An Effective Data Mining Technique for the Multi-Class Protein Sequence Classification

One way to understand the molecular mechanism of a cell is to understand the function of each protein encoded in its genome. The function of a protein is largely dependent on the three-dimensional structure the protein assumes after folding. Since the determination of three-dimensional structure experimentally is difficult and expensive, an easier and cheaper approach is for one to look at the primary sequence of a protein and to determine its function by classifying the sequence into the corresponding functional family. In this paper, we propose an effective data mining technique for the multi-class protein sequence classification. For experimentations, the proposed technique has been tested with different sets of protein sequences. Experimental results show that it outperforms other existing protein sequence classifiers and can effectively classify proteins into their corresponding functional families.