Prediction of homo-oligomeric proteins based on nearest neighbour algorithm

Nearest neighbour algorithm is applied to discriminating homo-oligomeric proteins from the protein primary structure. The algorithm describes protein primary sequences by their subsequence distributions, therefore, the information of protein sequences is sufficiently taken into account. When the length of subsequence increases, all performance measures have the tendency to ascend quickly for the tests of the method on the three data sets. Our tests demonstrate that the residue order along protein sequences plays an important role in recognition of the homo-oligomers, and nearest neighbour algorithm method is a simple and effective tool for classification of homo-oligomeric proteins. It is further confirmed that protein primary sequence encodes quaternary structure information.