Mongiardino Koch N et al. (2023), Confusion will be my epitaph: genome-scale disc...

ECB-ART-51306

Proc Biol Sci 2023 Jul 12;2902002:20230988. doi: 10.1098/rspb.2023.0988.

Show Gene links Show Anatomy links

Confusion will be my epitaph: genome-scale discordance stifles phylogenetic resolution of Holothuroidea.

Mongiardino Koch N , Tilic E , Miller AK , Stiller J , Rouse GW .

???displayArticle.abstract???
Sea cucumbers (Holothuroidea) are a diverse clade of echinoderms found from intertidal waters to the bottom of the deepest oceanic trenches. Their reduced skeletons and limited number of phylogenetically informative traits have long obfuscated morphological classifications. Sanger-sequenced molecular datasets have also failed to constrain the position of major lineages. Noteworthy, topological uncertainty has hindered a resolution for Neoholothuriida, a highly diverse clade of Permo-Triassic age. We perform the first phylogenomic analysis of Holothuroidea, combining existing datasets with 13 novel transcriptomes. Using a highly curated dataset of 1100 orthologues, our efforts recapitulate previous results, struggling to resolve interrelationships among neoholothuriid clades. Three approaches to phylogenetic reconstruction (concatenation under both site-homogeneous and site-heterogeneous models, and coalescent-aware inference) result in alternative resolutions, all of which are recovered with strong support and across a range of datasets filtered for phylogenetic usefulness. We explore this intriguing result using gene-wise log-likelihood scores and attempt to correlate these with a large set of gene properties. While presenting novel ways of exploring and visualizing support for alternative trees, we are unable to discover significant predictors of topological preference, and our efforts fail to favour one topology. Neoholothuriid genomes seem to retain an amalgam of signals derived from multiple phylogenetic histories.

???displayArticle.pubmedLink??? 37434530
???displayArticle.pmcLink??? PMC10336381
???displayArticle.link??? Proc Biol Sci

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

	Figure 1. . Representative holothuroid diversity included in this study. (a). Synapta sp. (b). Peniagone cf. vitrea. (c). Benthogone sp. (d). Pseudocolochirus violaceus. (e). Abyssocucumis albatrossi. (f). Colochirus robustus. (g). Ypsilothuria n. sp. (SIO-BIC E6221). (h). Molpadia amorpha. (i). Pseudostichopus cf. mollis. (j). Synallactidae. (k). Bathyplotes cf. moseleyi. The classification of these terminals can be found in electronic supplementary material, table S1. All photos except (g) are of the voucher specimens sequenced (catalogue numbers can be found in electronic supplementary material, table S1; further sampling information is accessible through the SIO-BIC online database, https://sioapps.ucsd.edu/collections/bi/). Images (b), (c), (i) and (k) are courtesy of the Schmidt Ocean Institute, and image (j) is courtesy of Monterey Bay Aquarium and Research Institute.
	Figure 2. . Summary of phylogenetic inference results. (a). Strict consensus of the nine inference conditions explored, varying both the number of loci and the method of inference. Nodes disagreeing between analyses are collapsed and labelled (see panels (b,c) for further details); branch lengths are otherwise taken from the CAT-PMSF analysis of the full dataset. (b). Monophyly of a clade composed of two cucumariid terminals, Colochirus robustus and Cucumaria georgiana, is rejected by ASTRAL-III, but upheld by the other methods (legend for support value grid is shown in (a)). (c). Systematic disagreement between all methods of inference regarding relationships among major neoholothuriid clades. The resolution favoured by each method is found across datasets of different sizes. Topologies, branch lengths and support values for each individual analysis are shown in electronic supplementary material, figures S2–S4.
	Figure 3. . Exploration of support for alternative neoholothuriid topologies across loci. (a). Principal components (PC) axes obtained from the three ΔGLS. Percentages of explained variance are shown on axis labels. Loci are colour coded depending on their favoured topology. (b). Relationship between the PC axes and the scores of individual ΔGLS. Trendlines correspond to LOESS smoothing curves, and ρ-values (Spearman's rank correlation coefficients) are shown when absolute values > 0.7, taken to represent strong correlations. The area included within ± 2 log-likelihood units is highlighted and considered an area of weak support. Note the markedly different scales of the y-axes for PCs 1 and 2. Topologies are colour coded as in (a).
	Figure 4. . Categorization of loci depending on their favoured topology, and exploration of potential determinants. Colouring scheme follows that of figure 3. (a). Most loci (615, 55.9% of the full dataset) can be considered uninformative regarding relationships among neoholothuriid clades. The remainder can be classified into those supporting a given topology (denoted using a plus sign, +) if they favour a given resolution against both alternatives (coloured section of bar chart) or only one (white section of bar chart) with a ΔGLS ≥ 2; or rejecting a given topology (denoted using a minus sign, −). The number of loci either supporting (right side of wheel) or rejecting (left side of wheel) the ASTRAL topology are comparable in number: 207 (18.8%) versus 278 (25.3%), respectively. Further details on loci categorization can be found in electronic supplementary material, figure S7. (b). Top: Exploration of 15 potential determinants of ΔGLS. Arrows indicate directions of maximum correlation between scores and determinants; their length is scaled to the strength of the correlation. Predictors mostly load onto PC 2. R2 and p-values are shown in electronic supplementary material, table S2, but no correlation is significant. Bottom: Smoothed surface of alignment length, the only significant determinant found using a classification tree. Longer loci are more likely to be informative, yet alignment length does not predict which topology is preferred (see electronic supplementary material, figure S9).

References [+] :

Al Jewari, Conflict over the Eukaryote Root Resides in Strong Outliers, Mosaics and Missing Data Sensitivity of Site-Specific (CAT) Mixture Models. 2023, Pubmed