Artificial neural networks and response surface methodology approach for optimization of an eco-friendly and detergent-stable lipase production from Actinomadura keratinilytica strain Cpt29

Abstract

This work mainly focused on the production of an efficient, economical, and eco-friendly lipase (AKL29) from Actinomadura keratinilytica strain Cpt29 isolated from poultry compost in north east of Algeria, for use in detergent industries. AKL29 shows a significant lipase activity (45 U/mL) towards hydrolyzed triacylglycerols, indicating that it is a true lipase. For maximum lipase production the modeling and optimization of potential culture parameters such as incubation temperature, cultivation time, and Tween 80 (v/v) were built using RSM and ANN approaches. The results show that both the two models provided good quality predictions, yet the ANN showed a clear superiority over RSM for both data fitting and estimation capabilities. A 4.1-fold increase in lipase production was recorded under the following optimal condition: incubation temperature (37.9 °C), cultivation time (111 h), and Tween 80 (3.27%, v/v). Furthermore, the partially purified lipase showed good stability, high compatibility, and significant wash performance with various commercial laundry detergents, making this novel lipase a promising potential candidate for detergent industries.