Study On The Key Technologies Of Building Virus Genome Bioinformation Analysis System

Along with the completion of human genome project and other genome projects, the key research of bioinformatics has transferred from collecting data to processing data. Then the research on bioinformation analysis system and data mining is becoming more important. However, due to the complexity of both bioinformation data and their applications, there has been no relatively common framework model so far to meet the development need of general bioinformation analysis system. The evolution of virus genome is much faster than other species, and it springs up bioinformation every day. So virus genome bioinformation analysis system needs to think about the problems of data updating and mining. This dissertation focuses on the virus genome bioinformation analysis system, and it gives a kind of bioinformation Multi-tier architecture system model (BIOCMSM) in common use, talks about the WEB-Based automatic bioinformation acquisition and density-based K-medoids clustering analysis. At last, it builds the Newcastle Disease Virus (NDV) bioinformation analysis system.The main research contents are as follows:1. The research of bioinformation analysis system model based on Multi-tier architecture. Due to the problems of data management, integration and application in the bioinformation analysis system, this dissertation firstly sums up the general process, and then gives a kind of bioinformation Multi-tier architecture system model (BIOCMSM) in common use. This model adds the data process layer to solve the data converting, processing, integrating and updating based on the common architecture model.2. The research of WEB-Based automatic bioinformation acquisition. Due to the problem of bioinformation automatic acquiring and updating, this dissertation gives a practicable scheme based on agent program, and then describting its implementation procedure in detail. Experiments have proved that this scheme could better resolve the problems of NDV bioinformation data updating. What is more important is that the scheme has better generality, it is not only limited to bioinformation updating.3. The research of density-based K-medoids clustering analysis. Due to the problem of clustering analysis in the nucleotide sequences, a new density-based K-medoids mothed is described in this dissertation, and it is applied into clustering analysis of NDV genes. Experiments have proved that this method has better initialization, less iterative times and satisfying results compared with the ordinary K-medoids mothed.