Figure 1.
Structure of the mitochondrial genomes ofP. trewavasae and T. moorii.
As the differences in genomic structures of both species are too minor to be resolved in this view, only one representative figure is shown. Direction of arrows denotes the strand on which a particular feature resides (H-strand clockwise); staggered arrows indicate an overlap of neighbouring features (for details see Table 1). The size of the genome is 16,588 bp for P. trewavasae and 16,590 bp for T. moorii, where both contain the known 13 protein-coding genes of the respiratory chain (ND1, ND2, COX1, COX2, ATP8, ATP6, COX3, ND3, ND4L, ND4, ND5, ND6 and CYTB), 22 interspersed transfer RNA genes, 2 ribosomal RNA genes (12S and 16S rRNA) and the non-coding control region (D-loop region). The marked position indicates the mutation in the ND4 gene of P. trewavasae (causing a frame shift), and the location of the almost immediately following induced stop codon.
Table 1.
Organization of the mitochondrial genome ofP. trewavasae/T. moorii.
Table 2.
Variable regions and conserved blocks in the non-coding D-loop region.
Figure 2.
Phylogenetic relationships among labroid families analyzed in this study.
Shown is a representative cladogram based on all sequences except those of tRNAs (i.e., data set #3), with numbers representing RAxML bootstrap values above and Bayesian posterior probabilities below the branches. Numbers at internal nodes relate to entries in the Supplementary Table S6 in File S1, where all calculated support values are given. The tree was rooted with the sequences of two outgroup species: Bajacalifornia megalops and Alepocephalus agassizii.
Figure 3.
Relative rate of molecular evolution.
Bars represent the coefficients of linear least squares regression, where regressions of the pairwise distances of all genes were calculated against the distances of the 12S rRNAs (for an example see Supplementary Figure S1 in File S1). Horizontal lines indicate group mean values (from left to right: 3.01, 2.76, 0.29, 0.68, 1.25, 1.12, 0.75). D-loop (4.38) represents the full non-coding region, whereas D-loop Gb (1.97) is based on a Gblocks-filtered multiple sequence alignment (InDels removed). For individual bar heights see Table S4A in File S1.
Figure 4.
Shown is the codon distribution of all merged protein-coding genes for all considered species (for gene-wise analysis results see Supplementary Figure S6 in File S1). Color key: The green region around one implies that a codon was observed about 1 time in 60 codons, higher red and lower blue values indicate deviations of uniform distribution (as factors); thereby also amino acid distribution is implicitly visible. A Hierarchical clustering (Lance-Williams [85]; average linkage method) of codon patterns (y-axis) approximately groups together the examined families. Closely related species (i.e. from the same tribe) generally have highly similar codon usage. Clustering on codon frequencies (x-axis) facilitates the visual identification of deviations. B Synonymous codons are listed sequentially to allow for quick evaluation of relative synonymous codon usage. For instance, the first 4 columns depict the codon usage for Alanine, where obviously the codon GCG is hardly used and codons GCU, GCC and GCA are used by all species with a preference for GCC.
Figure 5.
C-terminal end variation in Cox1.
A C-terminal amino acid sequences of Cox1 exhibit noticeable variations compared to the rather minor differences in all other mitochondrial genes examined; different variant groups can be identified. B Amino acid accessibility scale profiles indicate the potential for functional interaction sites in this region (for comparison see examples for full-length Cox1 amino acid accessibility in Supplementary Figure S4 in File S1). Position values are related to residue positions in the alignment, scores are normalized (0,1) with a higher score indicating higher accessibility.