasoning, and predicts reactant candidates by learning the changes of atomic environments associated with the chemical reaction. Atom environments stand as ideal, chemically meaningful building blocks, which together produce a high-resolution molecular representation. Describing a molecule with a set of atom environments establishes a clear relationship between translated product-reactant pairs due to the conservation of atoms in the reactions. Our model achieved a top-1 accuracy of 68.1% within the bioactively similar range for the USPTO test dataset, outperforming other state-of-the-art translation methods. Besides yielding a high level of overall accuracy, the proposed method solves the translation issues arising from the SMILES-based retrosynthesis planning methods effectively. Through careful inspection of reactant candidates, we demonstrated atom environments as promising descriptors for studying reaction route prediction and discovery. RetroTRAE provides fast and reliable retrosynthetic route planning for substances whose fragmentation patterns are revealed. Our methodology offers a novel way of devising a retrosynthetic planning model using fragmental and topological descriptors as natural inputs for chemical translation tasks.

Keywords

Retrosynthesis
Transformer
Neural Machin Translation
Atom Environments

Comments

Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.