Fig 1.
(A) A. ricciae individual in phase contrast, showing anterior–posterior axis and trophi (jaws). This genus does not possess the ciliated wheel-like organs (corona) on the head that distinguish other rotifers. (B) R. macrura showing eyes, extended corona used for feeding and locomotion, and a foot used for substrate attachment. (C) R. magnacalcarata with embryo developing internally. Scale bars indicate 50 μm; note: R. macrura is approximately the same size as R. magnacalcarata. Panels A and C courtesy of C. G. Wilson; panel B courtesy of M. Plewka (www.plingfactory.de).
Table 1.
Genome assembly and annotation metrics.
Fig 2.
Genome properties of bdelloid rotifers.
(A) Assembly contiguity. Cumulative assembly span for A. ricciae (red), R. macrura (blue), and R. magnacalcarata (green) reference assemblies, compared to the published sequence of A. vaga (orange). Scaffolds are arranged in descending length order along the x-axis, with cumulative span plotted along the y-axis. More contiguous assemblies achieve their total span with a smaller number of scaffolds, represented by a steeper line with a smaller tail. (B) Genome content. Proportion of each reference assembly covered by exons, introns, and identified repeat elements based on known metazoan TEs only (left-hand column) and TEs plus unclassified repeats detected using RepeatModeler (right-hand column). This shows that the inclusion of ab initio repeats results in substantially greater repeat content in all species, particularly in Rotaria. (C) Pseudogene detection. Distribution of percent identity for TBLASTN alignments (E-value ≤1 × 10−20) of predicted proteins to their own genome, discounting hits that overlapped with any existing predicted gene model. Only hits with a query coverage ≥95% are plotted. Vertical coloured bars indicate median values. (D) Intron length distributions for predicted genes (species coloured as previously). The inset shows detail of the upper tail of the main distribution (black box, truncated at ≥2.0). Vertical bars indicate median values for each species (of truncated distributions). Note log10 scale on y-axis. LINE, long interspersed element; LTR, long terminal repeat; SINE, short interspersed element; TE, transposable element.
Fig 3.
Collinearity in bdelloid genomes.
(A) Collinearity versus synonymous divergence. Points represent collinear blocks of genes, plotted based on the average pairwise KS between pairs of genes within the block (x-axis) and a ‘collinearity index’, defined as the number of collinear genes divided by the total number of genes within the genomic boundaries of that block (y-axis). Genes within Adineta species are clearly differentiated into 2 groups: homologs (low KS and high collinearity; open squares) and ohnologs (high KS and low collinearity; filled circles). In Rotaria species, however, genes within identified collinear blocks only show high KS and low collinearity, equivalent to that observed between ohnologs in Adineta. Note the different x-axis limit for A. ricciae, reflecting a higher synonymous divergence between homologs in this species. (B) Synonymous versus nonsynonymous divergence per collinear block. Genes within homologous blocks (open squares) show a low rate of synonymous (KS) and nonsynonymous (KA) substitution. Homologous gene copies are not resolved in Rotaria assemblies. Genes within ohnologous blocks (filled circles, found in all species) show relatively higher rates of both KS and KA. Right-hand panel shows elevated mean KS in both A. ricciae homologs (0.14 ± 0.036 [SD] versus A. vaga = 0.05 ± 0.026) and ohnologs (1.27 ± 0.146 versus A. vaga = 0.61 ± 0.062; R. macrura = 0.70 ± 0.080; and R. magnacalcarata = 0.70 ± 0.078). This elevation is not observed for KA (lower panel). Box-plots span the median (thick line), 50% of the values (box), and 95% of the values (whiskers).
Fig 4.
Distributions of MAF and read coverage.
(A) Folded MAF spectra for detected SNPs in the reference genomes of A. ricciae (red), A. vaga (orange), R. macrura (blue), and R. magnacalcarata (green) are distributed around a mode of 0.5 in all species. (B) Read and SNP coverage distributions. In each plot, the bar histogram represents the distribution of read coverage at SNP sites only, while the overlayed line shows the distribution of read coverage across all sites in the genome. The y-axes indicate proportion sites with given depth of each category (i.e., peak heights are relative to each category). The number of SNPs contributing to the bar histogram is indicated in parentheses (see legend). The cause of the secondary peak in SNP depth (at approximately 125x) for A. vaga (library ERR321927) is unknown. MAF, minor allele frequency; SNP, single-nucleotide polymorphism.
Fig 5.
Estimates of intragenomic divergence.
(A) Distribution of sequence identity (expressed as a divergence, i.e., proportion of nonidentical sites) for the top non-self BLAST hits from intragenomic comparisons for each species, showing highly similar gene copies within Adineta genomes but not within Rotaria genomes. Median values are 4.55% (A. ricciae, red), 1.42% (A. vaga, orange), 22.5% (R. macrura, blue), and 21.7% (R. magnacalcarata, green). (B) Homologous divergence estimated from SNP counts in coding regions in REF and MAXHAP assemblies for R. macrura and R. magnacalcarata. Mean estimates are 0.026% for R. macrura and 0.104% for R. magnacalcarata. MAXHAP, maximum haplotype; MP, mate pair; PE, paired end; REF, reference; SNP, single-nucleotide polymorphism.
Fig 6.
Unusual genomic features are not detected in A. ricciae.
Genomic scaffolds are shown as grey bars, GC content (green line), and read coverage (grey histogram) averaged over 5-kb bins are shown above each scaffold. Black lines within scaffolds represent gaps (Ns) introduced during scaffolding. (A) Example of collinearity breaks in A. vaga 2013 assembly scaffold AVAG00001. Links between homologous blocks are shown in light blue; downstream connections are shown in grey. Collinearity breaks are shown in dark blue. (B) The majority of homologous blocks encoded on the same scaffold in the A. vaga 2013 assembly are palindromes (dark red); tandem repeats are shown in pink. (C) Example of a single collinearity break on A. ricciae scaffold ARIC00001. (D) No genomic palindromes detected in A. ricciae. Only 2 homologous blocks were found on the same scaffold (ARIC00305 and ARIC00659), both arranged as tandem repeats. GC, guanine-cytosine.
Fig 7.
Bdelloid genome characteristics compared to other metazoans.
Phylogenetic relationships among 17 protostome species are shown on the left, major groups are marked in grey. (A) Gene count, defined as the number of protein-coding genes in each genome (taken from Ensembl Metazoa [90] or from the relevant literature). (B) Gene density, defined as the number of genes divided by the genome span (inferred from assembly). (C) Same-species OG membership, with bars from top to bottom representing 1-, 2-, 3-, and 4-member clusters. (D) HGT content, with bars from top to bottom representing the proportion of genes with hU ≥30 for each species (HGTC), in physical linkage with a known metazoan gene (‘Linked’), and with clan membership to nonmetazoan genes from phylogenetic analyses (‘Clan’). (E) Repeat content, with bars from top to bottom representing the proportion of each genome covered by LINE/SINEs, LTR elements, DNA elements, and simple/low-complexity repeats. HGT, horizontal gene transfer; HGTC, HGT candidates; LINE, long interspersed element; LTR, long terminal repeat; OG, orthologous group; SINE, short interspersed element.