We use a supervised machine learning strategy to systematically investigate the relative importance of study type, machine learning algorithm, and type of descriptor on predicting in vivo repeat-dose toxicity at the organ-level. A total of 985 compounds were represented using chemical structural descriptors, ToxPrint chemotype descriptors, and bioactivity descriptors from ToxCast in vitro high-throughput screening assays. Using ToxRefDB, a total of 35 target organ outcomes were identified that contained at least 100 chemicals (50 positive and 50 negative). Supervised machine learning was performed using Naïve Bayes, k-nearest neighbor, random forest, classification and regression trees, and support vector classification approaches. Model performnce was assessed based on F1 scores using five-fold cross-validation with balanced bootstrap replicates. Fixed effects modeling showed the variance in F1 scores was explained mostly by target organ outcome, followed by descriptor type, machine learning algorithm, and interactions between these three factors. A combination of bioactivity and chemical structure or chemotype descriptors were the most predictive. Model performance improved with more chemicals (up to a maximum of 24%) and these gains were correlated (ρ= 0.92) with the number of chemicals.

This dataset is associated with the following publication: Liu, J., G. Patlewicz, A. Williams, R. Thomas, and I. Shah. (Chemical Research in Toxicology) Predicting organ toxicity using in vitro bioactivity data and chemical structure. CHEMICAL RESEARCH IN TOXICOLOGY. American Chemical Society, Washington, DC, USA, 30: 2046−2059, (2017).