On Classification of PDZ Domains: A Computational Study

Abstract

Our goal in this present study is to introduce new wavelet based methods for differentiating and classifying Class I and Class II PDZ domains and compare the resulting signals. PDZ domains represent one of the most common protein homology regions playing key roles in several diseases. To perform the classification, we developed two methods. The first of our methods was comparable to the standard wavelet approaches while the second one surpasses it in recognition accuracy. Our models exhibited interesting results, and we anticipate that it can be used as a computational technique to screen out the misfit candidates and to reduce the search space, while achieving high classification and accuracy.