AI4ACEIP: A Computing\nTool to Identify Food Peptides\nwith High Inhibitory Activity for ACE by Merged Molecular Representation\nand Rich Intrinsic Sequence Information Based on an Ensemble Learning\nStrategy

Hypertension is a common chronic\ndisorder and a major\nrisk factor\nfor cardiovascular diseases. Angiotensin-converting enzyme (ACE) converts\nangiotensin I to angiotensin II, causing vasoconstriction and raising\nblood pressure. Pharmacotherapy is the mainstay of traditional hypertension\ntreatment, leading to various negative side effects. Some food-derived\npeptides can suppress ACE, named ACEIP with fewer undesirable effects.\nTherefore, it is crucial to seek strong dietary ACEIP to aid in hypertension\ntreatment. In this article, we propose a new model called AI4ACEIP\nto identify ACEIP. AI4ACEIP uses a novel two-layer stacked ensemble\narchitecture to predict ACEIP relying on integrated view features\nderived from sequence, large language models, and molecular-based\ninformation. The analysis of feature combinations reveals that four\nselected integrated feature pairs exhibit enhancing performance for\nidentifying ACEIP. For finding meta models with strong abilities to\nlearn information from integrated feature pairs, PowerShap, a feature\nselection method, is used to select 40 optimal feature and meta model\ncombinations. Compared with seven state-of-the-art methods on the\nsource and clear benchmark data sets, AI4ACEIP significantly outperformed\nby 8.47 to 20.65% and 5.49 to 14.42% for Matthew’s correlation\ncoefficient. In brief, AI4ACEIP is a reliable model for ACEIP prediction\nand is freely available at https://github.com/abcair/AI4ACEIP.