Abstract
Structure prediction tools for protein sequences benefit the biologists to search for the real protein interactive function sites. In the low sequence alignment similarity cases, increasing the accuracy of the structure prediction tools have become an important research issue in bio-informatics. In this thesis, we propose a new novel sequence matching tool using the secondary structure information of the protein in the similar families to predict the function sites of the unknown target. The algorithm of the matching tool is cutting the protein secondary structure, rearranging the topology, and then computing the best matching pairs. We take starch-binding domain (SBD) in the carbohydrate-binding module (CBM) protein family to be verified by matching tool. We also compare these tests with other tools, and our results correspond with the final biology experiments.