Three-dimensional Protein Shape Similarity Analysis Based On Hybrid Features

The analysis of protein similarity is a matter of concern in the bioinformatics field,since studying the protein similarity can help understand the protein structure-function relationship.To this aim,several methods have been proposed.But current protein similarity results are still not satisfactory.Here we presented a novel method for evaluating the similarity of 3D protein models based on hybrid features.The main contents are as follows:1.We introduce different structures of the protein and the method for obtaining 3D mesh models of proteins,and then we calculate some local surface descriptors of 3D protein models,such as curvature,shape index and significant geometric features.Lastly,we analyze the significance of these descriptors for the stability analysis of protein molecules,protein interaction mechanism and structure prediction.2.In order to capture the topological features of 3D protein model to compensate for the local features,we analyze some 3D skeleton extraction algorithms.We also provide an efficient and convenient method to extract the skeleton of 3D protein model,and calculate the local diameter of protein model based on the skeleton.The geodesic distance of the 3D protein model is calculated,and the extraction algorithm of the heat kernel signature is improved by using the geodesic distance to replace the Euclidean distance.The heat kernel signature is robust for the deformation and isometric invariance,and can retain the intrinsic geometric properties of the shape.3.It constructs a three order tensor descriptor for each 3D protein model based on the significant geometric feature(SGF),the local diameter(LD)and the heat kernel signature(HKS).The gray relational analysis method based on the matrix is extended to the three order tensor,and the similarity of 3D protein models is analyzed.The results of experiments show that our method has good adaptability and effectiveness for similarity analysis of 3D protein models with complex local surface features and different topological features.