Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network

Abstract As safety is one of the most important properties of drugs, chemical toxicology prediction has received increasing attentions in the drug discovery research. Traditionally, researchers rely on in vitro and in vivo experiments to test the toxicity of chemical compounds. However, not only are...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Jiarui Chen, Yain-Whar Si, Chon-Wai Un, Shirley W. I. Siu
Formato:	article
Lenguaje:	EN
Publicado:	BMC 2021
Materias:	Chemical toxicity Deep learning Graph convolutional neural network Semi-supervised learning Mean teacher Tox21 Information technology T58.5-58.64 Chemistry QD1-999
Acceso en línea:	https://doaj.org/article/9bc5fbb95f9047c3bbe067400ba1fe36
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:9bc5fbb95f9047c3bbe067400ba1fe36
record_format	dspace
spelling	oai:doaj.org-article:9bc5fbb95f9047c3bbe067400ba1fe362021-11-28T12:30:21ZChemical toxicity prediction based on semi-supervised learning and graph convolutional neural network10.1186/s13321-021-00570-81758-2946https://doaj.org/article/9bc5fbb95f9047c3bbe067400ba1fe362021-11-01T00:00:00Zhttps://doi.org/10.1186/s13321-021-00570-8https://doaj.org/toc/1758-2946Abstract As safety is one of the most important properties of drugs, chemical toxicology prediction has received increasing attentions in the drug discovery research. Traditionally, researchers rely on in vitro and in vivo experiments to test the toxicity of chemical compounds. However, not only are these experiments time consuming and costly, but experiments that involve animal testing are increasingly subject to ethical concerns. While traditional machine learning (ML) methods have been used in the field with some success, the limited availability of annotated toxicity data is the major hurdle for further improving model performance. Inspired by the success of semi-supervised learning (SSL) algorithms, we propose a Graph Convolution Neural Network (GCN) to predict chemical toxicity and trained the network by the Mean Teacher (MT) SSL algorithm. Using the Tox21 data, our optimal SSL-GCN models for predicting the twelve toxicological endpoints achieve an average ROC-AUC score of 0.757 in the test set, which is a 6% improvement over GCN models trained by supervised learning and conventional ML methods. Our SSL-GCN models also exhibit superior performance when compared to models constructed using the built-in DeepChem ML methods. This study demonstrates that SSL can increase the prediction power of models by learning from unannotated data. The optimal unannotated to annotated data ratio ranges between 1:1 and 4:1. This study demonstrates the success of SSL in chemical toxicity prediction; the same technique is expected to be beneficial to other chemical property prediction tasks by utilizing existing large chemical databases. Our optimal model SSL-GCN is hosted on an online server accessible through: https://app.cbbio.online/ssl-gcn/home .Jiarui ChenYain-Whar SiChon-Wai UnShirley W. I. SiuBMCarticleChemical toxicityDeep learningGraph convolutional neural networkSemi-supervised learningMean teacherTox21Information technologyT58.5-58.64ChemistryQD1-999ENJournal of Cheminformatics, Vol 13, Iss 1, Pp 1-16 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Chemical toxicity Deep learning Graph convolutional neural network Semi-supervised learning Mean teacher Tox21 Information technology T58.5-58.64 Chemistry QD1-999
spellingShingle	Chemical toxicity Deep learning Graph convolutional neural network Semi-supervised learning Mean teacher Tox21 Information technology T58.5-58.64 Chemistry QD1-999 Jiarui Chen Yain-Whar Si Chon-Wai Un Shirley W. I. Siu Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
description	Abstract As safety is one of the most important properties of drugs, chemical toxicology prediction has received increasing attentions in the drug discovery research. Traditionally, researchers rely on in vitro and in vivo experiments to test the toxicity of chemical compounds. However, not only are these experiments time consuming and costly, but experiments that involve animal testing are increasingly subject to ethical concerns. While traditional machine learning (ML) methods have been used in the field with some success, the limited availability of annotated toxicity data is the major hurdle for further improving model performance. Inspired by the success of semi-supervised learning (SSL) algorithms, we propose a Graph Convolution Neural Network (GCN) to predict chemical toxicity and trained the network by the Mean Teacher (MT) SSL algorithm. Using the Tox21 data, our optimal SSL-GCN models for predicting the twelve toxicological endpoints achieve an average ROC-AUC score of 0.757 in the test set, which is a 6% improvement over GCN models trained by supervised learning and conventional ML methods. Our SSL-GCN models also exhibit superior performance when compared to models constructed using the built-in DeepChem ML methods. This study demonstrates that SSL can increase the prediction power of models by learning from unannotated data. The optimal unannotated to annotated data ratio ranges between 1:1 and 4:1. This study demonstrates the success of SSL in chemical toxicity prediction; the same technique is expected to be beneficial to other chemical property prediction tasks by utilizing existing large chemical databases. Our optimal model SSL-GCN is hosted on an online server accessible through: https://app.cbbio.online/ssl-gcn/home .
format	article
author	Jiarui Chen Yain-Whar Si Chon-Wai Un Shirley W. I. Siu
author_facet	Jiarui Chen Yain-Whar Si Chon-Wai Un Shirley W. I. Siu
author_sort	Jiarui Chen
title	Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
title_short	Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
title_full	Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
title_fullStr	Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
title_full_unstemmed	Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
title_sort	chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
publisher	BMC
publishDate	2021
url	https://doaj.org/article/9bc5fbb95f9047c3bbe067400ba1fe36
work_keys_str_mv	AT jiaruichen chemicaltoxicitypredictionbasedonsemisupervisedlearningandgraphconvolutionalneuralnetwork AT yainwharsi chemicaltoxicitypredictionbasedonsemisupervisedlearningandgraphconvolutionalneuralnetwork AT chonwaiun chemicaltoxicitypredictionbasedonsemisupervisedlearningandgraphconvolutionalneuralnetwork AT shirleywisiu chemicaltoxicitypredictionbasedonsemisupervisedlearningandgraphconvolutionalneuralnetwork
_version_	1718407972241014784

Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network

Ejemplares similares