| Human genome project(HGP) promotes the research of post genome project and function genomics. In order to study the function of thousands of genes, especially relationship among genes in expression and regulation. Obviously, traditional research methods are low efficient and unable to fulfill the require of genome and function genomics. In this situation, microarray technology has arisen, which has been widely used in biology and medicine. Along with the development of this technology, data is constantly increasing, and researchers pay attention to how to deal effectively with the enormous data. The analysis of microarray need new algorithms, software and platform. This paper introduces microarray analysis tools and point out the existing problems. For example, some tools are used for special microarray; some are used only in certain operating system and have high demand in hardware; some are commercial and expensive; some need lots of parameters; some use the command line or has no graphical interface. Therefore, some biologists need to spend lots of time to study and using a number of software can get the ideal results.Along with the development of microarray technology, data increase. How to deal and manage the data? Microarray analysis need new algorithm, software and compute platform. First, this article analyze the current situation, and summarize problems. In view of the above problems we developed a microarray data analysis platform, which is user-friendly interface, full-featured and convenient for biologist. The platform using Browser/Server network structure, users can submit their oligonucleotide array data, cDNA array data and tissue microarray data, and parameters to platform server through the web Browser, then the server analysis the submitted data. After the analysis, results will return to user by Browser. The platform server is a PowerCluster8000IN computer with Linux operating system, network server using Apache HTTP, data management using MySQL, development language using Perl which integrates R language and Bioconductor, and HTML is used to design websites. The final results output in the form of table, text or graphics.R language is a computer programming language and open development platform, which has robust statistical analysis and some images display tools. Bioconductor is predicated on compute biology and bioinformatics, which provides lots of statistic methods and software packages. So we integrated several software packages using perl language.The main functions of platform are:1. The analysis of oligonucleotide arrays: read raw data, processing, quality assessment, analysis of different expressed genes and biological annotation.2. The analysis of cDNA arrays: processing, quality assessment and statistical analysis.3. The analysis of tissue microarray:classification using random forest algorithom.Through the platform, we analysis gene expression profiles in Mtb-stimulated macrophages with three clinical phenotypes, which are latent TB(LTB), pulmonary(PTB) and meningeal(TBM).It is important to identify tuberculosis susceptibility genes. We analysis the effect of INH treatment on Mycobacterium tuberculosis gene expression in various dormancy models, such as hypoxia and KatG mutant. The conclusions are that the set of genes that respond to INH treatment during Log phase growth would not be differentially regulated during INH treatment in the dormancy models, and that the overall number of differentially regulated genes would be reduced do to the low metabolic state of the cells. The application proves that the platform is very effective.The platform also has some shortcomings:microarray analysis has a lot of software and algorithm, but we only integrate some software packages in R and Bioconductor. Furthermore, we didn't use parallel computing. Our research can provide convenience for biologists who devote to microarray analysis. In addition, the platform could provide reference for designing and building other platforms. |
PDF Full Text Request