Current Opinion in Structura...

May 8, 2015 news / BIOON / -- with the rapid development of sequencing technology, a large number of genes have been revealed and a large number of protein sequences have been analyzed However, the sequence information has not been well annotated For example, less than 1% of protein sequences have known functions If we can use the method of computational biology to annotate these sequence information, we can accumulate a lot of knowledge about protein structure, and finally achieve the leap from quantitative change to qualitative change Scientists from Howard medical center in New York discussed a new method of annotation structure based on protein function In the past, the existing prediction methods are based on sequence homology, evolutionary relationship and position in the genome to predict the corresponding template and its function However, for 3D structure information, few of them are used in these predictions But that is changing Thanks to the increasingly mature and powerful homologous modeling technology, many unknown protein structures can be predicted with high reliability The update of function prediction based on template is based on such homologous modeling We can predict the structure of the protein sequence, and then find the corresponding template based on the similarity of the structure, so we can infer that this sequence may have similar functions with the template protein So, how do we define structural similarity? In addition to the similarity of secondary and tertiary structures of proteins, there are many other aspects of similarity to be considered For example, many proteins have structural cavities or pockets, which may be conservative In addition, some chemical properties, such as the electrostatic potential of protein surface, should also be fully considered In order to find the template of sequence in database, we must first compare the similarity of sequence, the global structure of protein and the local structure construction matrix Specifically, the modeling structure of the query sequence is overlapped with the protein in the protein database, and then the difference and similarity of the local structure and sequence are considered The functional similarity considered here mainly refers to the similar interaction characteristics For example, the queried proteins and templates can combine with the same type of molecules The authors also discussed how to improve the similarity of protein global structure and local region Then they also described the significance of protein family and template family in the exploration of protein similarity For example, machine learning model can be built to find different structure and physical and chemical properties of different protein template families Finally, the authors conclude that there are two main functions of protein sequence annotation based on structure First, structural similarity may mean similar protein binding properties Secondly, the similarity prediction based on structure can find more information, which can not be done by sequence alignment The method of computational biology can discover more information, which is of great significance for the future biological research and drug development This article is the original compilation of Biovalley, welcome to reprint! Please indicate the source of the reprint and attach the original link For more information, please download Biovalley app