Natural products remain one of the most productive sources of chemical inspiration for the development of new drugs. The structures of more than 250 000 natural products are available from public databases. At least 10% of these compounds are readily obtainable for experimental testing from commercial vendors and public research institutions. While the physicochemical properties of known natural products have been thoroughly studied and compared to those of drugs and other types of small molecules, the information available on the content, coverage, and relevance of individual virtual and physical natural product libraries is clearly limited. The aim of this study was the development of a detailed understanding of the coverage of chemical space by known and readily obtainable natural products and by individual natural product databases. For this purpose, we compiled comprehensive data sets of known and readily obtainable natural products from 18 virtual databases (including the Dictionary of Natural Products), nine physical libraries, and the Protein Data Bank (PDB). We also developed and employed an algorithm ("SugarBuster") for the removal of sugars and sugar-like moieties, which are generally not in the focus of interest for drug discovery, from natural products. In addition, we devised a rule-based approach for the automated classification of natural products into natural product classes (alkaloids, steroids, flavonoids, etc.). Among the most important results of this study is the finding that the readily obtainable natural products are highly diverse and populate regions of chemical space that are of high relevance to drug discovery. In some cases, substantial differences in the coverage of natural product classes and chemical space by the individual databases are observed. More than 2000 natural products are identified for which at least one X-ray crystal structure of the compound in complex with a biomacromolecule is available from the PDB.

Computational methods such as machine learning approaches have a strong track record of success in predicting the outcomes of in vitro assays. In contrast, their ability to predict in vivo endpoints is more limited due to the high number of parameters and processes that may influence the outcome. Recent studies have shown that the combination of chemical and biological data can yield better models for in vivo endpoints. The ChemBioSim approach presented in this work aims to enhance the performance of conformal prediction models for in vivo endpoints by combining chemical information with (predicted) bioactivity assay outcomes. Three in vivo toxicological endpoints, capturing genotoxic (MNT), hepatic (DILI), and cardiological (DICC) issues, were selected for this study due to their high relevance for the registration and authorization of new compounds. Since the sparsity of available biological assay data is challenging for predictive modeling, predicted bioactivity descriptors were introduced instead. Thus, a machine learning model for each of the 373 collected biological assays was trained and applied on the compounds of the in vivo toxicity data sets. Besides the chemical descriptors (molecular fingerprints and physicochemical properties), these predicted bioactivities served as descriptors for the models of the three in vivo endpoints. For this study, a workflow based on a conformal prediction framework (a method for confidence estimation) built on random forest models was developed. Furthermore, the most relevant chemical and bioactivity descriptors for each in vivo endpoint were preselected with lasso models. The incorporation of bioactivity descriptors increased the mean F1 scores of the MNT model from 0.61 to 0.70 and for the DICC model from 0.72 to 0.82 while the mean efficiencies increased by roughly 0.10 for both endpoints. In contrast, for the DILI endpoint, no significant improvement in model performance was observed. Besides pure performance improvements, an analysis of the most important bioactivity features allowed detection of novel and less intuitive relationships between the predicted biological assay outcomes used as descriptors and the in vivo endpoints. This study presents how the prediction of in vivo toxicity endpoints can be improved by the incorporation of biological information—which is not necessarily captured by chemical descriptors—in an automated workflow without the need for adding experimental workload for the generation of bioactivity descriptors as predicted outcomes of bioactivity assays were utilized. All bioactivity CP models for deriving the predicted bioactivities, as well as the in vivo toxicity CP models, can be freely downloaded from .

Disturbance of the thyroid hormone homeostasis has been associated with adverse health effects such as goiters and impaired mental development in humans and thyroid tumors in rats. In vitro and in silico methods for predicting the effects of small molecules on thyroid hormone homeostasis are currently being explored as alternatives to animal experiments, but are still in an early stage of development. The aim of this work was the development of a battery of in silico models for a set of targets involved in molecular initiating events of thyroid hormone homeostasis: deiodinases 1, 2, and 3, thyroid peroxidase (TPO), thyroid hormone receptor (TR), sodium/iodide symporter, thyrotropin-releasing hormone receptor, and thyroid-stimulating hormone receptor. The training data sets were compiled from the ToxCast database and related scientific literature. Classical statistical approaches as well as several machine learning methods (including random forest, support vector machine, and neural networks) were explored in combination with three data balancing techniques. The models were trained on molecular descriptors and fingerprints and evaluated on holdout data. Furthermore, multi-task neural networks combining several end points were investigated as a possible way to improve the performance of models for which the experimental data available for model training are limited. Classifiers for TPO and TR performed particularly well, with F1 scores of 0.83 and 0.81 on the holdout data set, respectively. Models for the other studied targets yielded F1 scores of up to 0.77. An in-depth analysis of the reliability of predictions was performed for the most relevant models. All data sets used in this work for model development and validation are available in the Supporting Information.