Unfolding the Locale of Membrane Proteins within Cellular Alcove by SVM

Abstract

The Membrane Proteins (MPs) on the cell membrane, is an ideal target for drug delivery owing to their distinct location on the cell membrane and intracellular structure. To locate the multitude of MPs with better accuracy, we implemented SVM on Benchmark MP Database, MemLoci Dataset. A total of 3000 proteins are selected with 1000 plasma membrane proteins, 1000 internal membrane proteins, and 1000 organelle membrane proteins. The amino acid sequence of the proteins is converted into Pseudo Amino Acid Code. After feature extraction, the data is trained and tested, yielding an overall accuracy of 78%. The output is displayed through Graphical User Interface (GUI) which discloses the category of MP based on the cellular site where they are residing.