Transfer learning for predicting human skin sensitizers

Chun Wei Tung, Yi Hui Lin, Shan Shan Wang

研究成果: 雜誌貢獻文章

1 引文 (Scopus)

摘要

Computational prioritization of chemicals for potential skin sensitization risks plays essential roles in the risk assessment of environmental chemicals and drug development. Given the huge number of chemicals for testing, computational methods enable the fast identification of high-risk chemicals for experimental validation and design of safer alternatives. However, the development of robust prediction model requires a large dataset of tested chemicals that is usually not available for most toxicological endpoints, especially for human data. A small training dataset makes the development of effective models difficult with insufficient coverage and accuracy. In this study, an ensemble tree-based multitask learning method was developed incorporating three relevant tasks in the well-defined adverse outcome pathway (AOP) of skin sensitization to transfer shared knowledge to the major task of human sensitizers. The results show both largely improved coverage and accuracy compared with three state-of-the-art methods. A user-friendly prediction server was available at https://cwtung.kmu.edu.tw/skinsensdb/predict. As AOPs for various toxicity endpoints are being actively developed, the proposed method can be applied to develop prediction models for other endpoints.

原文英語
頁(從 - 到)931-940
頁數10
期刊Archives of Toxicology
93
發行號4
DOIs
出版狀態已發佈 - 四月 1 2019

指紋

Skin
Toxicology
Computational methods
Research Design
Risk assessment
Learning
Toxicity
Servers
Transfer (Psychology)
Pharmaceutical Preparations
Testing
Datasets

ASJC Scopus subject areas

  • Toxicology
  • Health, Toxicology and Mutagenesis

引用此文

Transfer learning for predicting human skin sensitizers. / Tung, Chun Wei; Lin, Yi Hui; Wang, Shan Shan.

於: Archives of Toxicology, 卷 93, 編號 4, 01.04.2019, p. 931-940.

研究成果: 雜誌貢獻文章

Tung, Chun Wei ; Lin, Yi Hui ; Wang, Shan Shan. / Transfer learning for predicting human skin sensitizers. 於: Archives of Toxicology. 2019 ; 卷 93, 編號 4. 頁 931-940.
@article{c9e34ce295fd44568ceae92f09e60826,
title = "Transfer learning for predicting human skin sensitizers",
abstract = "Computational prioritization of chemicals for potential skin sensitization risks plays essential roles in the risk assessment of environmental chemicals and drug development. Given the huge number of chemicals for testing, computational methods enable the fast identification of high-risk chemicals for experimental validation and design of safer alternatives. However, the development of robust prediction model requires a large dataset of tested chemicals that is usually not available for most toxicological endpoints, especially for human data. A small training dataset makes the development of effective models difficult with insufficient coverage and accuracy. In this study, an ensemble tree-based multitask learning method was developed incorporating three relevant tasks in the well-defined adverse outcome pathway (AOP) of skin sensitization to transfer shared knowledge to the major task of human sensitizers. The results show both largely improved coverage and accuracy compared with three state-of-the-art methods. A user-friendly prediction server was available at https://cwtung.kmu.edu.tw/skinsensdb/predict. As AOPs for various toxicity endpoints are being actively developed, the proposed method can be applied to develop prediction models for other endpoints.",
keywords = "Adverse outcome pathway, Allergic contact dermatitis, Alternative method, ExtraTrees, Multitask learning, Skin sensitization",
author = "Tung, {Chun Wei} and Lin, {Yi Hui} and Wang, {Shan Shan}",
year = "2019",
month = "4",
day = "1",
doi = "10.1007/s00204-019-02420-x",
language = "English",
volume = "93",
pages = "931--940",
journal = "Archiv fur Toxikologie",
issn = "0003-9446",
publisher = "Springer Verlag",
number = "4",

}

TY - JOUR

T1 - Transfer learning for predicting human skin sensitizers

AU - Tung, Chun Wei

AU - Lin, Yi Hui

AU - Wang, Shan Shan

PY - 2019/4/1

Y1 - 2019/4/1

N2 - Computational prioritization of chemicals for potential skin sensitization risks plays essential roles in the risk assessment of environmental chemicals and drug development. Given the huge number of chemicals for testing, computational methods enable the fast identification of high-risk chemicals for experimental validation and design of safer alternatives. However, the development of robust prediction model requires a large dataset of tested chemicals that is usually not available for most toxicological endpoints, especially for human data. A small training dataset makes the development of effective models difficult with insufficient coverage and accuracy. In this study, an ensemble tree-based multitask learning method was developed incorporating three relevant tasks in the well-defined adverse outcome pathway (AOP) of skin sensitization to transfer shared knowledge to the major task of human sensitizers. The results show both largely improved coverage and accuracy compared with three state-of-the-art methods. A user-friendly prediction server was available at https://cwtung.kmu.edu.tw/skinsensdb/predict. As AOPs for various toxicity endpoints are being actively developed, the proposed method can be applied to develop prediction models for other endpoints.

AB - Computational prioritization of chemicals for potential skin sensitization risks plays essential roles in the risk assessment of environmental chemicals and drug development. Given the huge number of chemicals for testing, computational methods enable the fast identification of high-risk chemicals for experimental validation and design of safer alternatives. However, the development of robust prediction model requires a large dataset of tested chemicals that is usually not available for most toxicological endpoints, especially for human data. A small training dataset makes the development of effective models difficult with insufficient coverage and accuracy. In this study, an ensemble tree-based multitask learning method was developed incorporating three relevant tasks in the well-defined adverse outcome pathway (AOP) of skin sensitization to transfer shared knowledge to the major task of human sensitizers. The results show both largely improved coverage and accuracy compared with three state-of-the-art methods. A user-friendly prediction server was available at https://cwtung.kmu.edu.tw/skinsensdb/predict. As AOPs for various toxicity endpoints are being actively developed, the proposed method can be applied to develop prediction models for other endpoints.

KW - Adverse outcome pathway

KW - Allergic contact dermatitis

KW - Alternative method

KW - ExtraTrees

KW - Multitask learning

KW - Skin sensitization

UR - http://www.scopus.com/inward/record.url?scp=85062626096&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062626096&partnerID=8YFLogxK

U2 - 10.1007/s00204-019-02420-x

DO - 10.1007/s00204-019-02420-x

M3 - Article

C2 - 30806762

AN - SCOPUS:85062626096

VL - 93

SP - 931

EP - 940

JO - Archiv fur Toxikologie

JF - Archiv fur Toxikologie

SN - 0003-9446

IS - 4

ER -