Hugall AF et al. (2016), An Exon-Capture System for the Entire Class Oph...

ECB-ART-48263

Mol Biol Evol 2016 Jan 01;331:281-94. doi: 10.1093/molbev/msv216.

Show Gene links Show Anatomy links

An Exon-Capture System for the Entire Class Ophiuroidea.

Hugall AF , O'Hara TD , Hunjan S , Nilsen R , Moussalli A .

???displayArticle.abstract???
Exon-capture studies have typically been restricted to relatively shallow phylogenetic scales due primarily to hybridization constraints. Here, we present an exon-capture system for an entire class of marine invertebrates, the Ophiuroidea, built upon a phylogenetically diverse transcriptome foundation. The system captures approximately 90% of the 1,552 exon target, across all major lineages of the quarter-billion-year-old extant crown group. Key features of our system are 1) basing the target on an alignment of orthologous genes determined from 52 transcriptomes spanning the phylogenetic diversity and trimmed to remove anything difficult to capture, map, or align; 2) use of multiple artificial representatives based on ancestral state reconstructions rather than exemplars to improve capture and mapping of the target; 3) mapping reads to a multi-reference alignment; and 4) using patterns of site polymorphism to distinguish among paralogy, polyploidy, allelic differences, and sample contamination. The resulting data give a well-resolved tree (currently standing at 417 samples, 275,352 sites, 91% data-complete) that will transform our understanding of ophiuroid evolution and biogeography.

???displayArticle.pubmedLink??? 26474846
???displayArticle.pmcLink??? PMC4693979
???displayArticle.link??? Mol Biol Evol

Genes referenced: LOC100888042 LOC100893907 LOC115925415 LOC583082

???attribute.lit??? ???displayArticles.show???

	Fig. 1. Exon size distribution of the 425 gene data set, before and after selection of the exon-capture target.
	Fig. 2. The scale of the problem. (A) Exon distances among ophiuroids. Across the class most p-distances are well over the 12% benchmark, within families most are within 12%. (B) Diversity of capture probes required. The plot shows the cumulative distribution of the proportion of hybridization probes requiring a given number of representatives to ensure that no transcriptome sequence is more than 12% different. With 20 lineages 83% of probes fall within this limit across the candidate target of 425 genes.
	Fig. 3. Simple p-distance neighbor-joining tree of the aligned 1,552 exons of the original 52 transcriptome taxa and the subsequent 20 representative sequences (labeled SR). These are color-coded by the four kits into which they were combined. Taxa contributing to a representative are indicated by the first number in the label and gray brackets to the right. Note that kit 3 contains the clade 8 representative even though it is phylogenetically closer to the kit 2 sequences. This tree is essentially the same as the full transcriptome analysis in O’Hara et al. (2014).
	Fig. 4. General schema of read mapping strategies. Black arrows indicate primary input, mauve arrows indicate processing. Boxes indicate steps using BLAT sequence alignment software.
	Fig. 5. Summary of exon-capture and direct SR mapping performance. The plots show the number of samples (y-axes) for five key statistics (A)–(E) (x-axes). Blue and red refer to exons and COI (B and E), respectively. The sixth plot (F) shows proportion of comparisons (y-axis) against the difference in p-distance between sample and SR and sample and transcriptome exemplar, for clades with more than one exemplar (see fig. 3). Assembly-based mapping results appear similar.
	Fig. 6. Exon coverage distributions across a 44-sample test set. The four kits are color-coded as per figure 3.
	Fig. 7. Correlation of proportion of target recovered versus distance from SR. Analysis based on a test subset of 59 samples. Lines show two-degree polynomial best fit for direct SR mapping (red line, black squares) and assembly-based TASR mapping (blue line, blue crosses).
	Fig. 8. Distribution of exon polymorphism. The thick lines show average distribution of proportion of polymorphic sites per exon across a 44-sample test set: Blue direct SR mapping, gray assembly-based TASR mapping. The dashed line shows a log-linear coalescent expectation. The red lines show Ophiactis asperula F167536 before and after filtering of contaminating reads. The putative polyploid/hybrid Amphistigma minuta F173962 is shown in green.
	Fig. 9. Mitochondrial COI gene capture. Coverage along a long Trinity-assembled mtDNA contig containing the targeted COI gene (red line).This test sample (Ophiotreta valenciennesi UF8999) had 4.4 million reads and the COI gene was 17% different to the closest reference.
	Fig. 10. An ophiuroid exon-capture phylogeny. The tree is a RAxML codon-position GTR-CAT fast BS consensus with ML branch lengths, for 417 samples (380 species) from 275,352 sites in 1,490 exons, rooted according to O’Hara et al. (2014). Centre bar indicates divergence scale; taxon names have been omitted for simplicity; lineages are color-coded and labelled by mapping SR, and some large family groups indicated. The three higher-level nodes with BS support less than 95% are marked by asterisk; major difference between mapping pipelines in placement of the SR20 lineage is indicated by arrow. Genera mentioned in the text are denoted as follows: a, Astrophiura; b, Ophiomusium; c, Ophiosparte; d, Ophiomyces; e, Ophioscolex; f, Ophiocanops; g, Hemieuryale; h, Astrogymnotes; i, Ophiopsila; j, Amphilimna; k, Ophiopholis; l, Ophiothamnus.

References [+] :

Aberer, ExaBayes: massively parallel bayesian tree inference for the whole-genome era. 2014, Pubmed