PHYLOGENY by David H. A. Fitch (

For a more recent discussion of evolution and phylogeny of the nematodes see Evolution and Ecology (David H. A. Fitch, ed.) in the Wormbook.

Why a Phylogeny for C. elegans?- Relationship of Nematoda to Other Animals-Relationships within Nematoda- Relationships of Rhabditid Nematodes- Back to Contents

Why a Phylogeny for C. elegans?

As a "model system", we would like to think of C. elegans as being representative of biological systems in general and of animals in particular. However, along with Drosophila, this "model" has a derived mode of development quite different from other species, probably as an adaptation for rapid life cycles in response to ephemerality of resources (Hodgkin and Barnes, 1991). To determine what features are primitive and shared ancestrally with other species, and which features are derived (e.g., convergent or novel), a model needs a phylogenetic context.

An accurate phylogeny for nematodes including their relationship to other animals has major consequences for our understanding of how diversity has evolved in morphology, genes, developmental mechanisms, behavior, and life history, but also for studies of comparative functional genomics. We need species phylogenies to allow discrimination between orthologues and paralogues in gene phylogenies.Discriminating orthologues from paralogues enables us to make predictions about the functions of genes in a genome for which there is little or no experimental data. Whereas paralogues share homology through gene duplication and often have diverged functions or expression patterns, orthologues generally have retained similarity in function (Eisen, 1998). Orthologues are found by overlaying a gene tree with the species tree to ascertain which branching events were due to speciation events. For example, lin-12 and glp-1, genes encoding Notch-like receptors, are paralogues resulting from a gene duplication predating the divergence among taxa in the Elegans species group (Rudel and Kimble, 2001). In C. elegans, these paralogues have diverged in their developmental regulatory roles (Rudel and Kimble, 2001). According to the gene phylogeny and the assumption of parsimony (features are shared more often due to common ancestry than convergence, parallelism or reversal; i.e., heredity works), we would predict that there is likely to be more similarity between Cbr-lin-12 and Cel-lin-12 (orthologues in C. briggsae and C. elegans, respectively) than between Cbr-lin-12 and Cel-glp-1. With the availability of genome data for two closely related species of rhabditid nematodes (C. elegans and C. briggsae), growing genomic data for one diplogasterid species (Pristionchus pacificus), an important human parasite (Brugia malayi), and other model systems, phylogeny provides an essential predictive tool for bioinformatics.

Reconstructing a phylogeny also allows us to infer the ancestral states that were preadaptive to novel life histories. Although we know that parasitism has evolved several times in nematodes, it is often not clear how this has happened. Discovering the closest nonparasitic relatives provides data from which ancestral features can be inferred. These nonparasitic models are also likely to share important features (e.g., anthelminthic drug sensitivity) mostly with their parasitic relatives. Since the strongylid parasites of vertebrates derived from within Rhabditidae, the family containing C. elegans (Fitch and Thomas, 1997), easily culturable nonparasitic rhabditids (e.g., CEW1 Oscheius tipulae) could provide such models. As another example, both C. briggsae and C. elegans are self-fertile hermaphrodites, although the Caenorhabditis ancestor was gonochoristic (female-male). From this, it is clear that hermaphroditism evolved from gonochorism, but how many times? If C. briggsae is more closely related to C. elegans than to other, gonochoristic, species, we can be pretty confident that hermaphroditism evolved once in a common ancestor of these two species (See PhylFIG1; yellow star marks the point where the trees in PhylFig1, 3 and 6 converge)

On the other hand, if C. briggsae is more closely related to gonochoristic C. remanei than to C. elegans (See PhylFIG2 b) (as seems likely-Fitch et al., 1995; Haag and Kimble, 2000), hermaphroditism could have evolved independently (as indeed it has in other lineages of rhabditid nematodes) or gonochorism evolved secondarily in C. remanei.

A phylogeny is required for studying the evolution of developmental mechanisms. Although the gonadal "anchor cell" (AC) is of central importance in patterning the midbody C. elegans vulva through an EGF-type signaling mechanism, this is not the case in several species with a posterior vulva (Sommer and Sternberg, 1994). According to our phylogeny of rhabditids, posterior-vulva species originated a couple of times independently from ancestors with midbody vulvae (see PhylFIG3; yellow star marks the point where the trees in PhylFig1, 3 and 6 converge). In at least one instance, this transformation has been accompanied by a switch to independence from an anchor cell signal (Sommer and Sternberg, 1994).

A phylogenetic framework is also required for testing hypotheses about the evolutionary correlation of traits that address questions regarding developmental constraints, epistasis or pleiotropy, coevolution, and adaptive scenarios (Brooks and McLennan, 1993; Maddison, 2000 ). Such tests might show that the phylogenetic distribution of one trait depends significantly on the distribution of a second trait. For example, in the male copulatory bursa (male tail) of rhabditid species, evolutionary changes in the fates of at least three Rn.p seam cells are phylogenetically correlated. This correlation suggests a "developmental constraint" in the sense that these cells probably share a common fate determination pathway (Fitch, 1997). A phylogenetic correlation also exists between the overall form of the male bursa (See PhylFIG4) and the type of mating position (Fitch, 2000), suggesting an adaptive relationship between bursa form and mating behavior. Another interesting phylogenetic correlation is consistent with the notion that, when the posterior arm of a didelphic female gonad becomes reduced, it becomes adaptive to have a posterior vulva (presumably to accommodate an extension of the anterior gonad arm; Fitch et al., unpub.). Strict coevolution is tested by congruence between the phylogenies of host and parasite species. For example, particular instances of phylogenetic incongruence suggest that strongylid parasites invaded marsupials several times and derived from ancestors that were parasitic to placental mammals (Dorris et al., 1999).


Relationship of Nematoda to Other Animals

As a significant model for developmental genetics, genomics, and neurobiology, it is important to determine how C. elegans is related to other animals, and in particular to other model organisms (Fitch and Thomas, 1997; Sidow and Thomas, 1994). Traditional morphological (Hyman, 1940) and recent molecular analyses using a variety of genes, taxa, and methods (Blair et al, 2002 ; Brown et al., 2001; Sidow and Thomas, 1994) date the divergence of nematodes to a point before the Deuterostomia-Protostomia divergence (See PhylFIG5 a; red worm icon denotes that C. elegans is within this branch). Other analyses using primarily small subunit ribosomal RNA genes (18S rDNA), however, place nematodes close to other "moulting animals" (clade "Ecdysozoa" Aguinaldo et al., 1997; Giribet and Ribera, 1998; Peterson and Eernisse, 2001) (See PhylFIG5 b; red worm icon denotes that C. elegans is within this branch). This controversial position for Nematoda has major implications for our understanding of animal evolution. For example, Distalless (Dll) is required both for patterning the nervous system and for making "appendages" in bilaterian animals (noted by Peterson and Eernisse, 2001) (Panganiban et al., 1997). Depending on the phylogenetic position of nematodes (which do not have appendages patterned by Dll), the appendage-making role of Dll either could be ancestral and lost in nematodes or could have been independently and convergently recruited in the evolution of bilaterian appendages in different species. The latter hypothesis would suggest that an important developmental constraint exists for appendage formation in which the same molecular pathway is recruited for the same final function (in this case, bilaterian appendage formation).

In two analyses using a broad genomic sampling of putatively orthologous genes from C. elegans, Homo sapiens, Drosophila melanogaster, and Saccharomyces cerevisiae, different conclusions were reached about the evolutionary relationships of these model systems. In one analysis, (Mushegian et al., 1998), 36 protein-coding genes were phylogenetically analyzed. Most (i.e., 24) genes supported a closer relationship between D. melanogaster and H. sapiens than between C. elegans and D. melanogaster, although many of these genes were evolving at a significantly faster rate in the C. elegans lineage than in the other lineages. The authors interpreted this to mean that there was potential for "long-branch attraction" artefacts (though such artefacts were not demonstrated) and that C. elegans was actually more closely related to flies than humans (i.e., the authors felt the "Ecdysozoa" clade was not falsified). In a more recent study (Blair et al., 2002), sequences from 100 orthologous genes were subjected to multiple tests designed to detect biases due to taxon representation and long-branch attraction. This study falsified the "Ecdysozoa" hypothesis (e.g., PhylFIG5 b) and supported the traditional "Coelomata" hypothesis (See PhylFIG5 a) by significant margins. To minimize or eliminate the artefactual effects of long branches on the topology of a phylogeny, the representation of taxa needs to be significantly increased. Hence, it will be important to obtain comparable data for additional animal and nematode taxa to further test the position of nematodes within the animal Tree of Life.


Relationships within Nematoda

Nematode systematics has had an interesting and convoluted history (Andrássy, 1976; Andrássy, 1983; Andrássy, 1984; Chitwood and Chitwood, 1950; Inglis, 1983; Lorenzen,1994; Maggenti,1981; Malakhov, 1994; Thorne, 1961). A good review with an interesting proposal for a revised classification based on molecular as well as traditional systematic methods has been published recently (De Ley and Blaxter, 2002) (See PhylFIG6; yellow star marks the point where the trees in PhylFig1, 3 and 6 converge). Although Nematoda was originally split into two major groups, Phasmidia and Aphasmidia (more recently, Secernentea and Adenophorea) later systematists (e.g., Malakhov, 1994) recognized the paraphyly of Adenophorea and suggested further division into three major groups. De Ley and Blaxter base their proposed phylogeny largely on molecular (18S rDNA) data (primarily Blaxter et al., 1998), supplemented with some morphological considerations. The reader is directed to the original publications for details of the inference and comparisons to earlier phylogenies and classifications.

Important features of the overall phylogeny include the following: (1) Rhabditids are most closely related to plectids, reflecting their origin from within a group. The traditional composition of this relationship would now be considered as plectids being paraphyletic with respect to Rhabditida (similar to fish being paraphyletic to tetrapods-see Glossary) (2) Most of the remaining taxa fall into two, possibly monophyletic, groups: Enoplia and Dorylaimia. The root of the nematode tree is not yet clear and could fall between these two latter groups (thus resulting in paraphyly of Enoplia) (3) Major morphological or habitat differences must now be relegated to a less major role in considering phylogenetic and taxonomic splits. For example, the morphologically highly divergent parasitic taxa, such as tylenchs and strongylids, are derived from freeliving rhabditids with much less morphological difference across quite deep phylogenetic divisions. Certainly this is consistent with the power of natural selection causing major changes during short periods of adaptive radiation and maintaining stasis over long periods of time.


Relationships of Rhabditid Nematodes

Caenorhabditis is a taxon traditionally placed within a larger group of largely freeliving terrestrial nematodes (Rhabditidae, Rhabditina, Rhabditida). Sudhaus (in Sudhaus, 1976) proposed an important hypothesis of the phylogeny of this group based on detailed cladistic analysis of morphological characters. A quite different classification was proposed by Andrássy (in Andrássy, 1983) based primarily on gradistic differences. In light of recent molecular data (e.g. Fitch et al., 1995) and reconsiderations of morphological characters such as those of the male bursa (e.g., Fitch, 1997; Fitch and Emmons, 1995; Fitch, 2000), much of rhabditid systematics will have to be substantially revised (see Sudhaus and Fitch, 2001). A tentative phylogeny (based on 18S rDNA analyses; Fitch et al., unpub.) with incomplete taxon representation is presented in PhylFIG1. Although much remains to be worked out, some relationships are sufficiently resolved to make some important observations (see below). Note that in detail, the phylogeny of this group, and its relationship to other groups, are likely to show much greater paraphyly than was recognized even in the phylogeny of De Ley and Blaxter (in De Ley and Blaxter, 2002). The following text is taken without much modification from Fitch ( 2000) and incorporates results reviewed in appendices to Sudhaus and Fitch (Sudhaus and Fitch, 2001) and unpublished (Fitch et al.; Kiontke et al., in prep.).

(1) As Sudhaus (in Sudhaus, 1993) proposed, insect parasites belonging to family Heterorhabditidae are derived from within Rhabditidae and are closely related to a monophyletic "Eurhabditis" species group with a very similar composition as that proposed by Sudhaus (Sudhaus, 1976) (but excluding some species of his originally proposed Cephaloboides group which were later placed into subgenus Poikilolaimus (Sudhaus, 1980).
(2) Similar to the Mesorhabditinae subfamily group proposed by Andrássy, a strongly supported monophyletic group (the "Mesorhabditis group" in PhylFIG1 includes species of Teratorhabditis, Mesorhabditis, Crustorhabditis and Bursilla (the Monhystera group of Sudhaus (Sudhaus, 1976). Contrary to previous 18S rDNA phylogenies (Blaxter et al., 1998), and due to more inclusive taxon representation, Teratorhabditis is not less closely related to other Rhabditidae than are diplogasterids.
(3) Although Sudhaus (Sudhaus, 1976) believed that Pelodera was a sister taxon of Teratorhabditis, our phylogeny suggests Pelodera is more divergently related to the "Mesorhabditis group" (See PhylFIG1). Pelodera is clearly not monophyletically related to Caenorhabditis or other peloderan species (i.e., species without pointed tail tips extending posterior of the bursa velum or "fan" of the male tail) of the "Eurhabditis" group; Andrássy's (1983) subfamily Peloderinae is therefore polyphyletic and should be dropped.
(4) Consistent with recent molecular results (Baldwin et al. in Blaxter et al., 1998; Fitch and Thomas, 1997), parasites of the order Strongylida are actually derived from within Rhabditidae and are closely related to species of the Heterorhabditidae and "Eurhabditis" groups (See PhylFIG1).
(5) Not predicted by any prior systematic work, Protorhabditis and Diploscapter species are closely related to each other (Diploscapter may be derived from Protorhabditis) and are closely related to Caenorhabditis (See PhylFIG1). Sudhaus (1976) did not consider diploscapterids and Andrássy (Andrássy, 1983) put them into a separate family altogether, probably because of their highly derived mouthparts.
(6) It should also be noted that the Caenorhabditis group is one of the more well-resolved groups in the phylogeny (except within the Elegans species group) and includes a number of interesting, recently (and to be) described new species from desert habitats (e.g., Sudhaus and Kiontke, 1996).
(7) Previous authors thought that Protorhabditis and Parasitorhabditis groups diverged early from the rest of the Rhabditidae. This inference derived from the assumption that absence of a glottoid apparatus (metastomal ridges) was primitive and not easily lost. Our analysis suggests instead that Protorhabditis and Parasitorhabditis derived recently from different lineages (See PhylFIG1).
(8) Although some authors (e.g. Maggenti, 1981) have placed diplogasterids as far away from rhabditids as a separate subclass, this monophyletic group is derived from within Rhabditidae and is very closely related to Rhabditoides inermis ("Rhabditoides 1" of PhylFIG1).
(9) Finally, species of Poikilolaimus (sensu Sudhaus, 1980; nec Andrássy, 1983) may be the most anciently diverged of species traditionally included in Rhabditidae, contrary to all other systems suggested.

Glossary for Phylogeny
Clade: a group of taxa that includes all the descendents of a common ancestor (synonymous with "monophyletic group").
Cladistic analysis: inference of genealogical relationships among taxa. Often used to refer specifically to phylogenetic analysis using the objective criterion of parsimony.
Gradistic differences: amount of difference between taxa; because of the possibility that different lineages can diverge (accumulate change) at different rates, gradistic differences may not be proportionate to cladistic relationship.
Long-branch attraction: when branches (lineages leading to different taxa) in a phylogenetic tree are very long (due to the accumulation of much change since a common ancestor), homoplasy (similarity due to convergent, parallel, or reversal changes) can become greater than similarity due to common ancestral change. That is, the longer the period of time since common ancestry, the more chance there is for accumulated changes to obliterate the evidence of common ancestry. In such cases, the longest branches tend to artefactually "attract" each other in phylogenetic analyses (especially if no correction for superimposed changes has been made).
Monophyletic: all descendents of an ancestor are included in a taxonomic group (see Clade).
Paraphyly/Paraphyletic: when some of the descendents of a common ancestor are not included in a taxonomic group (e.g., "fish" is a paraphyletic group that does not include tetrapod descendents of a sarcopterygian (lobe-finned) ancestor). Although the process of evolution is paraphyletic, some taxonomists ("cladists") try to avoid paraphyly in classification ("traditional" or "evolutionary" taxonomists allow paraphyly in their classification systems).
Phylogenetic incongruence: when the topology (relationships) of two trees being compared are different. For example, if strict coevolution has occurred between parasites and hosts (with strict host-specificity), the phylogeny of the parasites should match that of the hosts. Such strict coevolution is falsified when the phylogenies are incongruent.
Polyphyletic: when a taxonomic group includes descendents of different ancestors (avoided by nearly all systematists).
Taxon representation: an attempt to include as many taxa as possible that represent major phylogenetic branches (lineages).

Suggested links to nematode phylogeny related web sites:

All contents are copyright ©2002-2004 Wormatlas unless otherwise noted. See copyright and use policy.