QSTR Modeling for Qualitative and Quantitative Toxicity Predictions of Diverse Chemical Pesticides in Honey Bee for Regulatory Purposes

Abstract

Pesticides are designed toxic chemicals for specific purposes and can harm nontarget species as well. The honey bee is considered a nontarget test species for toxicity evaluation of chemicals. Global QSTR (quantitative structure-toxicity relationship) models were established for qualitative and quantitative toxicity prediction of pesticides in honey bee (Apis mellifera) based on the experimental toxicity data of 237 structurally diverse pesticides. Structural diversity of the chemical pesticides and nonlinear dependence in the toxicity data were evaluated using the Tanimoto similarity index and Brock-Dechert-Scheinkman statistics. Probabilistic neural network (PNN) and generalized regression neural network (GRNN) QSTR models were constructed for classification (two and four categories) and function optimization problems using the toxicity end point in honey bees. The predictive power of the QSTR models was tested through rigorous validation performed using the internal and external procedures employing a wide series of statistical checks. In complete data, the PNN-QSTR model rendered a classification accuracy of 96.62% (two-category) and 95.57% (four-category), while the GRNN-QSTR model yielded a correlation (R(2)) of 0.841 between the measured and predicted toxicity values with a mean squared error (MSE) of 0.22. The results suggest the appropriateness of the developed QSTR models for reliably predicting qualitative and quantitative toxicities of pesticides in honey bee. Both the PNN and GRNN based QSTR models constructed here can be useful tools in predicting the qualitative and quantitative toxicities of the new chemical pesticides for regulatory purposes.

Predicting acetyl cholinesterase enzyme inhibition potential of ionic liquids using machine learning approaches: An aid to green chemicals designing

“…The model domain can be represented by ranges of the molecular descriptors for training set compounds [35]. In this study, a twodimensional descriptor space is considered.…”

Graph attention convolutional neural network model for chemical poisoning of honey bees’ prediction

“…The unpredictability of chemical poisoning of bees may be due to an incomplete understanding of the complex underlying toxicology of pesticides. In addition, thoroughly evaluating the effects of a large number of chemicals on bee poisoning using experimental tests is still challenging because of cost, time, and ethical concerns [10,11]. Computational prediction of bee poisoning (CPBP) may help scientists to identify toxic chemicals faster [12,13].…”

Towards the Development of Innovative Quantitative Structure-Activity Relationship Models for Human and Ecological Risk Assessment of Chemicals and their Mixtures

“…These QSAR models for species of ecological importance are often developed using the well-characterised relationships between the physico-chemical properties of chemicals, their persistence and toxicity as well as their global environmental fate (Domine et al, 1992;Devillers and Flatin, 2000). QSAR models have already been applied to the prediction of ecotoxicological endpoints in species of ecological importance including trout, daphnia, quail and bees within the project DEMETRA (Benfenati et al, 2011) and more recently for qualitative and quantitative toxici ty prediction in bees using a global quantitative structure-toxicity relationship model (QSTR) (Singh et al, 2014). To address the needs for QSAR models predicting toxicity of PPPs in honey bees, we developed an in-house software using databases on acute contact data in honey bees from different sources and a k-Nearest Neighbor algorithm (k-NN).…”

“…Although the authors identified numerous data gaps, such combined toxicity databases potentially allow developing predictive Quantitative Structure-Activity Relationships (QSAR) tools particularly because it is rather impossible to test all possible mixtures in bees for their acute or chronic effects. Such QSAR tools are only available for single chemicals but to date not for the prediction of combined toxicity (Venko et al, 2017;Singh et al, 2014;. Hence, this manuscript describes the development and application of three innovative predictive QSAR models for honey bees within the CORAL software namely (i) two regression-based QSAR models predicting acute (contact) mixtures potency (pLD 50-mix) in a quantitative manner, and (ii) a classification-based model predicting the nature of combined toxici ty for organic binary mixtures (i.e.…”

“…However, such tools are mostly based only on vertebrate information specifically fish toxicity data, thus limiting their application to aquatic environmental RA. Similarly, different QSAR models predicting toxicity for single chemicals in honey bees are available but to date these do not address the challenge of the integration of toxicity prediction together with MoA profiling (Venko et al, 2018;Singh et al, 2014;Toropov and Benfenati, 2007;Devillers et al 2002). MoA information together with information on adverse outcome pathways (AOPs) in honey bees can provide a sound understanding of the link between molecular targets , as molecular initiating event and key events leading to adverse effects at individual and colony level as recently for neonicotinoids targeting nicotinic acetylcholine receptors in honey bees (LaLone et al 2017).…”