ABSTRACT
Juvenoids are chemicals that mimic the activity of the juvenile hormone secreted by the corpora allata. These man-made chemicals selectively target and disrupt the endocrine system of insects, derailing their development irreversibly. Consequently, they are particularly suited as larvicides for the control of mosquitoes. In this context, an attempt was made to review all the structure–activity relationship (SAR) and quantitative structure–activity relationship (QSAR) models aiming at predicting the juvenoid activity of chemicals against mosquitoes. Analysis of these models shows that, to be correctly modeled, the juvenoid activity, even for structurally related molecules, needs to be encoded by different categories of descriptors, which are very often related to some structural characteristics of the molecules and to steric effects. The 1-octanol/water partition coefficient appears less significant except when it also encodes size and branching information such as the autocorrelation vector of lipophilicity. Last, this review reveals that the supervised nonlinear methods outperform the linear regression techniques in the design of powerful predictive QSAR models for juvenoids.
