克雷伯氏肺炎桿菌基因體序列之親緣關係分析

Abstract

在台灣，克雷伯氏肺炎桿菌常出現在糖尿病患者，進而引起肺炎、敗血症、肝膿瘍等等併發症。莢膜是主要的致病因子。莢膜多醣體屬於水溶性，偏酸性，莢膜多醣體之組成被發現有具有極高的變異度。 本篇論文希望能藉由比較10種菌株(NTUH-2044、K1、K2、K5、K9、K14、K52、NK8、NK29、NK245)的莢膜多醣體，並觀察10株中在哪些基因片段相似度較高；之後藉由不同演算法來建立演化樹，並且利用Robinson-foulds distance來找出在三種演算法中，何種演算法建立出來的為『相對最佳演化樹』。

由多重序列比較的結果來看，各序列在galf、wzi、gnd、和ugd的基因區段相似度較高。而在莢膜多醣體去除5’和3’端的情況下，普遍相似度不高。但在wzc和wcal區段有不錯的相似度。在計算演化樹的距離上，最大簡約法在未除去5’和3’跟已除去5’和3’狀態下中RF距離的數值是相對最小的，所以其產生的演化樹在三種演算法下是相對最佳演化樹。

In Taiwan, Klebsiella pneumonia is prominent in people who have diabetes and causes complications including pneumonia, sepsis, liver abscess, etc. Capsular is the main etiological factor. Also, capsular polysaccharide belongs to water-soluble and meta-acidic and its compositions are found to have highly variability. In this thesis, we want to find out in which algorithms the local optimal phylogentic tree is constructed by comparing ten strains of Klebsiella pneumonia of capsular polysaccharides, including NTUH-2044、K1、K2、K5、K9、K14、K52、NK8、NK29、NK245, observing similarities of genes among ten strains, constructing phylogenetic trees by various algorithms and then using the Robinson-foulds distance. According to the result of comparative sequence analysis overview, sequences in galf, wzi, gnd and ugd regions have higher similarities. However, the similarity in cps without 5' and 3' is low in general but in wzc and wcal region are higher. When calculating the distance of phylogenetic trees, the RF distance is relatively low under the situation of maximum parsimony method with whole cps and without 5' and 3'; therefore, the phylogenetic trees constructed by them are the local optimal phylogenetic tree among three algorithms.