Design and realization of a microarray data analysis platform / 南方医科大学学报

<p><b>OBJECTIVE</b>To design a platform for microarray data analysis and processing in the browser/server mode running in Linux operating system.</p><p><b>METHODS</b>Based on the Apache HTTP server, the platform, programmed with Perl language, integrated R language and Bioconductor packages for processing and analysis of the input data of oligonucleotide arrays and two-color spotted arrays. Users were allowed to submit data and parameter configurations to the platform via the web page, and the results of analysis were also returned via the web page.</p><p><b>RESULTS</b>With an easy operation and high performance, the platform fulfilled the functions of processing, quality assessment, biological annotation and statistical analysis of the data from oligonucleotide arrays and two-color spotted arrays. Using the platform, we analyzed the gene expression profiles in Mtb-stimulated macrophages of three clinical phenotypes, namely latent TB (LTB), pulmonary (PTB) and meningeal (TBM), and obtained valuable clues for identifying tuberculosis susceptibility genes. We also analyzed the effect of INH treatment on Mycobacterium tuberculosis gene expression in various dormancy models, such as hypoxia and KatG mutant, and found that a set of genes responded to INH treatment during exponential growth but not in dormancy, and that the overall number of differentially regulated genes was reduced in the cells in low metabolic state.</p><p><b>CONCLUSION</b>The platform we have constructed integrates comprehensive resources, and with a user-friendly interface, allows direct result visualization to facilitate microarray data analysis.</p>

Design and construction of the platform for comparative genomics / 南方医科大学学报

<p><b>OBJECTIVE</b>To design a versatile genome comparison and visualization platform based on browser/server mode supported by a local server.</p><p><b>METHODS</b>The server of the platform was Apache HTTP server. Perl was used to integrate such genome alignment package and algorithms as MUMmer, LAGAN, and Mauve for different comparison purposes, and the users could submit data and parameters to the platform via the web page. The results of analysis were also returned via the web page.</p><p><b>RESULTS</b>The platform could handle multiple data input formats, compare complete and draft genome sequence, alignment pair-wise or multi genome of more divergent species, identify regions of high similarity, locate local nucleotide mutations and large-scale recombination, and display the results by visualization interface. Analysis of the homology of 10 new strains of influenza A virus indicated that PB1 gene might evolve from human H3N2 viruses, PB2 and PA genes from avian H3N2 viruses, and HA and NS genes from swine H1N1 viruses. Alignment of Mycobacterium tuberculosis (H37Rv, CDC1551) and Mycobacterium bovis (AF2122/97) genomes revealed that sequence insertion/deletion and duplication were the major source of genomic differences.</p><p><b>CONCLUSION</b>The platform integrate comprehensive resources with a user-friendly interface and intuitive result visualization to facilitate conventional study of comparative genomics.</p>