Molecule generation using transformers and policy gradient reinforcement learning
Generating novel valid molecules is often a difficult task, because the vast chemical space relies on the intuition of experienced chemists. In recent years, deep learning models have helped accelerate this process. These advanced models can also help identify suitable molecules for disease treatmen...
Published in: | Scientific Reports |
---|---|
Main Authors: | , , , |
Format: | Text |
Language: | English |
Published: |
Nature Publishing Group UK
2023
|
Subjects: | |
Online Access: | http://www.ncbi.nlm.nih.gov/pmc/articles/PMC10232454/ http://www.ncbi.nlm.nih.gov/pubmed/37258546 https://doi.org/10.1038/s41598-023-35648-w |
id |
ftpubmed:oai:pubmedcentral.nih.gov:10232454 |
---|---|
record_format |
openpolar |
spelling |
ftpubmed:oai:pubmedcentral.nih.gov:10232454 2023-06-18T03:43:17+02:00 Molecule generation using transformers and policy gradient reinforcement learning Mazuz, Eyal Shtar, Guy Shapira, Bracha Rokach, Lior 2023-05-31 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC10232454/ http://www.ncbi.nlm.nih.gov/pubmed/37258546 https://doi.org/10.1038/s41598-023-35648-w en eng Nature Publishing Group UK http://www.ncbi.nlm.nih.gov/pmc/articles/PMC10232454/ http://www.ncbi.nlm.nih.gov/pubmed/37258546 http://dx.doi.org/10.1038/s41598-023-35648-w © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) . Sci Rep Article Text 2023 ftpubmed https://doi.org/10.1038/s41598-023-35648-w 2023-06-04T01:33:17Z Generating novel valid molecules is often a difficult task, because the vast chemical space relies on the intuition of experienced chemists. In recent years, deep learning models have helped accelerate this process. These advanced models can also help identify suitable molecules for disease treatment. In this paper, we propose Taiga, a transformer-based architecture for the generation of molecules with desired properties. Using a two-stage approach, we first treat the problem as a language modeling task of predicting the next token, using SMILES strings. Then, we use reinforcement learning to optimize molecular properties such as QED. This approach allows our model to learn the underlying rules of chemistry and more easily optimize for molecules with desired properties. Our evaluation of Taiga, which was performed with multiple datasets and tasks, shows that Taiga is comparable to, or even outperforms, state-of-the-art baselines for molecule optimization, with improvements in the QED ranging from 2 to over 20 percent. The improvement was demonstrated both on datasets containing lead molecules and random molecules. We also show that with its two stages, Taiga is capable of generating molecules with higher biological property scores than the same model without reinforcement learning. Text taiga PubMed Central (PMC) Scientific Reports 13 1 |
institution |
Open Polar |
collection |
PubMed Central (PMC) |
op_collection_id |
ftpubmed |
language |
English |
topic |
Article |
spellingShingle |
Article Mazuz, Eyal Shtar, Guy Shapira, Bracha Rokach, Lior Molecule generation using transformers and policy gradient reinforcement learning |
topic_facet |
Article |
description |
Generating novel valid molecules is often a difficult task, because the vast chemical space relies on the intuition of experienced chemists. In recent years, deep learning models have helped accelerate this process. These advanced models can also help identify suitable molecules for disease treatment. In this paper, we propose Taiga, a transformer-based architecture for the generation of molecules with desired properties. Using a two-stage approach, we first treat the problem as a language modeling task of predicting the next token, using SMILES strings. Then, we use reinforcement learning to optimize molecular properties such as QED. This approach allows our model to learn the underlying rules of chemistry and more easily optimize for molecules with desired properties. Our evaluation of Taiga, which was performed with multiple datasets and tasks, shows that Taiga is comparable to, or even outperforms, state-of-the-art baselines for molecule optimization, with improvements in the QED ranging from 2 to over 20 percent. The improvement was demonstrated both on datasets containing lead molecules and random molecules. We also show that with its two stages, Taiga is capable of generating molecules with higher biological property scores than the same model without reinforcement learning. |
format |
Text |
author |
Mazuz, Eyal Shtar, Guy Shapira, Bracha Rokach, Lior |
author_facet |
Mazuz, Eyal Shtar, Guy Shapira, Bracha Rokach, Lior |
author_sort |
Mazuz, Eyal |
title |
Molecule generation using transformers and policy gradient reinforcement learning |
title_short |
Molecule generation using transformers and policy gradient reinforcement learning |
title_full |
Molecule generation using transformers and policy gradient reinforcement learning |
title_fullStr |
Molecule generation using transformers and policy gradient reinforcement learning |
title_full_unstemmed |
Molecule generation using transformers and policy gradient reinforcement learning |
title_sort |
molecule generation using transformers and policy gradient reinforcement learning |
publisher |
Nature Publishing Group UK |
publishDate |
2023 |
url |
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC10232454/ http://www.ncbi.nlm.nih.gov/pubmed/37258546 https://doi.org/10.1038/s41598-023-35648-w |
genre |
taiga |
genre_facet |
taiga |
op_source |
Sci Rep |
op_relation |
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC10232454/ http://www.ncbi.nlm.nih.gov/pubmed/37258546 http://dx.doi.org/10.1038/s41598-023-35648-w |
op_rights |
© The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) . |
op_doi |
https://doi.org/10.1038/s41598-023-35648-w |
container_title |
Scientific Reports |
container_volume |
13 |
container_issue |
1 |
_version_ |
1769009614737637376 |