Computational analysis of NaV1.7 protein variants and tool for 3D visualization of protein structures

Abstract

This thesis is composed of two parts. The first part explores the possibility to use Graph Kernels to discriminate pathogenic versus non-pathogenic variants of a specific protein. All variants are represented as Residue Interaction Networks (RIN), where nodes are amino acids and edges represent non-covalent bonds between atoms of the two involved amino acids. This part is guided by a previous Master degree thesis that considered protein NaV1.7, which is responsible for the transmission of the pain signal from the peripheral nervous system to the brain. The thesis considered 85 genetic variants of the human NaV1.7, among which 30 are known to cause neuropathic diseases and 55 are instead neutral. The results of the first part highlight that some Graph Kernels are actually able to discriminate between pathogenic and neutral variants. This prompted the idea of realizing a 3D viewer able to show the three-dimensional structure of a protein and also its non-covalent bonds. The second part of the thesis describes Spheremole, an application for the visualization of the three-dimensional structure of a protein. In particular, Spheremole allows the visualization of two proteins structures and their visual comparison, also based on their non-covalent bonds.