Using PyCaret to model Chlorella vulgaris's growth response to salinity and oil contamination for crude oil bioremediation.

Abstract

Crude oil spills significantly impact aquatic ecosystems, necessitating innovative remediation strategies. Microalgae-based bioremediation, particularly with Chlorella vulgaris, offers a promising solution. This study introduces a novel framework that evaluates the combined effects of selected environmental stressors on microalgal adaptability, advancing beyond traditional isolated factor analyses. By integrating a factorial experimental design with a machine learning approach using PyCaret AutoML and SHAP values, we provide a detailed examination of how crude oil concentration, salinity, and exposure duration affect C. vulgaris growth. The Extra Trees Regressor model emerged as highly accurate in predicting biomass concentration, a crucial adaptability indicator, achieving an MAE of 0.0202, RMSE of 0.029, and an R² of 0.8875. SHAP analysis highlighted salinity and crude oil as significant growth influencers, with exposure duration playing a minor role. Notably, C. vulgaris exhibited more sensitivity to salinity than to crude oil, indicating potential high-salinity challenges but also a strong tolerance to oil pollutants. These findings enhance our understanding of microalgal responses in polluted environments and suggest improved bioremediation approaches for saline waters affected by oil spills, leveraging the synergy of environmental factors and machine learning insights.