New Publication: Structural Alerts and Random Forest Models in a Consensus Approach for Receptor Binding Molecular Initiating Events

A molecular initiating event (MIE) is the gateway to an adverse outcome pathway (AOP), a sequence of events ending in an adverse effect. In silico predictions of MIEs are a vital tool in a modern, mechanism-focused approach to chemical risk assessment. For 90 biological targets, each representing an important human MIE, structural alert-based models have been constructed with an automated procedure that uses Bayesian statistics to iteratively select substructures. These models give impressive average performance statistics across biological targets (77% Sensitivity, 94% Specificity, 92% Accuracy, and 0.75 Matthews Correlation Coefficient in test sets), significantly improving on previous models. Random Forest models have been constructed from 200 physicochemical features for the same targets, giving similarly impressive average performance statistics (81% Sensitivity, 92% Specificity, 93% Accuracy, 0.77 Matthews Correlation Coefficient in test sets). The structural alert models are transparent and easy to interpret, whilst the most important physicochemical features can be identified with Random Forest models. The two complementary models have been combined in a consensus model, improving performance compared to each individual model (92% of test set predictions agree between models, 92% Sensitivity, 92% Specificity, 94% Accuracy, 0.86 Matthews Correlation Coefficient) and increasing confidence in predictions. Variation in model performance has been explained by calculating a modelability index, using Tanimoto coefficient between Morgan fingerprints to identify nearest neighbor chemicals. This work significantly improves upon previous structural alert-based models and shows how they can be used in a consensus approach with complementary models to increase confidence in MIE predictions. This represents an important step towards building confidence in in silico tools for assessment of toxicity.

View the publication here.


Comments are closed.