Machine learning aided experimental approach for evaluating the growth kinetics of Candida antarctica for lipase production

A hybrid machine learning (ML) aided experimental approach was proposed in this study to evaluate the growth kinetics of Candida antarctica for lipase production. Different ML models were trained and optimized to predict the growth curves at various substrate concentrations. Results on comparison demonstrate the superior perfor-mance of the Gradient boosting regression (GBR) model in growth curves prediction. GBR-based growth kinetics was found to be matching well with the results of the conventional experimental approach while significantly reducing the experimental effort, time, and resources by ~ 50%. Further, the activity and enzyme kinetics of lipase produced in this study was investigated on hydrolysis of p-nitrophenyl butyrate resulting in a maximum lipase activity of 24.07 U at 44 h. The robustness and significance of developed kinetic models were ensured through detailed statistical analysis. The application of the proposed hybrid approach can be extended to any other microbial process.

Automated summary

In this study, a hybrid machine learning aided experimental approach was proposed to evaluate the growth kinetics of Candida antarctica for lipase production. The Gradient boosting regression (GBR) model showed superior performance in predicting the growth curves. The application of this hybrid approach significantly reduced the experimental effort, time, and resources, while matching well with the results of the conventional experimental approach. Additionally, the activity and enzyme kinetics of the lipase produced were investigated and the robustness of the kinetic models were ensured through statistical analysis.