微陣列基因體比較雜合法之統計方法整合

Abstract

微陣列實驗技術是一個具有能夠同時偵測數千甚至上萬基因表現的特性且被廣泛的運用於生物領域上。微陣列基因體比較雜合法則是一種用來偵測在單一實驗中，DNA序列拷貝數目改變的技術。技術的原理是透過比較經由螢光標定後的螢光強度比值，藉以偵測大片段基因區域改變的幅度，並將這些變異的程度藉由圖形對應的方式表現出來。在統計方法的應用上，是為了要能夠有效的辨別出染色體的擴增與缺失現象。因此，將焦點置於分析染色體的擴增與缺失現象。沿用Lai et al. (2005)文中所模擬與比較的過程，透過整合既有的統計方析方法，如循環二元分割法、適應性權重平滑法與CGH分割法等方法。並使用PERL、PHP與Apache等網頁服務程式，開發出一套以R語言為後端演算的統計分析平台，提供資料的正規化、統計分析與染色體區域擴增與缺失圖形，最後藉由UCSC Genome Browser與ID Converter的註解，供使用者針對感興趣的生物問題來做探討。

The DNA microarray is widely used to investigate gene expression profiles of many thousands of genes simultaneously. And it has become a common tool for exploring various questions in many areas of biological and medical sciences. Specifically, array-based comparative genomic hybridization (Array CGH) is applied to screen alteration of DNA copy numbers genomewide. The main purpose of such application is to detect the altered DNA segments among genome sequences from a control (reference) treatment to a test treatment. Typically, efficient statistical tools are developed to compare the intensity ratios of spots representing the competitive hybridization between the control mRNA sample and the test mRNA sample, which are separately labeled with red (Cy5) and green (Cy3) fluorescence dyes. Users usually focus on the gain region and the loss region on each chromosome. In consequence, the differentially altered regions are displayed by graphical plots. From the simulation results presented in Lai et al. (2005), several competing statistical methods are selected for analysis of Array CGH data, including Adaptive Weights Smoothing method, Circular Binary Segmentation method and CGH Segmentation method. Furthermore, we use Perl, PHP programming language and Apache web server to integrate the chosen statistical methods into an analysis platform under R language environment. The proposed platform offers normalization, identification of the differentially altered regions and plotting of the gain and loss regions genomewide. In addition, users can annotate information through UCSC Genome Browser and ID Converter for advanced analyses.