loading page

Machine learning aided experimental approach for evaluating the growth kinetics of Candida antarctica for lipase production
  • +4
  • Nipon Sarmah,
  • Vazida Mehtab,
  • Pratyusha Bugata,
  • James Tardio,
  • Suresh Bhargava,
  • Rajarathinam Parthasarathy,
  • Sumana Chenna
Nipon Sarmah
Indian Institute of Chemical Technology CSIR
Author Profile
Vazida Mehtab
Indian Institute of Chemical Technology CSIR
Author Profile
Pratyusha Bugata
Indian Institute of Chemical Technology CSIR
Author Profile
James Tardio
Royal Melbourne Institute of Technology
Author Profile
Suresh Bhargava
Royal Melbourne Institute of Technology
Author Profile
Rajarathinam Parthasarathy
Royal Melbourne Institute of Technology
Author Profile
Sumana Chenna
Indian Institute of Chemical Technology CSIR

Corresponding Author:[email protected]

Author Profile

Abstract

Lipase derived from Candida antractica is the most widely used enzyme for catalyzing various reactions. This paper reports the growth and enzyme kinetics of Candida antarctica MTCC-2706 for lipase production, which is one of the fundamental steps in bioprocess design, optimization, and scale-up studies. A hybrid machine learning (ML) aided experimental approach is proposed here to evaluate growth kinetics in which, different ML models were built to predict the growth curves at various substrate concentrations using a substantially smaller set of experimental samples. Comparative assessment of model performances revealed the superiority of gradient boosting regression (GBR) in predicting the growth curves. GBR-based growth kinetics was found to be fitted well with Monod’s model. Further, the activity and enzyme kinetics of lipase produced was investigated by considering the hydrolysis of p-nitrophenyl butyrate. The maximum lipase activity resulted was 24.07 U at 44 hrs. The deviation and R2 values of Monod’s and Michaelis-Menten’s models were 1.4% and 2.25%, and 0.96 and 0.99, respectively. The proposed ML-based approach is found to be efficient in predicting the growth kinetics with reduced experimental effort, time and resources (~50%) as compared to conventional method and its application can be extended to any other microbial processes.