MBE Advance Access originally published online on February 14, 2008
Molecular Biology and Evolution 2008 25(4):620-623; doi:10.1093/molbev/msn035
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letters |
Rare Genomic Characters Do Not Support Coelomata: Intron Loss/Gain
* National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD
Departament de Genètica, Universitat de Barcelona, Barcelona, Spain
E-mail: scottwroy{at}gmail.com.
 |
Abstract |
|---|
 TOP Abstract AcknowledgementsReferences |
|---|
Recently, a new phylogenetic method employing intron position sharing across species was proposed and support for a Coelomate clade reported (Zheng et al. 2007
. A rigorous analysis of the pattern of intron conservation supports the Coelomata clade of animals. Mol Biol Evol. 24:2583–2592.). Here, we show that the previous analysis depends on: 1) an idiosyncratic definition of "conserved" introns, 2) exclusion of all phylogenetically informative introns present in outgroups, 3) incorrect inference of change along the critical branch, and 4) lack of variation in rates across branches. The method thus seems unlikely to give accurate results. In addition, we address differences in rates of loss across intron sites, which Zheng et al. claimed invalidates our previous analysis that supported Ecdysozoa (Roy and Gilbert. 2005a. Resolution of a deep animal divergence by the pattern of intron conservation. Proc Natl Acad Sci USA. 102:4403–4408.). Instead, we show that our conclusions are likely to be robust to such concerns.
Key Words: Ecdysozoa • Coelomata • phylogeny • long branches • rare genomic characters
The phylogenetic position of nematodes continues to be debated. Many studies have concluded support for the grouping of nematodes and arthropods, consistent with the Ecdysozoa hypothesis (Aguinaldo et al. 1997; Dopazo H and Dopazo J 2005; Philippe et al. 2005; Roy and Gilbert 2005a; Delsuc et al. 2006; Irimia et al. 2007). Other studies have instead suggested the grouping of arthropods and deuterostomes to the exclusion of nematodes (consistent with the Coelomata hypothesis; Field et al. 1988; Wolf et al. 2004; Brinkmann et al. 2005; Philip et al. 2005; Rogozin et al. 2007a, 2007b). In several cases, support for Coelomata may reflect long-branch attraction and/or insufficient taxon sampling (Philippe et al. 2004; Baurain et al. 2007; Irimia et al. 2007; Lartillot et al. 2007).
Recently Zheng et al. (2007) studied spliceosomal intron positions across 11 eukaryotic genomes (fig. 1). For the studied species, the implications of the Ecdysozoa/Coelomata distinction reduce to whether arthropods (represented by Anopheles gambiae and Drosophila melanogaster) group with deuterostomes (represented by humans) or nematodes (represented by Caenorhabditis elegans). Their manuscript presents 1) a new method, which they claim supports Coelomata and 2) evidence for differences in the rate of intron loss across sites, which they claim nullifies our previous support for Ecdysozoa (Roy and Gilbert 2005a). We address these 2 lines of argument in turn.
|
The new method performs parsimony on a subset of "conserved" intron positions, defined as those present in at least 3 out of 5 taxa: D. melanogaster, A. gambiae, C. elegans, humans, and collective nonanimal outgroups. Because most phylogenetic patterns fulfilling this criterion are not informative (table 1), this reduces to introns present in both dipteran species and either C. elegans or humans (but not both). Thus, the authors’ major finding is: parsimony infers more intron gains (excluding those subsequently lost in a dipteran) in a hypothetical Coelomata ancestor than in an Ecdysozoan ancestor (as previously seen, e.g., compare figure 3b of Rogozin et al. [2003
] and figure 2 of Roy and Gilbert [2005b]), which they interpret as evidence for Coelomata. However, there are several problems.
|
First, support for Coelomata is entirely dependent on their idiosyncratic definition of "groups." Under their division whereas each dipteran species (D. melanogaster and A. gambiae, diverged 250–300 MYA; Holt et al. 2002
) constitutes a separate group, the deeply diverged nonanimals (plants, apicomplexans, and fungi) share a single group. As such, all phylogenetically informative conserved positions are animal specific, whereas all phylogenetically informative introns shared with nonanimals are deemed "variable" and excluded. When more intuitive phylogenetic groupings are used (e.g., 3 main animal groups, fungi, and plants + apicomplexans or animals, fungi, plants, and apicomplexans), and only introns present in at least three groups considered, the results consistently support Ecdysozoa over both Coelomata and the third alternative (arthropods as an outgroup to a clade including nematodes and deuterostomes, termed "Bizarre" by Rogozin et al. [2007a]). For every phylogenetically reasonable alternative grouping we could think of, more characters support Ecdysozoa (table 2). (Note that these results do not by themselves provide strong support for Ecdysozoa because parallel losses in relatively intron-poor nematodes and dipterans must be corrected for [as in previous methods; see below].)
|
Notably, all the support for Coelomata relies on the "Coelomata" introns (those specific to humans and dipterans) having been gained along the internal branch (otherwise these introns are ancestral bilaterian introns that have been lost along an external branch and so are not informative). However, it has been repeatedly shown that such introns are not gains in Coelomata, but ancestral bilaterian ancestors that have been lost in C. elegans. Of the 87 Coelomata gains inferred by parsimony in a data set directly related to that of Zheng et al. (Rogozin et al. 2003
), maximum likelihood reanalysis estimated than none (0) were true gains (Nguyen et al. 2005). More recently Carmel et al. (2007) estimated no gains along the Coelomata branch, even though parsimony infers 42 (calculated from their raw data, provided to us by Liran Carmel). Therefore, it seems most unlikely that many of these introns represent changes in a Coelomata ancestor.
That the Coelomata intron positions are not true Coelomata synapomophies becomes even clearer when the intron-rich basal animal Nematostella vectensis is considered. In a similar data set (Sullivan et al. 2006), 69% of intron positions shared between dipterans and humans (but absent from C. elegans and nonanimals and thus seemingly supportive of a Coelomata clade) are present in N. vectensis (calculated from their raw data, provided to us by Jim Sullivan), clear evidence that these introns are largely not true Coelomata synapomorphies. (We were unable to perform the analogous analysis on the Zheng et al. data set as the raw protein alignments were not preserved by the original authors.)
Further, the Zheng et al. method is expected to systematically place the taxon that has experienced the most intron loss as the outgroup. Regardless of topology, in order to be included in the study, an ancestral bilaterian intron must be retained all the way to dipterans (implying presence in the Coelomata or Ecdysozoan ancestor), as well as either 1) lost in humans and retained in C. elegans (supporting Ecdysozoa) or 2) lost in C. elegans and retained in humans (Coelomata). Because the C. elegans branch has experienced far more intron loss than humans (Rogozin et al. 2003; Banyai and Patthy 2004; Csurös 2005; Nguyen et al. 2005; Raible et al. 2005; Roy and Gilbert 2005a, 2005b; Carmel et al. 2007; Csuros et al. 2007), we a priori expect more ancestral bilaterian introns to artifactually support the Coelomata topology. The relative probabilities of loss along the 2 branches appear to be around 3 to 1: Among introns shared between a dipteran and a nonanimal, 24.0% and 63.5% are absent in humans and C. elegans, respectively. Thus, the relative probabilities of a character artifactually supporting Coelomata or Ecdysozoa are expected to be perhaps an order of magnitude different, consistent with the findings of Zheng et al.
The final point is more subtle. Zheng et al. introduce a modified Dollo parsimony argument that takes into account differences in probabilities of loss along branches—formally, differences in the cost for an intron loss along different branches. Different relative costs for loss along the human and C. elegans branches lead to different relative probabilities of Coelomata and Ecdysozoa. They show that the data support Coelomata only as long as the probability of loss in nematodes is at least 9.3 times higher than for humans (for details, see Zheng et al. 2007). However, the observed probability of loss along the nematode branch is only around 3 times higher than for humans (see above), suggesting a similar relative cost, well with the range over which the data support Ecdysozoa. In total, then, for several reasons the Zheng et al. results may not provide strong support for Coelomata.
The second claim of Zheng et al. was that variations in rates of intron loss across sites fully explain and nullify our previous findings (Roy and Gilbert 2005a). Zheng et al. first confirm previous findings of such differences (Sverdlov et al. 2004; Roy and Gilbert 2005c), which violate an assumption of our proposed intron position–based phylogenetic method (Rogozin et al. 2005; Roy and Gilbert 2005a). They conclude that our findings of support of Ecdysozoa are "fully explained by parallel loss of introns in nematodes and arthropods" (Zheng et al. 2007). However, no quantitative comparison of expected and observed magnitudes of signal is presented, and, in general, they do not study the specific expected effects of violations of the assumption on our method. It is therefore difficult to evaluate their claims.
How are such rate variations likely to affect our conclusions? Our method consists of a set of 6 comparisons between ratios of introns retained to lost (Roy and Gilbert 2005a). For instance, all introns shared between deuterostomes and nematodes (but not in the outgroup plants) were present in a presumed Coelomata ancestor; thus, the relative number of these that are retained versus lost in arthropods (ADN/DN in our notation, with the absence of "P" indicating absence in plants) is just given by the probability of retention along the arthropod branch. Similarly, among introns present in deuterostomes and plants (and present or absent in nematodes), the relative numbers present/absent (ADP[N]/DP[N], with "[N]" indicating either presence or absence in nematodes) in arthropods are again given by the probability of retention along the arthropod branch. Thus, assuming equal loss rates and no parallel insertion, we expect ADN/DN = ADP[N]/DP[N] (" = " in the Coelomata column for the first line of table 3). By contrast, assuming Ecdysozoa, presence in arthropods is still determined by retention/loss along the arthropod branch for the first group (ADN/DN) but is less likely for the second group, as it requires retention both along the internal branch leading to the Ecdysozoan ancestor and subsequent retention in arthropods for the second group (ADP[N]/DP[N]); thus, we expect ADN/DN > ADP[N]/DP[N] (">" in the Ecdysozoa column). Under the assumption of equal rates across sites, the 3 different possible topologies give qualitatively different predictions about the relationship between each pair of compared ratios—whether greater than, less than, or equal to (table 3). Please see (Roy and Gilbert 2005a) for a full development of the method.
|
How are these expectations likely to be affected by differences in rates across sites? In each comparison, the left-hand ratio concerns animal-specific introns, whereas the right-hand ratio concerns introns shared between plants and animals. Given their retention over deep phylogenetic distances, we might expect introns in the right-hand ratio to experience lower overall rates of loss and thus to have a higher retained-to-lost ratio. That is, the logic of the method likely underestimates the right-hand ratio (relative to the left) because introns on the right will experience lower rates of loss on average. Thus, the predictions should probably be revised toward less than: those that are nominally greater than are perhaps also truly consistent with equal to (under certain conditions); those that are nominally equal to might be consistent with less than.
Two of the 3 comparisons supporting Ecdysozoa are robust to these concerns, as they fulfill greater than predictions, which cannot be explained by lower rates of loss among right-hand introns (table 3). Therefore, the support for Ecdysozoa is likely to be robust to differences in rates of loss across sites. Although this in turn implies that some findings of " = " may be consistent with nominal greater than Coelomata predictions, there remain no clear comparisons that exclusively support Coelomata. That difference in loss rates across sites leads to right-hand ratios being relatively higher than predicted was confirmed by simulations and by analysis of expected ratios under various distributions across sites (2 site classes, a general gamma distribution; data not shown).
Fully harnessing the phylogenetic potential of spliceosomal intron positions over deep evolutionary distances requires improved evolutionary models of intron loss/gain. Two recent major studies make significant theoretical and empirical progress (Csuros et al. 2007; Carmel et al. 2007). An important remaining issue involves the prevalence of rate differences across sites, which have been repeatedly found (Sverdlov et al. 2004; Roy and Gilbert 2005b; Zheng et al. 2007) but which were not recovered by the most statistically sophisticated reconstruction to date (Carmel et al. 2007). Availability of large numbers of metazoan genomes affords the opportunity for significant model improvements.
In total, it is not clear what Zheng et al. have shown about the performance of our method. On the other hand, owing to several concerns raised here, in our opinion, their analyses do not provide convincing evidence for Coelomata.
|
Acknowledgements |
|---|
|
TOPAbstractAcknowledgementsReferences |
|---|
M.I. was funded by the Spanish Ministerio of Educación y Ciencia, through the FPI grant (BFU2005-00252), and S.W.R. by the Intramural Research Program of the National Library of Medicine at National Institutes of Health/DHHS. We thank Eugene Koonin and Jordi Garcia-Fernandez and their groups for intellectual support and stimulation, for financial support, and for fostering environments of open intellectual exploration in their respective groups.
|
Footnotes |
|---|
Peter Lockhart, Associate Editor
|
References |
|---|
|
TOPAbstractAcknowledgementsReferences |
|---|
Aguinaldo AMA, Turbeville JM, Linford LS, Rivera MC, Garey JR, Raff RA, Lake JA. Evidence for a clade of nematodes, arthropods and other moulting animals. Nature (1997) 387:489–493.[CrossRef][Medline]
Banyai L, Patthy L. Evidence that human genes of modular proteins have retained significantly more ancestral introns than their fly or worm orthologues. FEBS Lett (2004) 565:127–132.[CrossRef][Web of Science][Medline]
Baurain D, Brinkmann H, Philippe H. Lack of resolution in the animal phylogeny: closely spaced cladogeneses or undetected systematic errors? Mol Biol Evol (2007) 24:6–9.
Brinkmann H, van der Giezen M, Zhou Y, Poncelin de Raucourt G, Philippe H. An empirical assessment of long-branch attraction artefacts in deep eukaryotic phylogenomics. Syst Biol (2005) 54:743–757.
Carmel L, Wolf YI, Rogozin IB, Koonin EV. Three distinct modes of intron dynamics in the evolution of eukaryotes. Genome Res (2007) 17:1034–1044.
Csurös M. Likely scenarios of intron evolution. In: Comparative Genomics, Vol. 3678 of LNBI, Heidelberg (Germany): Springer—McLysaght A, Huson DH, eds. (2005) 47–60.
Csurös M, Holey JA, Rogozin IB. In search of lost introns. Bioinformatics (2007) 23:i87–i96.
Delsuc F, Brinkmann H, Chourrout D, Philippe H. Tunicates and not cephalochordates are the closest living relatives of vertebrates. Nature (2006) 439:965–968.[CrossRef][Medline]
Dopazo H, Dopazo J. Genome-scale evidence of the nematode-arthropod clade. Genome Biol (2005) 6:R41.[CrossRef][Medline]
Field K, Olsen G, Lane D, Giovannoni S, Ghiselin M, Raff E, Pace N, Raff R. Molecular phylogeny of the animal kingdom. Science (1988) 239:748–753.
Holt RA, Subramanian GM, Halpern A, et al. The genome sequence of the malaria mosquito Anopheles gambiae. Science (2002) 298:129–149. (123 co-authors).
Irimia M, Maeso I, Penny D, Garcia-Fernandez J, Roy SW. Rare coding sequence changes are consistent with Ecdysozoa, not Coelomata. Mol Biol Evol (2007) 24:1604–1607.
Lartillot N, Brinkmann H, Philippe H. Suppression of long-branch attraction artefacts in the animal phylogeny using a site-heterogeneous model. BMC Evol Biol (2007) 7:S4.[CrossRef][Medline]
Nguyen H, Yoshihama M, Kenmochi N. New maximum likelihood estimators for eukaryotic intron evolution. PLoS Comput Biol (2005) 1:e79.[CrossRef][Medline]
Philip GK, Creevey CJ, McInerney JO. The Opisthokonta and the Ecdysozoa may not be clades: stronger support for the grouping of plant and animal than for animal and fungi and stronger support for the Coelomata than Ecdysozoa. Mol Biol Evol (2005) 22:1175–1184.
Philippe H, Lartillot N, Brinkmann H. Multigene analyses of bilaterian animals corroborate the monophyly of Ecdysozoa, Lophotrochozoa and Protostomia. Mol Biol Evol (2005) 22:1246–1253.
Philippe H, Snell E, Bapteste E, Lopez P, Holland P, Casane D. Phylogenomics of eukaryotes: the impact of missing data on large alignments. Mol Biol Evol (2004) 21:1740–1752.
Raible F, Tessmar-Raible K, Osoegawa K, et al. Vertebrate-type intron-rich genes in the marine annelid Platynereis dumerilii. Science (2005) 310:1325–1326. (123 co-authors).
Rogozin I, Sverdlov A, Babenko V, Koonin E. Analysis of evolution of exon-intron structure of eukaryotic genes. Brief Bioinform (2005) 6:118–134.
Rogozin IB, Wolf YI, Carmel L, Koonin EV. Ecdysozoan clade rejected by genome-wide analysis of rare amino acid replacements. Mol Biol Evol (2007a) 24:1080–1090.
Rogozin IB, Wolf YI, Carmel L, Koonin EV. Analysis of rare amino acid replacements supports the Coelomata clade. Mol Biol Evol (2007b) 24:2594–2597.
Rogozin IB, Wolf YI, Sorokin AV, Mirkin BG, Koonin EV. Remarkable interkingdom conservation of intron positions and massive, lineage-specific intron loss and gain in eukaryotic evolution. Curr Biol (2003) 13:1512–1517.[CrossRef][Web of Science][Medline]
Roy SW, Gilbert W. Resolution of a deep animal divergence by the pattern of intron conservation. Proc Natl Acad Sci USA (2005a) 102:4403–4408.
Roy SW, Gilbert W. Rates of intron loss and gain: implications for early eukaryotic evolution. Proc Natl Acad Sci USA (2005b) 102:5773–5778.
Roy SW, Gilbert W. The pattern of intron loss. Proc Natl Acad Sci USA (2005c) 102:713–718.
Sullivan JC, Reitzel AM, Finnerty JR. A high percentage of introns in human genes were present early in animal evolution: evidence from the basal metazoan Nematostella vectensis. Genome Inform (2006) 17:219–229.[Medline]
Sverdlov A, Babenko V, Rogozin I, Koonin E. Preferential loss and gain of introns in 3' portions of genes suggests a reverse-transcription mechanism of intron insertion. Gene (2004) 338:85–91.[CrossRef][Web of Science][Medline]
Wolf Y, Rogozin I, Koonin E. Coelomata and not Ecdysozoa: evidence from genome-wide phylogenetic analysis. Genome Res (2004) 14:29–36.
Zheng J, Rogozin IB, Koonin EV, Przytycka TM. Support for the Coelomata clade of animals from a rigorous analysis of the pattern of intron conservation. Mol Biol Evol (2007) 24:2583–2592.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. W. Roy and M. Irimia Origins of Human Malaria: Rare Genomic Changes and Full Mitochondrial Genomes Confirm the Relationship of Plasmodium falciparum to Other Mammalian Parasites but Complicate the Origins of Plasmodium vivax Mol. Biol. Evol., June 1, 2008; 25(6): 1192 - 1198. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||