Structural characterization of functional peptides by extending the hybrid orbital theory

Abstract

Abstract In recent years, food‐derived functional peptides have received more and more attention because of their low toxicity, which is different from drugs. This study was based on the theory that hybrid orbitals play an important role in chemical reactions. While preserving the original physical information, a descriptor called hybrid orbitals and atomic characteristics was designed to explore the relationship between hybrid orbitals, electronegative atoms, and the function of food‐derived peptides. The classification effects of support vector machines and K‐nearest neighbor are compared and selected by machine learning KNN algorithm to predict the function of food‐derived functional peptides, including inhibitors of angiotensin‐converting enzyme and dipeptidyl peptidase IV, antioxidant peptides, and antibacterial peptides, the accuracy of prediction and the area under curve value were about 0.8. Comparing the result of using hybrid orbital and electronegative atoms to predict the function of peptides, respectively, it was found that the hybrid orbital is more closely related to the function of the peptide. This study revealed that the application of hybrid orbital theory to the research of food‐derived peptides is of positive significance.

Keywords

• electronegative atom
• food‐derived functional peptides
• hybrid orbital theory
• K‐nearest neighbor algorithm
• peptide structure
• QSAR