Process optimization, multi-gene genetic programming modeling and reliability assessment of bioactive extracts recovery from Phyllantus emblica

Abstract

This study investigates the feasibility of extracting bioactive antioxidants from Phyllantus emblica leaves using a combination of ethanol-water mixture (0-100%) and heat-assisted extraction technology (HAE-T). Operating temperature (30-50°C), solid-to-liquid ratio (1:20-1:60 g/mL), and extraction time (45-180 min) were varied to determine their effects on extract total phenolic content (TPC), yield (EY), and antioxidant activity (AA). The Box-Behnken experimental design (BBD) within response surface methodology (RSM) was employed, with multi-objective process optimization using the desirability function algorithm to find the optimal process variables for maximizing TPC, EY, and AA simultaneously. The extraction process was modeled using BBD-RSM and multi-gene genetic programming (MGGP) algorithm, with model reliability assessed via Monte Carlo simulation. HPLC characterization identified betulinic acid, gallic acid, chlorogenic acid, caffeic acid, ellagic acid, and ferulic acid as bioactive constituents in the extract. The study found that a 50% ethanol solution yielded the best extraction efficiency. The optimal process parameters for maximum EY (21.6565%), TPC (67.116 mg GAE/g), and AA (3.68583 µM AAE/g) were determined as OT of 41.61°C, S:L of 1:60 g/mL, and ET of 180 min. Both BBD-RSM and MGGP-based models satisfactorily predicted the observed process responses, with BBD-RSM models showing slightly better performance. Reliability analysis indicated high certainty in the predictions, with BBD-RSM models achieving 99.985% certainty for TPC, 97.569% for EY, and 98.661% for AA values.