Using Algorithm5 for classification of new protein structures

The goal of all research related to structural alignment is how accurately can we predict protein structures from the PDB which are close to a newly discovered protein structure $P_{new}$ . This can be easily addressed by computing all the parwise Z-scores's between the new protein $P_{new}$ and existing proteins in the PDB database, and rank all of the alignments according to the Z-Score, then we select top-K (K is typically 10) and findout which class and which superfamily occurs the maximum and classify the new protein's class and superfamily. We give extensive experimental details of our algorithm5, and compare our algorithms average accuracy with that of some of the latest algorithms in this area, one such algorithm is PSIST[3] we give more details about existing algorithms in the upcoming sections.