A nanoparticle–plant dataset with atomistic descriptors was developed and used to train an optimized XGBoost model to predict nanoparticle effects on plant length without experimental data. Image designed with https://canva.com.

The ubiquity and environmental persistence of per- and polyfluoroalkyl substances (PFASs) have raised significant concerns about their detrimental effects on human health. Collective scientific efforts are increasingly focused on elucidating PFAS toxicity mechanisms and identifying potential low-impact PFAS structures that retain the exceptional properties of this chemical class. To advance the use of in silico methods in PFAS toxicity assessment, we developed a robust modelling framework for predicting PFAS acute oral toxicity class (high or low) in rats, leveraging the enhanced capabilities of the in-house Isalos Analytics Platform. The automated machine learning (autoML) functionality was employed to optimise four ML models—k-nearest neighbours (kNN), Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and fully connected neural network (NN)—using Mold2 molecular descriptors, and to identify the top-performing model through five-fold cross-validation. The selected kNN model (k = 3) was used for predictions on the held-out testing set, achieving an accuracy of 81.5%, while a Shapley values analysis provided valuable insights into the factors influencing toxicity predictions. Furthermore, the nearest-neighbour-based methodology enabled a read-across structural analysis of PFAS similarity groups consisting of each testing set instance and its three closest neighbours in the training set. This analysis revealed a consistent association between polyaromatic and heterocyclic structural features and high acute oral toxicity. The developed, thoroughly validated read-across model is freely accessible through the INSIGHT RatTox web application as well as the INSIGHT Cheminformatics Platform in Enalos Cloud, supporting high-throughput screening of PFAS compounds and investigation of structural similarities with their nearest neighbours for enriched structural interpretation.

Thyroid-disrupting chemicals (TDCs) pose a significant risk to human health and have been associated with adverse neurological and cardiovascular effects. Among these chemicals, per- and polyfluoroalkyl substances (PFAS) have been shown to bind to transthyretin (TTR), disrupting thyroid hormone homeostasis and leading to adverse outcomes. In this study, we present a refined workflow for modelling and predicting PFAS activity toward TTR, where training compounds were labelled as Active or Inactive based on experimental activity values (binding affinities) exceeding 50% relative to the control, thyroxine, the biological binding partner of TTR. A curated dataset comprising experimental data from published studies was processed, resulting in 131 unique PFAS structures for modelling. The dataset was further augmented with molecular descriptors and molecular docking derived scores, followed by model optimisation and evaluation through an automated machine learning (AutoML) approach. Among the tested classification algorithms, XGBoost demonstrated the best performance, achieving an external test accuracy of 0.818. To provide mechanistic insights, molecular docking calculations and Shapley value analysis were employed, revealing that increased PFAS lipophilicity enhances binding within the hydrophobic TTR pocket, in agreement with experimental observations. In accordance with the Findable, Accessible, Interoperable, Reusable (FAIR) principles for data and models, the curated dataset is publicly available through the ChemPharos database (https://db.chempharos.eu/datasets/Datasets.zul?datasetID=ds21). The final model was implemented as a user-friendly web application on the Enalos Cloud Platform, enabling rapid screening of existing or novel PFAS and PFAS-like compounds and supporting chemical safety management (https://enaloscloud.novamechanics.com/scenarios/thyrotox/), and is fully documented via a QSAR model report format (QMRF).

Consensus modelling for the prediction of PFAS acute oral toxicity classes in mice and rats, with the validated consensus models implemented as a web application.