Predictive Model for Ecotoxicity in Daphnia magna: Impact of Wildlife Retardants on Aquatic Life

By Revathi Mekkoth | Valley Christian High School, San Jose, California, United States

I. Abstract

Recently, the extent of wildfires and subsequent firefighting efforts with retardants have brought a renewed attention to its impact on freshwater aquatic life. Non animal testing approaches like those based on quantitative structure–activity relationships (QSARs) maximize the information contained in existing experimental data and predict missing information. In this study, a QSAR model was developed from a data set consisting of 1516 organic compounds, to predict acute aquatic toxicity toward Daphnia magna. The goal of the research was to predict the aquatic toxicity in the form of an LC50 value for triethyl phosphate (TEP) - an organic chemical used in some fire retardant formulations as a corrosion inhibitor and flame retardant enhancer. Random Forest ML algorithm, which uses an ensemble of decision trees, was used as the regression method for this data set. Ninety percent of the data was used for training and the remaining was used for validation. The final model showed a coefficient of determination R^2 = 0.5685. The model was built with a tree/estimator size of 500 and a max depth of 50 with a MAPE of 15.44% (Model Size: 62.46 MB). The use of TEP as a test case predicted the pLC50 value with an accuracy of 87.41%. The model used ten molecular descriptors that encoded information about lipophilicity, the formation of H-bonds, polar surface area, polarizability, and electrophilicity. This study shows that the proposed model can be used for reliably predicting the toxicity of an organic chemical similar to the training set.

II. Research Poster

The research poster published, “Predictive Model for Ecotoxicity in Daphnia magna: Impact of Wildlife Retardants on Aquatic Life,” was received on June 18, 2025, and was reviewed and accepted on July 21, 2025. To contact editors and reviewers please click here.