贝叶斯推论重建乳头瘤病毒的进化系统

基于90种乳头瘤病毒的L1基因和E2基因的DNA全序列，应用似然比检验和贝叶斯推论重建病毒的系统发育树，并在此基础上讨论病毒分子系统学研究的一般原则和方法。同时比较研究贝叶斯推论和过去常用的最小进化、最大简约、最大似然三种建树算法的优劣。结果表明：乳头瘤病毒可分为4超群，不同超群的乳头瘤病毒宿主和组织特异性不同。引起人宫颈癌的高危和低危乳头瘤病毒在进化上分别具有相近的亲缘关系。高危类群起源于共同祖先。算法的比较研究表明，贝叶斯推论作为进化生物学的最新进展，在计算速度和可靠性上明显优于其它算法。

Resolution of the Evolutionary Relationship of Papillomaviruses Using Bayesian Phylogenetics

Our aim is to analyze the phylogenetic relationships of 90 r epresentative papillomaviruses(PVs )on the basis of complete sequence s of the L1 and E2 genes.We employ ed likelihood ratio tests and Baye sian inference to reconstruct the ph ylogenetic trees of these PVs and e stimated the trees by Bayesian pos terior probability and bootstrap a nalyses.The results of this study resolved the 90 PVs into four supe rgroups and clinical studies showe d these 4 supergroups had different host ranges and tissue t ropisms.It also showed there was a close kinship within high-risk HP Vs group and low-risk HPVs group. All the PVs of high-risk HPVs group had a common evolutionary origin.The s tudy on methods of reconstructing tr ees showed that Bayesian inference is more advanced and robust than m aximum likelihood,maximum parsimon y and minimum evolution.