Molecular simulation and machine learning assisted in exploring betaine-based deep eutectic solvent extraction of active compounds from peony petals

Microwave-assisted deep eutectic solvent extraction (MAE-DES) method was proposed to extract of active compounds from peony petals. Four different components of DES were used as extraction solvent, and the extraction effect was evaluated by total phenolic content (TPC), total flavonoid content (TFC) and total anthocyanin content (TAC). Response surface methodology (RSM) was used to optimize the reaction temperature, reaction time and liquid–solid ratio. The results showed that under optimized conditions, the extraction yields of TPC, TFC and TAC were 321.59 mg GAE/g, 61.65 mg RE/g and 2.15 mg C3GE/g, respectively. The characterization of solid residue showed that DES effectively disrupted the structure of peony petals. In addition, the importance of variables affecting extraction efficiency was investigated by machine learning (ML). Finally, the extraction mechanism of DES for active compounds was explored by density functional theory (DFT) and molecular dynamics (MD) simulation. The results showed that the extraction efficiency of active compounds was affected by the hydrogen bond interaction with DES. This study provides efficient extraction protocol and mechanism research for the extraction of active compounds from multiple methods and insights.