Non-linear regression in insecticide bioassays: Benefits and applications of the drc package

Insecticide bioassay is a critical component of entomological research. Traditionally, probit analysis has been used as the standard method for dose-response analysis in pesticide bioassays due to its historical use and accessibility. However, this method has multiple limitations, including: (a) assumptions of linearity, (b) challenges with log-transformation of zero doses, and (c) increased uncertainty in lethal dose (LD) determination due to large standard errors around the estimate. Using data from several case studies, this project outlines the procedure for non-linear regression analysis utilizing the R software and drc (dose response curve) package. Non-linear regression models presented in drc package can handle binary and continuous responses and also generate more precise estimates of LD values and associated thresholds, such as the non-observable effect level (NOEL) and the lowest observable effect level (LOEL) or resistance ratios. The improvements generated by non-linear regression models can help with resistance monitoring efforts, and ecological risk assessment, which could improve pest management strategies. With this project, we want to promote the adoption of this method for dose-response modelling in entomology.