Artificial intelligence-guided optimization of Suillus lakei extracts enhancing bioactive profile and functional properties for future food applications

This study aimed to comparatively evaluate the biological properties of Suillus lakei (Murrill) A.H. Sm. & Thiers extracts optimized using response surface methodology (RSM) and artificial neural network-genetic algorithm (ANN-GA) approaches. The extracts were systematically analyzed in terms of their antioxidant potential, anticholinesterase capacity, antiproliferative effects, and phenolic composition. The ANN-GA-optimized extract exhibited significantly higher DPPH and FRAP values and significantly lower TOS values than the RSM-optimized extract, whereas the differences in TAS and OSI were not statistically significant. Enzyme inhibition assays further demonstrated that these extracts exhibited stronger inhibitory activity against both acetylcholinesterase (AChE) and butyrylcholinesterase (BChE). In vitro MTT assays showed that the ANN-GA-derived extract produced a greater reduction in MTT absorbance, particularly at higher concentrations, in A549, MCF-7, and DU-145 cancer cell lines. Additionally, phenolic profiling suggested that these extracts were enriched in bioactive compounds, with gallic acid and quercetin occurring at the highest measured concentrations, followed by protocatechuic and caffeic acids. Overall, under the tested experimental conditions, the ANN-GA-optimized extract exhibited a distinct chemical composition and more favorable values for the measured in vitro biological endpoints than the RSM-optimized extract. This work represents one of the early detailed investigations into the bioactivity profile of S. lakei and provides preliminary in vitro evidence that warrants further validation before its extracts can be considered for potential functional food-related applications.