A novel procedure for selection of molecular descriptors: QSAR model for mutagenicity of nitroaromatic compounds

Nitroaromatic compounds (NACs) stand out as pervasive organic pollutants, prompting an imperative need to investigate their hazardous effects. Computational chemistry methods play a crucial role in this exploration, offering a safer and more time-efficient approach, mandated by various legislations. In this study, our focus lay on the development of transparent, interpretable, reproducible, and publicly available methodologies aimed at deriving quantitative structure-activity relationship models and testing them by modelling the mutagenicity of NACs against the Salmonella typhimurium TA100 strain. Descriptors were selected from Mordred and RDKit molecular descriptors, along with several quantum chemistry descriptors. For that purpose, the genetic algorithm (GA), as the most widely used method in the literature, and three alternative algorithms (Boruta, Featurewiz, and ForwardSelector) combined with the forward stepwise selection technique were used. The construction of models utilized the multiple linear regression method, with subsequent scrutiny of fitting and predictive performance, reliability, and robustness through various statistical validation criteria. The models were ranked by the Multi-Criteria Decision Making procedure. Findings have revealed that the proposed methodology for descriptor selection outperforms GA, with Featurewiz showing a slight advantage over Boruta and ForwardSelector. These constructed models can serve as valuable tools for the quick and reliable prediction of NACs mutagenicity.

Accurate Prediction of Rat Acute Oral Toxicity and Reference Dose for Thousands of Polycyclic Aromatic Hydrocarbon Derivatives Based on Chemometric QSAR and Machine Learning

Acute oral toxicity is currently not available for most polycyclic aromatic hydrocarbons (PAHs), especially their derivatives, because it is cost-prohibitive to experimentally determine all of them. Here, quantitative structure-activity relationship (QSAR) models using machine learning (ML) for predicting the toxicity of PAH derivatives were developed, based on oral toxicity data points of 788 individual substances of rats. Both the individual ML algorithm gradient boosting regression trees (GBRT) and the stacking ML algorithm (extreme gradient boosting + GBRT + random forest regression) provided the best prediction results with satisfactory determination coefficients for both cross-validation and the test set. It was found that those PAH derivatives with fewer polar hydrogens, more large-sized atoms, more branches, and lower polarizability have higher toxicity. Software based on the optimal ML-QSAR model was successfully developed to expand the application potential of the developed model, obtaining reliable prediction of pLD50 values and reference doses for 6893 external PAH derivatives. Among these chemicals, 472 were identified as moderately or highly toxic; 10 out of them had clear environment detection or use records. The findings provide valuable insights into the toxicity of PAHs and their derivatives, offering a standard platform for effectively evaluating chemical toxicity using ML-QSAR models.

Uncovering the toxicity mechanisms of a series of carboxylic acids in liver cells through computational and experimental approaches

Hepatotoxicity poses a significant concern in drug design due to the potential liver damage that can be caused by new drugs. Among common manifestations of hepatotoxic damage is lipid accumulation in hepatic tissue, resulting in liver steatosis or phospholipidosis. Carboxylic derivatives are prone to interfere with fatty acid metabolism and cause lipid accumulation in hepatocytes. This study investigates the toxic behaviour of 24 structurally related carboxylic acids in hepatocytes, specifically their ability to cause accumulation of fatty acids and phospholipids. Using high-content screening (HCS) assays, we identified two distinct lipid accumulation patterns. Subsequently, we developed structure-activity relationship (SAR) and quantitative structure-activity relationship (QSAR) models to determine relevant molecular substructures and descriptors contributing to these adverse effects. Additionally, we calculated physicochemical properties associated with lipid accumulation in hepatocytes and examined their correlation with our chemical structure characteristics. To assess the applicability of our findings to a wide range of chemical compounds, we employed two external datasets to evaluate the distribution of our QSAR descriptors. Our study highlights the significance of subtle molecular structural variations in triggering hepatotoxicity, such as the presence of nitrogen or the specific arrangement of substitutions within the carbon chain. By employing our comprehensive approach, we pinpointed specific molecules and elucidated their mechanisms of toxicity, thus offering valuable insights to guide future toxicology investigations.

The role of trust in the use of artificial intelligence for chemical risk assessment

Risk assessment of chemicals is a time-consuming process and needs to be optimized to ensure all chemicals are timely evaluated and regulated. This transition could be stimulated by valuable applications of in silico Artificial Intelligence (AI)/Machine Learning (ML) models. However, implementation of AI/ML models in risk assessment is lagging behind. Most AI/ML models are considered 'black boxes' that lack mechanistical explainability, causing risk assessors to have insufficient trust in their predictions. Here, we explore 'trust' as an essential factor towards regulatory acceptance of AI/ML models. We provide an overview of the elements of trust, including technical and beyond-technical aspects, and highlight elements that are considered most important to build trust by risk assessors. The results provide recommendations for risk assessors and computational modelers for future development of AI/ML models, including: 1) Keep models simple and interpretable; 2) Offer transparency in the data and data curation; 3) Clearly define and communicate the scope/intended purpose; 4) Define adoption criteria; 5) Make models accessible and user-friendly; 6) Demonstrate the added value in practical settings; and 7) Engage in interdisciplinary settings. These recommendations should ideally be acknowledged in future developments to stimulate trust and acceptance of AI/ML models for regulatory purposes.

Computational strategies for the discovery of biological functions of health foods, nutraceuticals and cosmeceuticals: a review

Scientific and consumer interest in healthy foods (also known as functional foods), nutraceuticals and cosmeceuticals has increased in the recent years, leading to an increased presence of these products in the market. However, the regulations across different countries that define the type of claims that may be made, and the degree of evidence required to support these claims, are rather inconsistent. Moreover, there is also controversy on the effectiveness and biological mode of action of many of these products, which should undergo an exhaustive approval process to guarantee the consumer rights. Computational approaches constitute invaluable tools to facilitate the discovery of bioactive molecules and provide biological plausibility on the mode of action of these products. Indeed, methodologies like QSAR, docking or molecular dynamics have been used in drug discovery protocols for decades and can now aid in the discovery of bioactive food components. Thanks to these approaches, it is possible to search for new functions in food constituents, which may be part of our daily diet, and help to prevent disorders like diabetes, hypercholesterolemia or obesity. In the present manuscript, computational studies applied to this field are reviewed to illustrate the potential of these approaches to guide the first screening steps and the mechanistic studies of nutraceutical, cosmeceutical and functional foods.

In vivo toxicity of nitroaromatic compounds to rats: QSTR modelling and interspecies toxicity relationship with mouse

Nitroaromatic compounds (NACs) in the environment can cause serious public health and environmental problems due to their potential toxicity. This study established quantitative structure-toxicity relationship (QSTR) models for the acute oral toxicity of NACs towards rats following the stringent OECD principles for QSTR modelling. All models were assessed by various internationally accepted validation metrics and the OECD criteria. The best QSTR model contains seven simple and interpretable 2D descriptors with defined physicochemical meaning. Mechanistic interpretation indicated that van der Waals surface area, presence of C-F at topological distance 6, heteroatom content and frequency of C-N at topological distance 9 are main factors responsible for the toxicity of NACs. This proposed model was successfully applied to a true external set (295 compounds), and prediction reliability was analysed and discussed. Moreover, the rat-mouse and mouse-rat interspecies quantitative toxicity-toxicity relationship (iQTTR) models were also constructed, validated and employed in toxicity prediction for true external sets consisting of 67 and 265 compounds, respectively. These models showed good external predictivity that can be used to rapidly predict the rat oral acute toxicity of new or untested NACs falling within the applicability domain of the models, thus being beneficial in environmental risk assessment and regulatory purposes.

Computational Evaluation and In Vitro Validation of New Epidermal Growth Factor Receptor Inhibitors

BACKGROUND

The Epidermal Growth Factor Receptor (EGFR) is a transmembrane protein that acts as a receptor of extracellular protein ligands of the epidermal growth factor (EGF/ErbB) family. It has been shown that EGFR is overexpressed by many tumours and correlates with poor prognosis. Therefore, EGFR can be considered as a very interesting therapeutic target for the treatment of a large variety of cancers such as lung, ovarian, endometrial, gastric, bladder and breast cancers, cervical adenocarcinoma, malignant melanoma and glioblastoma.

METHODS

We have followed a structure-based virtual screening (SBVS) procedure with a library composed of several commercial collections of chemicals (615,462 compounds in total) and the 3D structure of EGFR obtained from the Protein Data Bank (PDB code: 1M17). The docking results from this campaign were then ranked according to the theoretical binding affinity of these molecules to EGFR, and compared with the binding affinity of erlotinib, a well-known EGFR inhibitor. A total of 23 top-rated commercial compounds displaying potential binding affinities similar or even better than erlotinib were selected for experimental evaluation. In vitro assays in different cell lines were performed. A preliminary test was carried out with a simple and standard quick cell proliferation assay kit, and six compounds showed significant activity when compared to positive control. Then, viability and cell proliferation of these compounds were further tested using a protocol based on propidium iodide (PI) and flow cytometry in HCT116, Caco-2 and H358 cell lines.

RESULTS

The whole six compounds displayed good effects when compared with erlotinib at 30 μM. When reducing the concentration to 10μM, the activity of the 6 compounds depends on the cell line used: the six compounds showed inhibitory activity with HCT116, two compounds showed inhibition with Caco-2, and three compounds showed inhibitory effects with H358. At 2 μM, one compound showed inhibiting effects close to those from erlotinib.

CONCLUSION

Therefore, these compounds could be considered as potential primary hits, acting as promising starting points to expand the therapeutic options against a wide range of cancers.

Norm index-based QSTR model to predict the eco-toxicity of ionic liquids towards Leukemia rat cell line

The evaluation of eco-toxicity of ionic liquids (ILs) in the aquatic environment is essential for their safe utilization and QSTR approach plays an important role in obtaining the eco-toxicity data of ILs with diverse structures. Usually, the descriptors used to build QSTR model were made up of anion and cation descriptors, and their interactions were often neglected to some extent. In this work, based on the optimization of the ILs structure, a new set of descriptors were proposed to describe the interaction between anions and cations, and some new atomic distribution matrices were constructed to calculate norm descriptors of ILs, anion and cation. A norm index-based QSTR model was built to predict the eco-toxicity of ILs toward Leukemia rat cell line (IPC-81). This model has satisfactory statistical results with the R² of 0.954 and RMSE of 0.241, respectively. Furthermore, leave-one-out cross-validation and applicability domain results showed good stability and predictability of this model. This approach showed that the interaction between cations and anions could be reflected by optimizing the whole structure of ILs which might play an important role for describing the eco-toxicity of ILs. Therefore, it is further suggested that the norm descriptors would be applicable to predict the eco-toxicity of ILs towards IPC-81.