18S rDNA sequence-structure phylogeny of the eukaryotes simultaneously inferred from sequences and their individual secondary structures

Abstract

Objective: The eukaryotic tree of life has been subject of numerous studies ever since the nineteenth century, with more supergroups and their sister relations being decoded in the last years. In this study, we reconstructed the phylogeny of eukaryotes using complete 18S rDNA sequences and their individual secondary structures simultaneously. After the sequence-structure data was encoded, it was automatically aligned and analyzed using sequence-only as well as sequence-structure approaches. We present overall neighbor-joining trees of 211 eukaryotes as well as the respective profile neighbor-joining trees, which helped to resolve the basal branching pattern. A manually chosen subset was further inspected using neighbor-joining, maximum parsimony, and maximum likelihood analyses. Additionally, the 75 and 100 percent consensus structures of the subset were predicted.

Results: All sequence-structure approaches show improvements compared to the respective sequence-only approaches: the average bootstrap support per node of the sequence-structure profile neighbor-joining analyses with 90.3, was higher than the average bootstrap support of the sequence-only profile neighbor-joining analysis with 73.9. Also, the subset analyses using sequence-structure data were better supported. Furthermore, more subgroups of the supergroups were recovered as monophyletic and sister group relations were much more comparable to results as obtained by multi-marker analyses.

Keywords: 18S rDNA; Eukaryotes; Phylogenetics; Secondary structure.

Fig. 1

Left: Encoding of sequence-structure information. Scoring matrices and substitution models have been adapted accordingly. The figure shows an RNA sequence with its individual secondary structure in the bracket-dot-bracket notation. The respective 2D structure, the 12-letter translation table as well as the one-letter-encoded pseudoprotein sequence are depicted. Right: Different alignments are shown. They differ in terms of informational content (exemplarily highlighted in red). Only the sequence-structure-alignments as derived from 4SALE [–22] include information about individual secondary structures whereas the guided-sequence alignment is guided only by a consensus structure

Fig. 2

Overall sequence-structure NJ tree using the 18S rDNA of all 211 taxa. ClustalW [19], as implemented in 4SALE [20, 21], was used for the global multiple sequence-structure alignment. The tree was reconstructed using ProfDistS [25, 26] and midpoint rooted. The scale bar shows evolutionary distances. Taxa names are accompanied by their corresponding GenBank accession number. Clades and respective singular taxa are marked in a color-scheme based on the eukaryotic tree of life published by Keeling and Burki [2]. If clades and singular taxa do not form one monophyletic group, they are numbered consecutively. If a group is only represented by one taxon, the taxon is marked in red. Taxa which were manually selected for the subsampling are marked bold. Supergroups are indicated according to Burki et al. [3] and Keeling and Burki [2], the names of the supergroups are adapted based on Adl. et al. [27]. With regards to readability the supergroups Amorphea, Obazoa and Opisthokonta are only named once near the biggest monophyletic subgroup. The three supergroups are marked with quotation marks since they are not monophyletic. The supergroup Opisthokonta includes Fungi, Metazoa, Choanoflagellata and Ichthyosporea. Amoebozoa are classified as Obazoa

Fig. 3

Two-times iterated sequence-structure PNJ tree with BS values (A) and original BL (B). Profiles were predefined according to Fig. 2; singletons were not included. The scale bar shows evolutionary distances. The trees were reconstructed using ProfDistS [25, 26] and rooted according to the overall sequence-structure NJ tree (Fig. 2). In each iteration, super-profiles of profiles have been built based on BS values (> 75). At internodes, the BS values from 100 pseudo-replicates have been mapped. The numbers in the triangles in front of the taxa represent the quantity of taxa included in the profile