A Multiscale-Information-Embedded Universal Toxicity Prediction Framework DOI
Song He, Lianlian Wu, Fanmeng Wang

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 25, 2024

Abstract The inherent toxicity of compounds poses a significant threat to human health and the environment throughout their entire lifecycle, encompassing production, use, storage, disposal. Compound-induced evaluation is critical concern across various fields, including drug discovery environmental studies. Recent advances in deep learning hold promise for predicting compound-induced toxicities. However, existing works often exhibit poor predictive accuracy generalizability, especially rare toxicities with limited data. Most fail capture three-dimensional (3D) spatial arrangement stereochemical properties compounds, which are crucial understanding toxicological profiles. And interrelated nature has been overlooked. Here we propose ToxScan, novel SE(3)-equivariant multiscale model, as universal prediction framework address these issues. A two-level representation protocol, molecular- atomic-level information, introduced better incorporate geometry information 3D conformation. parallel modelling multi-task scheme applied learn characteristics multiple categories. Through comprehensive analysis scenarios such prediction, module effectiveness testing, generalization testing on new small-scale endpoints, distinguishing ability structurally similar opposing toxicities, vitro experimental validation predictions, practical application data, ToxScan demonstrates convincing capability surpassing state-of-the-art by remarkable 7.8–37.6% performance boost four metrics medium- endpoints. Typical case studies demonstrate that can identify subtle structural variations capturing molecular features at levels. To facilitate user accessibility, have integrated trained developed user-friendly online platform (https://app.bohrium.dp.tech/toxscan), allowing researchers obtain endpoint predictions quickly conveniently. We envision valuable tool evaluation, effectively reducing cycles costs while providing utility discovery, toxicology

Language: Английский

From molecular descriptors to the developmental toxicity prediction of pesticides/veterinary drugs/bio-pesticides against zebrafish embryo: Dual computational toxicological approaches for prioritization DOI
Yutong Wang, Peng Wang, Tengjiao Fan

et al.

Journal of Hazardous Materials, Journal Year: 2024, Volume and Issue: 476, P. 134945 - 134945

Published: June 17, 2024

Language: Английский

Citations

18
Quantitative prediction of toxicological points of departure using two-stage machine learning models: A new approach methodology (NAM) for chemical risk assessment DOI

Vaisali Chandrasekar,

Saad Mohammad,

Omar M. Aboumarzouk

et al.

Journal of Hazardous Materials, Journal Year: 2025, Volume and Issue: 487, P. 137071 - 137071

Published: Jan. 10, 2025

Language: Английский

Citations

10
Occurrence and risks of pharmaceuticals, personal care products, and endocrine-disrupting compounds in Chinese surface waters DOI
Yuhang Chen, Mengyuan Li,

Weichun Gao

et al.

Journal of Environmental Sciences, Journal Year: 2023, Volume and Issue: 146, P. 251 - 263

Published: Oct. 28, 2023

Language: Английский

Citations

18
ToxMPNN: A deep learning model for small molecule toxicity prediction DOI

Yini Zhou,

Chao Ning,

Yijun Tan

et al.

Journal of Applied Toxicology, Journal Year: 2024, Volume and Issue: 44(7), P. 953 - 964

Published: Feb. 26, 2024

Abstract Machine learning (ML) has shown a great promise in predicting toxicity of small molecules. However, the availability data for such predictions is often limited. Because unsatisfactory performance models trained on single endpoint, we collected toxic molecules with multiple endpoints from previous study. The dataset comprises 27 categorized into seven classes, namely, carcinogenicity and mutagenicity, acute oral toxicity, respiratory irritation corrosion, cardiotoxicity, CYP450, endocrine disruption. In addition, binary classification Common‐Toxicity task was added based aforementioned dataset. To improve models, marketed drugs as negative samples. This study presents predictive model, ToxMPNN, message passing neural network (MPNN) architecture, aiming to predict results demonstrate that ToxMPNN outperforms other capturing features within molecular structure, resulting more precise ROC_AUC testing score 0.886 Toxicity_drug Furthermore, it observed adding samples not only improves but also enhances stability model prediction. It shows graph‐based deep (DL) algorithms this can be used trustworthy effective tool assess molecule development new drugs.

Language: Английский

Citations

7
Machine Learning‐Enabled Drug‐Induced Toxicity Prediction DOI Creative Commons
Changsen Bai, Lianlian Wu, Ruijiang Li

et al.

Advanced Science, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 3, 2025

Abstract Unexpected toxicity has become a significant obstacle to drug candidate development, accounting for 30% of discovery failures. Traditional assessment through animal testing is costly and time‐consuming. Big data artificial intelligence (AI), especially machine learning (ML), are robustly contributing innovation progress in toxicology research. However, the optimal AI model different types usually varies, making it essential conduct comparative analyses methods across domains. The diverse sources also pose challenges researchers focusing on specific studies. In this review, 10 categories drug‐induced examined, summarizing characteristics applicable ML models, including both predictive interpretable algorithms, striking balance between breadth depth. Key databases tools used prediction highlighted, toxicology, chemical, multi‐omics, benchmark databases, organized by their focus function clarify roles prediction. Finally, strategies turn into opportunities analyzed discussed. This review may provide with valuable reference understanding utilizing available resources bridge mechanistic insights, further advance application drugs‐induced

Language: Английский

Citations

1
Implementing comprehensive machine learning models of multispecies toxicity assessment to improve regulation of organic compounds DOI
Ying He, Guohong Liu, Song Hu

et al.

Journal of Hazardous Materials, Journal Year: 2023, Volume and Issue: 458, P. 131942 - 131942

Published: June 27, 2023

Language: Английский

Citations

15
A deep learning based multi-model approach for predicting drug-like chemical compound’s toxicity DOI
Konda Mani Saravanan,

Jiang-Fan Wan,

Liujiang Dai

et al.

Methods, Journal Year: 2024, Volume and Issue: 226, P. 164 - 175

Published: May 1, 2024

Language: Английский

Citations

5
Fabrication of D-α-tocopheryl polyethylene glycol 1000 succinates and human serum albumin conjugated chitosan nanoparticles of bosutinib for colon targeting application; in vitro-in vivo investigation DOI
Laxmi Manthalkar, Sankha Bhattacharya, Ketan Hatware

et al.

International Journal of Biological Macromolecules, Journal Year: 2023, Volume and Issue: 253, P. 127531 - 127531

Published: Oct. 18, 2023

Language: Английский

Citations

9
Multi-task multi-view and iterative error-correcting random forest for acute toxicity prediction DOI
Jie Gao, Lianlian Wu,

Guangyi Lin

et al.

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 126972 - 126972

Published: Feb. 1, 2025

Language: Английский

Citations

0
Prediction of acute toxicity of organic contaminants to fish: model development and a novel approach to identify reactive substructures DOI

Shangyu Li,

Mingming Zhang, Peizhe Sun

et al.

Journal of Hazardous Materials, Journal Year: 2025, Volume and Issue: 491, P. 137917 - 137917

Published: March 13, 2025

Language: Английский

Citations

0