Identification of relevant physico chemical properties of amino acids with respect to protein glycosylation prediction

Abstract

Glycosylation is the process of adding carbohydrates to a protein residue. It is an important part of post-translational modification undergone by protein chains. Over 40 disorders have been linked to improper glycosylation bonds, over 80% of which affect the nervous system. Our aim is to study glycosylation in proteins and to understand the properties that affect this change. A detailed study using support vector machines has given us a computer that accurately predicts 79% percent of the possibility of positive glycan-bond. In this paper we attempt to discuss the entire study and our conclusions. We target to find out which of the physicochemical properties of amino acids are relevant for glycosylation.