Virtual-screening of xanthine oxidase inhibitory peptides: Inhibition mechanisms and prediction of activity using machine-learning.
Journal:
Food chemistry
PMID:
39128372
Abstract
Xanthine oxidase (XO) inhibitory peptides can prevent XO-mediated hyperuricemia. Currently, QSAR about XO inhibitory peptides with different lengths remains to be enriched. Here, XO inhibitory peptides were obtained from porcine visceral proteins through virtual-screening. A prediction model was established by machine-learning. Virtual-screening retained four lengths of peptides, including 3-6. Molecular-docking recognized their binding sites with XO and showed residues W, F, and G were the key amino acids. Datasets of XO inhibitory peptides therewith were established. The optimal model was used to generalize the peptides reported. Results showed that the R of the tripeptide, tetrapeptide, pentapeptide and hexapeptide in the generalisation test were R = 0.81, R = 0.82, R = 0.83 and R = 0.83, respectively. Overall, this work can serve as a reference for explaining the activity mechanism of XO inhibitory peptides and predicting the activity of XO inhibitory peptides.