2009-08-012024-05-16https://scholars.lib.ntu.edu.tw/handle/123456789/669487摘要：細胞的生命現象是綜合各種動態基因調控機制造成的結果。因此，建立細胞內各 基因的調控網絡，有助於了解外來刺激對細胞造成的影響。近年發展出的微陣列技 術能同時觀測數以千計的基因表現，進階提供了系統化了解基因調控網路的機會， 也有許多透過微陣列基因表現資料推論基因調控網路的演算法（或稱逆向工程演算 法）被相繼提出。然而，進行文獻回顧時，不難發現各演算法結果不全一致。不一 致的結果可能肇因於：(1) 各演算法在搜尋不同形式的基因調控效果各有優劣，及 (2) 用作比較基準的模擬或實驗資料不足以代表基因表現的所有真實情況。本計畫 的目的即是在探討上述兩者是否為造成各演算法結果不一致的因素，研究目的又可 細分為二。首先，我們將彙整並比較用於偵測在網絡中具有活動力基因的統計方法, 並探討這些統計方法用於區分基因間直接或間接關連性的可能性。再來，我們將透 過不同方式模擬並收集可公開取得的微陣列基因表現資料，以了解不同統計方法是 否在不同型態的資料有差異性的分析結果。<br> Abstract: Activities of living cells are consequences of dynamic gene regulatory events. The elucidation of gene regulatory network is thus important to divine the cell’s behavior in response to perturbations. About a decade ago, microarray technology were developed to simultaneously monitor expressions of thousands of genes and opened a window of opportunity to conceptualize the cell’s regulatory machinery. Efforts to infer, or reverse engineer, gene regulatory network based on microarray expression data have been made constantly ever since. However, through literature review, it can be observed that different types of reverse-engineering algorithms yield inconsistent results. The inconsistency may be due to (1) different methods may have their own strength and weakness in application to identify different types of association, and (2) lack of realistic set of benchmark data to assess the performance of the methods. Hence, the objectives of this research have twofold. Firstly, we will review and compare statistical methods that can be used to identify involved genes and their associations in the network. Secondly, we will make a few attempts to simulate nearly realistic microarray expression data on which the reverse-engineering algorithms will be exercised. Several public-available experimental data will be tested, too.基因調控網絡微陣列逆向工程演算法gene regulatory networkmicroarrayreverse-engineering algorithms