AI-Driven Rapid Screening and Characterization of Dipeptidyl Peptidase-IV (DPP-IV) Inhibitory Peptides from Goat Blood Proteins: An Integrative In Silico and Experimental Strategy

To enhance the screening efficiency of bioactive peptides, an AI-driven approach was employed to screen DPP-IV inhibitory peptides from goat blood proteins by an integrated <i>in silico</i>, <i>in vitro</i>, and machine learning strategy. Furthermore, the inhibitory mechanism of DPP-IV inhibitory peptides was elucidated by kinetics, molecular docking and simulation. Additionally, their <i>in vitro</i> digestive stability was assessed. <i>In silico</i> results revealed that goat blood proteins were promising precursors of DPP-IV inhibitory peptides, while bromelain was the optimal protease. Their peptide sequences were further identified by peptidomics and predicted by self-developed machine learning models (LightGBM) to identify the potent DPP-IV inhibitory peptides. Two novel DPP-IV inhibitory peptides were identified (FPL and FPHFDL). Enzyme kinetics, molecular docking and molecular simulation data indicated that FPL served as a competitive inhibitor, whereas FPHFDL was a non-competitive inhibitor. Overall, the integrative <i>in silico</i> and <i>in vitro</i> strategy is feasible for rapid screening of DPP-IV inhibitory peptides from goat blood proteins.