Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features

Cytotoxicity, usually represented by cell viability, is a crucial parameter for evaluating drug safety in vitro. Accurate prediction of cell viability/cytotoxicity could accelerate drug development in the early stage. In this study, by integrating cellular transcriptome and cell viability data using...

Full description

Saved in:
Bibliographic Details
Main Authors: You Wu, Ke Tang, Chunzheng Wang, Hao Song, Fanfan Zhou, Ying Guo
Format: Article
Language:English
Published: Elsevier 2025-03-01
Series:Acta Pharmaceutica Sinica B
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2211383525000528
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849733150784094208
author You Wu
Ke Tang
Chunzheng Wang
Hao Song
Fanfan Zhou
Ying Guo
author_facet You Wu
Ke Tang
Chunzheng Wang
Hao Song
Fanfan Zhou
Ying Guo
author_sort You Wu
collection DOAJ
description Cytotoxicity, usually represented by cell viability, is a crucial parameter for evaluating drug safety in vitro. Accurate prediction of cell viability/cytotoxicity could accelerate drug development in the early stage. In this study, by integrating cellular transcriptome and cell viability data using four machine learning algorithms (support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost), and light gradient boosting machine (LightGBM)) and two ensemble algorithms (voting and stacking), highly accurate prediction models of 50% and 80% cell viability were developed with area under the receiver operating characteristic curve (AUROC) of 0.90 and 0.84, respectively; these models also showed good performance when utilized for diverse cell lines. Concerning the characterization of the employed Feature Genes, the models were interpreted, and the mechanisms of bioactive compounds with a narrow therapeutic index (NTI) can also be analyzed. In summary, the models established in this research exhibit superior capacity to those of previous studies; these models enable accurate high-safety substance screening via cytotoxicity prediction across cell lines. Moreover, for the first time, Cytotoxicity Signature (CTS) genes were identified, which could provide additional clues for further study of mechanisms of action (MOA), especially for NTI compounds.
format Article
id doaj-art-ee3ec495497243d79ad943fe51c23b62
institution DOAJ
issn 2211-3835
language English
publishDate 2025-03-01
publisher Elsevier
record_format Article
series Acta Pharmaceutica Sinica B
spelling doaj-art-ee3ec495497243d79ad943fe51c23b622025-08-20T03:08:06ZengElsevierActa Pharmaceutica Sinica B2211-38352025-03-011531344135810.1016/j.apsb.2025.02.009Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome featuresYou Wu0Ke Tang1Chunzheng Wang2Hao Song3Fanfan Zhou4Ying Guo5Beijing Key Laboratory of New Drug Mechanisms and Pharmacological Evaluation Study, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Department of Pharmacology, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, ChinaBeijing Key Laboratory of New Drug Mechanisms and Pharmacological Evaluation Study, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Department of Pharmacology, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, ChinaBeijing Key Laboratory of New Drug Mechanisms and Pharmacological Evaluation Study, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Department of Pharmacology, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, ChinaBeijing Key Laboratory of New Drug Mechanisms and Pharmacological Evaluation Study, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Department of Pharmacology, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, ChinaBeijing Key Laboratory of New Drug Mechanisms and Pharmacological Evaluation Study, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Department of Pharmacology, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, ChinaBeijing Key Laboratory of New Drug Mechanisms and Pharmacological Evaluation Study, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Department of Pharmacology, Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100050, China; Corresponding author.Cytotoxicity, usually represented by cell viability, is a crucial parameter for evaluating drug safety in vitro. Accurate prediction of cell viability/cytotoxicity could accelerate drug development in the early stage. In this study, by integrating cellular transcriptome and cell viability data using four machine learning algorithms (support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost), and light gradient boosting machine (LightGBM)) and two ensemble algorithms (voting and stacking), highly accurate prediction models of 50% and 80% cell viability were developed with area under the receiver operating characteristic curve (AUROC) of 0.90 and 0.84, respectively; these models also showed good performance when utilized for diverse cell lines. Concerning the characterization of the employed Feature Genes, the models were interpreted, and the mechanisms of bioactive compounds with a narrow therapeutic index (NTI) can also be analyzed. In summary, the models established in this research exhibit superior capacity to those of previous studies; these models enable accurate high-safety substance screening via cytotoxicity prediction across cell lines. Moreover, for the first time, Cytotoxicity Signature (CTS) genes were identified, which could provide additional clues for further study of mechanisms of action (MOA), especially for NTI compounds.http://www.sciencedirect.com/science/article/pii/S2211383525000528Interpretable modelDrug safetyCell viabilityWeak cytotoxicityMachine learningTranscriptome
spellingShingle You Wu
Ke Tang
Chunzheng Wang
Hao Song
Fanfan Zhou
Ying Guo
Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
Acta Pharmaceutica Sinica B
Interpretable model
Drug safety
Cell viability
Weak cytotoxicity
Machine learning
Transcriptome
title Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
title_full Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
title_fullStr Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
title_full_unstemmed Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
title_short Establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
title_sort establishment of interpretable cytotoxicity prediction models using machine learning analysis of transcriptome features
topic Interpretable model
Drug safety
Cell viability
Weak cytotoxicity
Machine learning
Transcriptome
url http://www.sciencedirect.com/science/article/pii/S2211383525000528
work_keys_str_mv AT youwu establishmentofinterpretablecytotoxicitypredictionmodelsusingmachinelearninganalysisoftranscriptomefeatures
AT ketang establishmentofinterpretablecytotoxicitypredictionmodelsusingmachinelearninganalysisoftranscriptomefeatures
AT chunzhengwang establishmentofinterpretablecytotoxicitypredictionmodelsusingmachinelearninganalysisoftranscriptomefeatures
AT haosong establishmentofinterpretablecytotoxicitypredictionmodelsusingmachinelearninganalysisoftranscriptomefeatures
AT fanfanzhou establishmentofinterpretablecytotoxicitypredictionmodelsusingmachinelearninganalysisoftranscriptomefeatures
AT yingguo establishmentofinterpretablecytotoxicitypredictionmodelsusingmachinelearninganalysisoftranscriptomefeatures