The Construction And Application Of Protein-protein Interaction Database System

More and more data have been drawn from the studies on protein interactions as they gradually develop in depth. Therefore, a unified standard of protein-protein interaction databases is in urgent need for collecting, collating the existing data and mining for useful information from them. In this study, the database maintained by DIP is used as the source of basic data. According to the contents of study on protein interaction, the database is designed to consist of seven parts:protein, interactive protein, interaction, source of species, evidence of identification, references, and cross index. Based, on the development environment of Apache, MySQL, and CakePHP, our protein interaction database system is constructed. The main contents and results are as follows:(1) A set of data model of protein-protein interaction is designed and a protein interaction database is created, by studying the format of database data and referring to the data specification of biological database. The database covers 268 species,20683 protein records,60699 interaction records, and 3852 document records.(2) The protein interaction database system has achieved high-level retrieval, data statistics, data download, data submission and other functions. High-level retrieval can be performed by three retrieval methods, including protein, interaction, and reference. Data statistics can provide statistical information of database from different levels to help users to know the data range of the database. Data download module has the function for a unified download of current datum collection to meet the download need of special users. The system provides a platform for data submission to import the normally formatted protein-protein interaction data into the database.(3) A protein interaction prediction platform based on the similarity of protein is created by combining with a new editing program of BLAST, which can effectively find similar-sequenced proteins and information of interaction. (4) The retrieval function and analysis&prediction system are tested and validated by giving two specific examples, HSP70 and PRK2a, and practical results are obtained. The PRK2a example shows that the prediction system has the basic function to predict the protein-protein interaction, namely, after the sequence of a protein is imported, it can export a list of similar-sequenced proteins and proteins interacted with the protein. This study could promote the study on protein interaction.