Identification of antihypertensive peptides from lupine using a machine learning approach

¹ Department of Mathematics, Physics and Information technologies, University of Food Technologies, 26, Maritza blvd, Plovdiv, Bulgaria
² Department of Computer Science and Mathematics, Trakia University, Stara Zagora, Bulgaria
³ Department of Analytical Chemistry and Physical Chemistry, University of Food Technologies, 26, Maritza blvd, Plovdiv, Bulgaria
⁴ Department of Biotechnology, University of Food Technologies, 26, Maritza blvd, Plovdiv, Bulgaria

^* Corresponding author: mterziyska@uft-plovdiv.bg

Abstract

Bioactive products with antihypertensive biological activity, isolated from natural sources, have been the subject of growing interest in recent years. This is due to their widespread use in medicine for the treatment and prevention of various diseases, as well as dietary supplements for athletes or their inclusion in diets for overweight people. One such source is Lupine. Lupine beans are delicious and useful. They can be used in food as a nutritional source of vegetable proteins. They are also rich in polyphenols, carotenoids, and phytosterols. The approaches to screen antihypertensive peptides, based on information technologies and more concretely on machine learning, doubtlessly have higher throughput and rapid speed than the in vivo and in vitro procedures. Therefore, the scientific literature abounds with articles offering various artificial intelligence algorithms for predicting food-derived antihypertensive peptides. In this study, an Adaptive Boosting (AdaBoost) algorithm was developed for these purposes. The results showed that the AdaBoost model as a novel auxiliary tool is feasible to screen for antihypertensive peptides derived from food, with high throughput and high efficiency.