Comparison of null models predicting mixture toxicity of synthetic insecticides against Musca domestica L. (Diptera:...

The increasing prevalence of insecticidal products containing mixtures of active ingredients has highlighted the need for computational approaches to predict the insecticidal activity at specific dose combinations. Several models predicting the interactions have been proposed, each based on different assumption regarding mixture toxicity. However, the lack of consensus on which model to adopt for insecticidal activity has hindered the accurate prediction of mixture toxicity and the identification of synergistic or antagonistic interactions. In the present study, we compared the dose-responses of 14 synthetic insecticides in 3 modes of action groups assigned by the Insecticide Resistance Action Committee. The structure-activity relationships of dose-response parameters were analyzed, with a particular focus on LD50 and Hill slope values, which play pivotal roles in some models. Four widely adopted models, Bliss, Loewe, Highest Single Agent, and Schindler, were evaluated to predict mixture toxicity using Musca domestica L. (Diptera: Muscidae) as the model insect. The Loewe and Bliss models demonstrated 79.1% and 76.9% predictability, respectively, for 91 binary mixtures. To improve predictive accuracy, a 2-step framework was devised. Combinations were grouped based on the similarity of their modes of action, and then the models were selectively applied to the corresponding group. These findings contribute to a better understanding of the development and assessment of mixture-based insecticidal products.