AryNet-以微陣列數據進行基因網絡視覺化的方式來比較化學物質與精神病關聯性之網路應用系統

Abstract

生物晶片研究因為擁有再現性、樣本重複性，可藉由大量數據彙整進行統計分析來進行基因體交互作用相關研究。現今關於生物晶片數據分析與視覺化的平台，除市面上付費軟體外，網路上開源相關軟體套件和網站系統亦有提供一些免費資源，但這些免費平台大多容易因為功能限制而造成不便。 本研究針對幾種生物晶片平台建立一個MVC模式資訊整合資料庫--AryNet，資料庫收集來自Reactome的基因轉譯後蛋白質交互作用資訊，以及自GEO下載的常見精神疾病、神經退化疾病以及環境化學物質的相關晶片數據，特別是「基因表現」、「核糖核酸甲基化」兩種類型的晶片原始數據。資料庫後端透過R-Engine配置多種Bioconductor所提供的數據標準化、統計校正演算法功能。系統將這些演算法相關參數以JavaScript方式呈現於網頁上供使用者自由調整，並以互動式基因網絡圖形將分析結果即時回傳使用者。 我們使用AryNet收錄資料庫統計躁鬱症、思覺失調症、重度憂鬱症等疾病的可能相關基因，然後結合基因調控網絡資訊計算這些疾病與幾種環境化學物質造成相似外表型的可能相關性。

As the Internet develops, microarray data shows a powerful potential for detecting the novel genomic interaction with its powerful repeatability and reproducibility. There are a lot of web tools and open source packages for analysis and visualization of microarray data. But most of them are costly or too technical to use. For the purpose of developing a tool to operate data processing, statistics, genomic comparison, and visualization on microarray data in a friendly usage. We built a web system named AryNet. Applying the MVC framework with an interactive controlling panel on web page, AryNet possesses a SQL database storing epigenetic and gene expression profiles of microarray data downloaded from GEO database including samples with mental diseases, neural diseases, and chemical exposure. The information of protein-protein interactions from Reactome is also installed. The Java-based controller was armed with a plugin named R-Engine to drive the R-package of data processing and statistics obtain from Bioconductor. The resulting analysis will be retrieved back to the user view and generates a gene networks diagram on time. We obtain the differential expression genes (DEGs) profiles from bipolar disorder, schizophrenia, major depression and chemical exposure by AryNet. And then we generated the gene network diagrams combining the DEGs with genes which have relationships in protein-protein interaction of each profiles. Comparison with the gene networks shows that there might be some resemblance between the phenotypes of diseases and endocrine disruptors.