Machine learning aided experimental approach for studying the growth
kinetics of Candida antarctica for lipase production
Abstract
This paper reports a systematic study for evaluating the growth and
enzyme kinetics of Candida antarctica MTCC 2706 for the first time. A
hybrid machine learning (ML) aided experimental approach is proposed
here to evaluate the microbial growth kinetics in which different ML
models were built with a smaller set of experimental samples to predict
the growth curve at different substrate concentrations and their
performance was compared. Gradient boosting regression model was found
to predict the growth curves effectively. Further, growth kinetics was
evaluated and the data was found to be fitted well with Monod’s model.
In addition, the activity of lipase produced in this study was
evaluated, and the resulting enzyme kinetics were studied. Further, the
statistical significance of models developed was quantified using
average absolute deviation percentage (AAD %) and R2 parameters, and
robustness of the model parameters is ensured through sensitivity
analysis. The results of the proposed ML-based approach were found to be
in good agreement with that of the conventional method while
successfully reducing the experimental efforts, time and resources by
nearly 50%. Also, the application of the proposed method can be
extended to any other microbial process.