Dynamic modelling of tetrazolium-based microbial toxicity assay—a parametric proxy of traditional dose-response relationship

Abstract

Microbial toxicity of test substances in tetrazolium assay is often quantified while referring to IC50 values. However, the implication of such estimates is very limited and can differ across studies depending on prevailing test conditions. In this work, a factorial design-based end-point microbial assay was adopted, which suggests significant interaction (P= 0.041) between inoculum and tetrazolium dose on formazan production. A dynamic model framework was proposed to be incorporated in toxicity assay, that not only captures the nonlinear interdependency between biomass, substrate, formazan content but also measure the toxicity in terms of inhibition parameter. Microbial growth, glucose uptake, formazan production in presence and absence of heavy metal toxicant (Cu 2+ ) in designed batch studies were utilized for sequential estimation of model parameters and their bootstrap confidence intervals. A logistic growth model (R 2 >0.96) with multiplicative inhibition terms fit the experimental data reasonably well. Dynamic relative sensitivity analysis revealed that both microbial growth and formazan production profiles were sensitive to toxicant inhibition parameter. The adoption of a dynamic model framework as a stable index for the toxic potential of test substances can be extended to design a versatile, robust in-vitro assay system.

Structural variability, implementational irregularities in mathematical modelling of polyhydroxyalkanoates (PHAs) production—A state‐of‐the‐art review

“…Mathematical consistency, that is, boundedness of the variables and asymptotic behaviour must be ascertained. Model discretization strategy can be adopted, where parameters are sequentially estimated using experimental data designed to explore incremental model subspace (Philus & Mahanty, 2021). Uncertainty propagation from experimental data to parameters should be supported with confidence intervals estimate as an assessment for model accuracy.…”