Optimizing bioactive compounds extraction from different medicinal plants and prediction through nonlinear and linear models

Abstract

Optimizing the extraction methods of bioactive compounds are of great interest for research and development in the food and pharmaceutical industries. This study investigated five medicinal plants: lavender (Lavandula x hybrida L.), lemon balm (Melissa officinalis L.), mint (Metha piperita L.), sage (Salvia officinalis L.), and thyme (Thymus serpyllum L.), with the focus on the extraction conditions and analysis of the total polyphenolic content, antioxidant activity, conductivity and extraction yield. In order to optimise the extraction conditions the Taguchi method was used. Nonlinear and piecewise linear regression models were developed for prediction of physical and chemical properties of medicinal plant extracts based on experimental process conditions. For extracts of all five medicinal plants the expected total polyphenolic content and the antioxidant activity can be predicted with high determination coefficients r2 > 0.9) by piecewise linear regression models, while the prediction of conductivity and extraction yield resulted with determination coefficients slightly under 0.9, which is still highly acceptable considering that the models present the prediction for all five plants. Proposed nonlinear models can be used as qualitative models in relation of the prediction of bioactive compounds, while the piecewise linear regression models confirmed they dominance in the quantification models.