Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning
Deep reinforcement learning methods have been shown to be potentially powerful tools for de novo design. Recurrent-neural-network-based techniques are the most widely used methods in this space. In this work we examine the behaviour of recurrent-neural-network-based methods when there are few (or no...
Main Authors: | , , , , , |
---|---|
Format: | Journal article |
Language: | English |
Published: |
Springer Nature
2023
|
_version_ | 1797110301969088512 |
---|---|
author | Mokaya, M Imrie, F van Hoorn, WP Kalisz, A Bradley, AR Deane, CM |
author_facet | Mokaya, M Imrie, F van Hoorn, WP Kalisz, A Bradley, AR Deane, CM |
author_sort | Mokaya, M |
collection | OXFORD |
description | Deep reinforcement learning methods have been shown to be potentially powerful tools for de novo design. Recurrent-neural-network-based techniques are the most widely used methods in this space. In this work we examine the behaviour of recurrent-neural-network-based methods when there are few (or no) examples of molecules with the desired properties in the training data. We find that targeted molecular generation is usually possible, but the diversity of generated molecules is often reduced and it is not possible to control the composition of generated molecular sets. To help overcome these issues, we propose a new curriculum-learning-inspired recurrent iterative optimization procedure that enables the optimization of generated molecules for seen and unseen molecular profiles, and allows the user to control whether a molecular profile is explored or exploited. Using our method, we generate specific and diverse sets of molecules with up to 18 times more scaffolds than standard methods for the same sample size; however, our results also point to substantial limitations of one-dimensional molecular representations, as used in this space. We find that the success or failure of a given molecular optimization problem depends on the choice of simplified molecular-input line-entry system (SMILES). |
first_indexed | 2024-03-07T07:53:03Z |
format | Journal article |
id | oxford-uuid:6d93cdbb-f3d2-4bb8-b2b8-42467d876fc3 |
institution | University of Oxford |
language | English |
last_indexed | 2024-03-07T07:53:03Z |
publishDate | 2023 |
publisher | Springer Nature |
record_format | dspace |
spelling | oxford-uuid:6d93cdbb-f3d2-4bb8-b2b8-42467d876fc32023-07-27T09:33:36ZTesting the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learningJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:6d93cdbb-f3d2-4bb8-b2b8-42467d876fc3EnglishSymplectic ElementsSpringer Nature2023Mokaya, MImrie, Fvan Hoorn, WPKalisz, ABradley, ARDeane, CMDeep reinforcement learning methods have been shown to be potentially powerful tools for de novo design. Recurrent-neural-network-based techniques are the most widely used methods in this space. In this work we examine the behaviour of recurrent-neural-network-based methods when there are few (or no) examples of molecules with the desired properties in the training data. We find that targeted molecular generation is usually possible, but the diversity of generated molecules is often reduced and it is not possible to control the composition of generated molecular sets. To help overcome these issues, we propose a new curriculum-learning-inspired recurrent iterative optimization procedure that enables the optimization of generated molecules for seen and unseen molecular profiles, and allows the user to control whether a molecular profile is explored or exploited. Using our method, we generate specific and diverse sets of molecules with up to 18 times more scaffolds than standard methods for the same sample size; however, our results also point to substantial limitations of one-dimensional molecular representations, as used in this space. We find that the success or failure of a given molecular optimization problem depends on the choice of simplified molecular-input line-entry system (SMILES). |
spellingShingle | Mokaya, M Imrie, F van Hoorn, WP Kalisz, A Bradley, AR Deane, CM Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning |
title | Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning |
title_full | Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning |
title_fullStr | Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning |
title_full_unstemmed | Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning |
title_short | Testing the limits of SMILES-based de novo molecular generation with curriculum and deep reinforcement learning |
title_sort | testing the limits of smiles based de novo molecular generation with curriculum and deep reinforcement learning |
work_keys_str_mv | AT mokayam testingthelimitsofsmilesbaseddenovomoleculargenerationwithcurriculumanddeepreinforcementlearning AT imrief testingthelimitsofsmilesbaseddenovomoleculargenerationwithcurriculumanddeepreinforcementlearning AT vanhoornwp testingthelimitsofsmilesbaseddenovomoleculargenerationwithcurriculumanddeepreinforcementlearning AT kalisza testingthelimitsofsmilesbaseddenovomoleculargenerationwithcurriculumanddeepreinforcementlearning AT bradleyar testingthelimitsofsmilesbaseddenovomoleculargenerationwithcurriculumanddeepreinforcementlearning AT deanecm testingthelimitsofsmilesbaseddenovomoleculargenerationwithcurriculumanddeepreinforcementlearning |