比較老鼠心肌組織施用醛固酮與否之基因調控網路

Abstract

RNA序列定序所得的基因表現量，標準分析程序是求得差異顯著基因並對顯著基因進行基因分類分析。近年來有越來越多結合網路的分析方式，像是繪製基因的相關性網路，尋找網路中的團塊化基因等等。但這些方法大多都都只能概括性的描述網路特徵，很難進行統計上的檢定，也不太能進行不同處理下網路的差異比較。因此我們以比較兩種不同處理下網路的變化為目標，並且跳脫以往以節點為出發點的分析方式，探討不同情況下顯著基因之間連線上的改變，以達到偵測哪些基因再施打藥劑之後產生了擾動現象。找出在不同形況之下有所擾動的基因後，我們也進行基因分類分析，並且結合了較常運用在社會科學的雙模式網路分析法，視覺化的描述基因分類之間的往來是來自哪些基因。本篇研究建立系統化的網路比較流程，並且結合雙模式網路分析法，說明基因分類之間的關係，期許能替將來相關研究帶來一點貢獻。

The gene expression levels estimated from RNA sequencing can be subject to statistical analyses in finding the differentially expressed genes and the gene ontology terms. In recent years, there have been more and more analytical methods that incorporate networks, such as construction of gene association networks, searching for gene modules in networks, and so on. However, most of these methods can only describe the characteristics of the network in a general way. It is difficult to carry out statistical hypotheses tests, and it is not possible to compare the differences between the networks under different conditions. In this study, we aim to compare the networks under two different treatments. Instead of conventional node-based analysis, we propose to explore changes in the connections between significant genes in different situations so as to detect which genes are being disturbed. After identifying genes that are perturbed under different conditions, we also perform GO analysis and adopt the two-mode network method, which is commonly used in the social sciences, to visualize the genes that belong to more than one gene classifications. This study has established a systematic protocol for network comparison. Combined with the two-mode network analysis method, the protocol can assist to explain the relationship between GO terms. The results may contribute to related research in the future.