A rapid 3D protein Structural Classification using Pseudo 2D HMMs

In this paper, we introduce a rapid method to classify protein structures. Instead of using the 3D structures of proteins directly, we used 2D inner-distance matrices converted from the 3D protein structures and extracted feature vectors from the 2D inner-distance matrices. We divided the matrices into a training dataset and a testing dataset. We then employed a 2D Hidden Markov Model (2DHMM) to classify the protein structures. First, we obtained a 2DHMM protein structure template from training the dataset of the 2D inner-distance matrices. Second, we applied the 2DHMM template to the testing dataset. The results show that our approach obtains a competitive performance to other approaches.