A System for Protein Classification Based on Protein 3D Structure

The classification of proteins based on their structure plays an important role in the deduction or discovery of protein function. Furthermore, the large number of potential classes causes problems for many classification strategies, increasing the likelihood that the classifier will reach local optima while trying to distinguish between all of the possible structural categories. In this paper, we present an efficient system for protein classification by using 3D structure content representation. We use a 3D structure-based approach for the efficient classification of protein molecules. The method relies on descriptors extracted from the known protein structure. These descriptors integrate geometry-based and biological features of the protein. An ART neural network algorithm is introduced to achieve dimensionality reduction, thus improving the overall performance of the system. In this work, a hierarchical strategy, using Boosted C4.5 algorithm, is applied for structural classification based on the SCOP (Structural Classification of Proteins) hierarchy. The SCOP database was used to evaluate the effectiveness of the multi-level approach of this system.