Quantitative prediction of the growth inhibition of various harmful chemicals by statistical analysis of delayed fluorescence decay curves obtained from the green alga Raphidocelis subcapitata

Abstract

The toxicity of chemical substances to algal growth is generally measured by the 72–96 h algal growth inhibition test. We have developed a method to assess the toxicity of chemicals in aquatic environments more quickly and simply than conventional testing methods by delayed fluorescence (DF), which reflects the photosynthetic capacity of algae. The DF method is based on a technique for evaluating the amount of change in the decay curve due to the effects of chemicals (${\varDelta}{{{\mathrm{DF}}}}$, DF inhibition). Various studies on DF have been reported; however, few reports have evaluated the decay curve of DF by approach using inductive modeling based on measurement data such as principal component analysis (PCA) and partial least squares regression analysis (PLS). Therefore, the purpose of this study is to examine methods for estimating the magnitude and type of toxicity of chemicals by means of a principal component model (PC model) and multiple regression model (MR model) derived from changes in the decay curves of DF of algae exposed to a wide range of 37 toxic substances that have an effect of clear magnitude on algal growth. The changes in the DF decay curves due to exposure the 37 toxic substances to algae were summarized in the PC model composed of eigenvectors and scores of four principal components. For validation of usefulness, a hierarchical cluster analysis (HCA) of the amount of change in four PC scores revealed that the growth inhibition rate was more influential than the chemical type. We also found the possibility of quantitatively predicting the growth inhibition of chemicals by MR model by the amount of change in the PC scores.