Intelligent Protein 3D Structure Retrieval System

Since the 3D structure of a protein determines its function, the protein structural identification and comparison system is very important to biologists. In this paper, an intelligent protein 3D structure retrieval system is described. The system is intelligent since it integrates the moment feature extraction technology and the relevant feedback method in Artificial Intelligence (AI). As there is no universal agreement on the similarity of proteins structures, the major advantage of our system compared to other previous systems is that we use the relevance feedback technology to aid the biologists to find the similar protein structures more effectively. The similarity metric formula is improved dynamically by biologists’ interaction through relevance feedback. The experimental results show that the proposed approach can capture the biologists’ intentions in real-time and obtain good performance in the protein 3D structure retrieval. The ratio of total improvement is about 15.5% on average, which is quite significant compared to the improvements obtained in some previous work.