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 in silico, in vitro, and machine learning strategy. Furthermore, the inhibitory mechanism of DPP-IV inhibitory peptides was elucidated by kinetics, molecular docking and simulation. Additionally, their in vitro digestive stability was assessed. In silico 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 in silico and in vitro strategy is feasible for rapid screening of DPP-IV inhibitory peptides from goat blood proteins.