Whole-proteome phylogeny of prokaryotes by feature frequency profiles: An alignment-free method with optimal feature resolution
Whole-proteome phylogeny of prokaryotes by feature frequency profiles: An alignment-free method with optimal feature resolution
复制标题
DOI：
10.1073/pnas.0913033107
复制
发表时间：
2010-01-05
影响因子：
11.1
通讯作者：
Kim, Sung-Hou
中科院分区：
文献类型：
--
作者：
Jun, Se-Ran;Sims, Gregory E.;Kim, Sung-Hou
我们提出了一种原核生物的全蛋白质组系统发育树，它是通过比较全蛋白质组的特征频率图谱（FFPs）构建的。特征是氨基酸的l - 聚体，每个生物体都由所有特征的频率图谱来表示。在FFP方法中，特征长度的选择至关重要，我们已经开发了一种程序来确定用于推断原核生物系统发育（严格来说是蛋白质组系统发育）的最佳特征长度。我们的FFP树是用884种原核生物、16种单细胞真核生物以及2条随机序列的全蛋白质组构建的。为了突出主要类群的分支顺序，我们展示了一个带有分支支持的单系类或门的简化蛋白质组FFP树。在我们的全蛋白质组FFP树中：（i）古菌、细菌、真核生物以及作为外类群的随机序列被清晰地分开；（ii）当以随机序列为根时，古菌和细菌形成姐妹群；（iii）具有细胞内膜区室的浮霉菌门被置于细菌域的基部位置；（iv）在大多数分类级别上，原核生物中几乎所有类群都是单系的，但在我们的蛋白质组FFP树和用其他方法构建的树之间，主要类群的分支顺序存在许多差异；（v）先前“未分类”的基因组可能被归入最有可能的分类单元。我们描述了我们的FFP树与基于其他方法的树在原核生物的分组和系统发育方面的显著相似之处和差异。
We present a whole-proteome phylogeny of prokaryotes constructed by comparing feature frequency profiles (FFPs) of whole proteomes. Features are l-mers of amino acids, and each organism is represented by a profile of frequencies of all features. The selection of feature length is critical in the FFP method, and we have developed a procedure for identifying the optimal feature lengths for inferring the phylogeny of prokaryotes, strictly speaking, a proteome phylogeny. Our FFP trees are constructed with whole proteomes of 884 prokaryotes, 16 unicellular eukaryotes, and 2 random sequences. To highlight the branching order of major groups, we present a simplified proteome FFP tree of monophyletic class or phylum with branch support. In our whole-proteome FFP trees (i) Archaea, Bacteria, Eukaryota, and a random sequence outgroup are clearly separated; (ii) Archaea and Bacteria form a sister groupwhen rooted with random sequences; (iii) Planctomycetes, which possesses an intracellular membrane compartment, is placed at the basal position of the Bacteria domain; (iv) almost all groups are monophyletic in prokaryotes at most taxonomic levels, but many differences in the branching order of major groups are observed between our proteome FFP tree and trees built with other methods; and (v) previously "unclassified" genomes may be assigned to the most likely taxa. We describe notable similarities and differences between our FFP trees and those based on other methods in grouping and phylogeny of prokaryotes.