SAR and QSAR modeling of a large collection of LD50 rat acute oral toxicity data

Abstract

The median lethal dose for rodent oral acute toxicity (LD50) is a standard piece of information required to categorize chemicals in terms of the potential hazard posed to human health after acute exposure. The exclusive use of in vivo testing is limited by the time and costs required for performing experiments and by the need to sacrifice a number of animals. (Quantitative) structure–activity relationships [(Q)SAR] proved a valid alternative to reduce and assist in vivo assays for assessing acute toxicological hazard. In the framework of a new international collaborative project, the NTP Interagency Center for the Evaluation of Alternative Toxicological Methods and the U.S. Environmental Protection Agency’s National Center for Computational Toxicology compiled a large database of rat acute oral LD50 data, with the aim of supporting the development of new computational models for predicting five regulatory relevant acute toxicity endpoints. In this article, a series of regression and classification computational models were developed by employing different statistical and knowledge-based methodologies. External validation was performed to demonstrate the real-life predictability of models. Integrated modeling was then applied to improve performance of single models. Statistical results confirmed the relevance of developed models in regulatory frameworks, and confirmed the effectiveness of integrated modeling. The best integrated strategies reached RMSEs lower than 0.50 and the best classification models reached balanced accuracies over 0.70 for multi-class and over 0.80 for binary endpoints. Computed predictions will be hosted on the EPA’s Chemistry Dashboard and made freely available to the scientific community.

Uncertainty-Informed Deep Transfer Learning of PFAS Toxicity

“…These results suggest that DNN with Mordred descriptors input outperforms other models with an R 2 of 0.65. While variations in datasets prevent direct oneto-one comparison to Table2, our DNN-Mordred model yields similar performance to that reported by Zhu et al50 and Gadaleta et al26 , justifying the evaluation of these models when further developed for the PFAS domain. Result of 5-fold cross-validation and mean test fold metrics.…”

Machine learning models for classification tasks related to drug safety

“…In our study, we have summarized the relevant classification models [85][86][87][88]. Different guidelines help in the categorization of the compounds in the different toxicity classes, such as the four-class system of the U.S. Environmental Protection Agency (EPA) [89] or the five-class version of the United Nations Globally Harmonized System of Classification and Labelling (GHS) [90].…”

“…SVM and neural network-based algorithms are also very common, and only a little amount of models contained algorithms other than the first five group, like logistic regression, LDA, self-organizing maps, SIMCA, etc. [72,86,114].…”

In silico prediction of toxicity and its applications for chemicals at work

“…Acute oral toxicity is the most widely studied in computational predictions, and several models were developed to predict acute oral toxicity. Several machine learning methods were developed and applied to construct classification and regression models to predict LD 50 or its toxicity categories [16].…”