Support Vector Machines for Protein Family Identification Using Surface Invariant Coordinates

Identification of protein family or classification of proteins is one of the trying and important problems in bioinformatics and modern day biology. It's important to determine the family of a protein because of sheer importance of proteins in our life. In the past many methods based on the structure and sequence composition of proteins have been developed to classify proteins and to determine the family of proteins. But no work has been done in this regard by taking into account the roughness of the protein surfaces. Here we demonstrate a method using support vector machines and exploiting the surface roughness of the protein to identify the family of a protein. This method can also be used to classify proteins into different families.