Eukaryotic Plasma Cholesterol Prediction from Human GPCRs Using K-Means with Support Vector Machine

In cell physiology, eukaryotic membrane components have played significant role in every aspect of human body. Amid all, membrane cholesterol is one who has the power to drop off the permeability of biological membranes to a variety of solutes. Generally, cholesterol is distributed heterogeneously in membrane protein. Additionally, the length of membrane protein is laid from extracellular region to intracellular region, and membrane cholesterol is always targeted in both forward and backward direction. We know that G-protein-coupled receptor is a superfamily in mammalian cells and it contains more than 800 proteins in it. Among these 800 proteins, half of them are olfactory proteins. So it is very difficult to predict exact signature motif of cholesterol with GPCR protein. So we have focused our research except on olfactory receptor. In case of data mining applications, clustering plays a crucial role. Clustering is nothing but grouping of similar objects in one cluster points. There are numerous algorithms which have been utilized for clustering the same data points. Among all K-means is an important algorithm which is used when we have unlabeled data. Support vector machine (SVM) plays a significant role for prediction and classification among dissimilar objects. For that purpose we have developed a hybrid approach, that is, K-means and support vector machine, for prediction of membrane protein GPCR and membrane cholesterol, and this experiment provides enhanced prediction results according with our data.