A transcriptomic analysis of the phylum Nematoda.

Abstract

The phylum Nematoda occupies a huge range of ecological niches, from free-living microbivores to human parasites. We analyzed the genomic biology of the phylum using 265,494 expressed-sequence tag sequences, corresponding to 93,645 putative genes, from 30 species, including 28 parasites. From 35% to 70% of each species' genes had significant similarity to proteins from the model nematode Caenorhabditis elegans. More than half of the putative genes were unique to the phylum, and 23% were unique to the species from which they were derived. We have not yet come close to exhausting the genomic diversity of the phylum. We identified more than 2,600 different known protein domains, some of which had differential abundances between major taxonomic groups of nematodes. We also defined 4,228 nematode-specific protein families from nematode-restricted genes: this class of genes probably underpins species- and higher-level taxonomic disparity. Nematode-specific families are particularly interesting as drug and vaccine targets.

Full text

Page 1

A transcriptomic analysis of the phylum Nematoda
John Parkinson1,2, Makedonka Mitreva3, Claire Whitton2, Marian Thomson2, Jennifer Daub2, John Martin3,
Ralf Schmid2, Neil Hall4,6, Bart Barrell4, Robert H Waterston3,6, James P McCarter3,5& Mark L Blaxter2
The phylum Nematoda occupies a huge range of ecological niches, from free-living microbivores to human parasites. We
analyzed the genomic biology of the phylum using 265,494 expressed-sequence tag sequences, corresponding to 93,645 putative
genes, from 30 species, including 28 parasites. From 35% to 70% of each species’ genes had significant similarity to proteins
from the model nematode Caenorhabditis elegans. More than half of the putative genes were unique to the phylum, and 23%
were unique to the species from which they were derived. We have not yet come close to exhausting the genomic diversity of the
phylum. We identified more than 2,600 different known protein domains, some of which had differential abundances between
major taxonomic groups of nematodes. We also defined 4,228 nematode-specific protein families from nematode-restricted genes:
this class of genes probably underpins species- and higher-level taxonomic disparity. Nematode-specific families are particularly
interesting as drug and vaccine targets.
Nematodes, or roundworms, are a highly diverse group of organisms1.
What nematodes lack in obvious morphological disparity, they make
up for in abundance, accounting for 80% of all individual animals on
earth2, and diversity, with estimates ranging from 100,000 to 1 million
extant species3. They exploit a wide variety of niches and include free-
living terrestrial and marine microbivores, meiofaunal predators,
herbivores, and plant and animal parasites. On the basis of small
subunit ribosomal RNA (SSU rRNA) phylogenetics1,4, nematodes can
be divided into three major clades: Dorylaimia (clade I)1,4, Enoplia
(clade II) and Chromadorea (which includes Rhabditida, also known as
Secernentea). Rhabditida can be further divided into Spirurina (clade
III), Tylenchina (clade IV) and Rhabditina (clade V; Fig. 1). Parasitism
of both animals and plants seems to have arisen multiple times during
nematode evolution, and all major clades include parasites.
Most nematode diseases are intractable problems. Infections of
humans by nematodes result in substantial human mortality and
morbidity, especially in tropical regions of Africa, Asia and the
Americas: 2.9 billion people are infected. Morbidity from nematodes
is substantial and rivals diabetes and lung cancer in worldwide
disability adjusted life year measurements5. Although mortality is
low in proportion to the huge number of infections, deaths may
still total 100,000 annually. The most important parasites include
hookworms, Ascaris and whipworms (41 billion infections each) and
the filarial nematodes that cause elephantiasis and African river
blindness (120 million infections). Parasitic nematodes also cause
substantial losses in livestock and companion animals and are respon-
sible for $80 billion in annual crop damage worldwide6.
Much of what we know about the molecular and developmental
biology of nematodes stems from the study of the free-living soil
rhabditine nematode Caenorhabditis elegans (Fig. 1). C. elegans is a
versatile and tractable model organism, contributing substantially to
understanding of important medical fields including cancer, ageing,
neurobiology and parasitic diseases7–9. C. elegans was the first multi-
cellular organism whose genome was completely assembled7. Despite
the wealth of information available for C. elegans, and its sister species
Caenorhabditis briggsae10, comparatively little is known about other
members of this important phylum.
Two projects were initiated to generate new sequence data for
nematode parasites spanning the phylogenetic disparity of the phy-
lum11. We used expressed-sequence tags (ESTs; sequences derived
from randomly selected cDNA clones) as they are a cost-effective route
to gene discovery12. We generated 265,494 sequences from 30 different
species of nematode, the largest collection of ESTs representing the full
diversity of a single phylum. In addition to identifying traits that may
be species- or phylum-specific, this collection offers an unparalleled
opportunity to explore and elucidate evolutionary and functional
relationships. Here we present an overview of the sequence data
arising from the parasitic nematode EST project and place them in
the context of C. elegans genomic biology.
RESULTS
265,494 ESTs from nematodes other than Caenorhabditis
Nematode EST projects have generated more than 250,000 ESTs
from 30 target species (Fig. 1 and Table 1 online; refs. 11,13–20 and
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
Published online 14 November 2004; doi:10.1038/ng1472
1Hospital for Sick Children, 555 University Avenue; Departments of Biochemistry and Medical Genetics and Microbiology, University of Toronto, Toronto, Ontario
M5G 1X8, Canada.2School of Biological Sciences, University of Edinburgh, Edinburgh, EH9 3JT, UK.3Genome Sequencing Center, Washington University School
of Medicine, St Louis, Missouri 63108, USA.4Pathogen Sequencing Unit, Wellcome Trust Sanger Institute, Cambridge, CB10 1SA, UK.5Divergence, St Louis,
Missouri 63141, USA.6Present addresses: The Institute for Genomic Research, 9712 Medical Center Drive, Rockville, Maryland 20850, USA (N.H.); Department
of Genome Sciences, University of Washington, Seattle, Washington 98195, USA (R.H.W.). Correspondence should be addressed to J.P. (jparkin@sickkids.ca).
NATURE GENETICS ADVANCE ONLINE PUBLICATION1
ARTICLES

Page 2

J.P.M. and M.L.B., unpublished data). For each species, we grouped
ESTs into clusters and predicted consensus sequences for each cluster
(putative gene). These sequences together form the ‘partial genome’ of
each species. Figure 2a shows the level of redundancy (ESTs per gene)
associated with each partial genome. We observed diminishing
returns, in terms of new gene discovery, as we sequenced more
ESTs from one species. Redundancy was greatest for Ascaris suum,
the most heavily sampled species. The number of genes per species
ranged from 208 for the smallest EST set (Zeldia punctata, 388 ESTs)
to more than 9,500 (Brugia malayi, 25,067 ESTs). We defined 93,645
putative genes. This is probably a slight overestimate, as the clustering
process may split some allelic variation into distinct genes (most
parasitic nematode populations used were outbred), different splice
forms may not have been clustered together, and nonoverlapping
ESTs derived from the same mRNA will not have clustered. This
inflation is probably minor (B5%), based on comparisons to the
complete C. elegans proteome17,18and previous analyses of subsets of
these data18,19. If, as seems likely, most nematodes have B20,000
protein-coding genes7,10,21, we have tags for 1–50% of the expected
gene complement for each species, with a mean of B16%. The
total number of putative genes triples the number of nematode
genes defined22,23.
Genomic disparity across the phylum Nematoda
The different genes found in the genome(s) of an organism or group
of organisms can be thought of as occupying a ‘genespace’. More
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
Haemonchus contortus
Ostertagia ostertagi
Teladorsagia circumcincta
Necator americanus
Nippostrongylus brasiliensis
Ancylostoma caninum
Ancylostoma ceylanicum
Caenorhabditis briggsae
Caenorhabditis elegans
Pristionchus pacificus
Strongyloides ratti
Strongyloides stercoralis
Parastrongyloides trichosuri
Globodera pallida
Globodera rostochiensis
Heterodera glycines
Meloidogyne arenaria
Meloidogyne chitwoodii
Meloidogyne hapla
Meloidogyne incognita
Meloidogyne javanica
Pratylenchus penetrans
Zeldia punctata
Ascaris suum
Ascaris lumbricoides
Toxocara canis
Brugia malayi
Onchocerca volvulus
Dirofilaria immitis
Trichinella spiralis
Trichuris muris
Trichuris vulpis
0255075100%
Human parasite
Domestic animal parasite
Model animal parasite
Plant parasite
Free-living
(other Chromadorea)
Enoplia (clade II)
SSU rRNA phylogeny
Tylenchina (clade IV)
Tylenchomorpha
Spirurina (clade III)
Rhabditida
Chromadorea
Dorylaimia (clade I)
Ascaridomorpha
Spiruromorpha
Strongyloidea
Diplogasteromorpha
Panagrolaimomorpha
Cephalobomorpha
Rhabditoidea
Taxa studiedPercentage of partial genome with significant similarity to C. elegans
Trophic mode
Rhabditina (clade V)
Trichinellida
10,000
9,000
8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
Number of ESTs
Number of genes
Number of unique genes
Total number of genes
AscaridomorphaSpiruromorphaStrongyloidea
Diplogasteromorpha
Panagrolaimomorpha
Cephalobomorpha
TrichinellidaTylenchomorpha
50,000
40,000
30,000
20,000
10,000
60,000
0
025,00050,00075,000100,000125,000
ab
010,00020,00030,00040,000
Figure 2 Gene discovery in nematode EST data sets. (a) Gene discovery rates in nematode EST data sets. This graph shows the relationship between the
number of ESTs sequenced and the number of genes discovered. Each point represents an individual organism’s data set. See Table 1 for details.
(b) Exploration of genespace in the phylum Nematoda. The cumulative number of different genes (those that have no significant similarity to any other gene)
in the EST and proteome data sets from the phylum Nematoda. Each point represents the addition of one nematode species. The first point represents the
B22,000 C. elegans proteins. As each partial genome data set was added, increasing the total number of genes (x axis), the number of different genes
(y axis) increased. There was no apparent fall-off in the rate of discovery of new genes, suggesting that nematode genespace may be very large. The colors
indicate the systematic origin of each species group (see Fig. 1).
Figure 1 EST data sets from across the phylum
Nematoda. (a) Species are grouped into major
taxonomic groups based on SSU rRNA
phylogeny1,4. This differs from ‘traditional’
phylogenies but is consistent with current
morphological and developmental evidence. The
trophic biology of each targeted species is
indicated by a small icon. (b) The proportion of
each species’ partial genome that has significant
similarity (a match with a raw BLASTX score
Z50) to the complete proteome of C. elegans.
Owing to the difference in criteria used to define
significant similarity, these numbers differ slightly
from those previously reported17,18.
2ADVANCE ONLINE PUBLICATION NATURE GENETICS
ARTICLES

Page 3

complex genomes occupy a larger genespace, in general, and larger
groups of organisms (e.g., phyla) have a genespace that is the union of
the constituent species’ genespaces. Analysis of bacterial genespace
from complete genomes showed that sequencing additional eubacter-
ial genomes has yielded diminishing returns in terms of novelty24. If
nematode genespace is similarly limited, the fact that the genomes of
C. elegans and C. briggsae have been completely sequenced means that
sampling additional genomes will result in a low rate of gene discovery
as additional genomes are sampled. We carried out an exhaustive
series of cross-species BLAST analyses to estimate the extent of
nematode genes. We found that 30–70% of the genes from each
species had no non-nematode homolog (Table 1). The partial nature
of the consensus sequences derived from the ESTs may preclude
finding matches: short sequences will not reach our score cutoff,
and some consensuses may cover only 3¢ untranslated regions. Of the
64,685 cluster consensuses longer than 400 bp, 29,118 (45%) had no
significant match to non-nematode sequences; for consensuses less
than 400 bp in length, 78% seemed to be new. Thus, even excluding
short sequences, nearly half the predicted genes seem to be new. The
rate of discovery of genetic novelty has not yet started to decline with
the analysis of new genomes (Fig. 2b), implying that nematode
genespace may be much larger than bacterial genespace.
Of the 93,645 putative genes identified in this study, 14,630 (B15%)
had significant sequence similarity to putative genes in all of the five
major nematode taxonomic groups (i.e., homologs of these genes were
identified in all major clades; Table 1). Most of these (13,368; 91%) also
had homologs outside Nematoda and therefore are probably involved
in core metabolic or structural pathways. We found that 1,262 genes are
nematode-specific but widely represented within the phylum. These
genes may have roles unique to the nematode body plan and life
history and are good targets for pan-nematode control drugs.
These findings raise an important issue: if organisms from the same
major taxonomic group share only B60% of their genes, then
individual taxa may have widely divergent biology. Within the genus
Caenorhabditis, C. elegans and C. briggsae share B90% of their genes
at the level of discrimination used here10. Our survey suggests that this
level of genetic novelty may be universal across the Nematoda.
Lineage-specific genes could have completely new functions, could
have similar functions to genes in other organisms but use a
completely different mechanism (analogous genes), or could have
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
Table 1 Summary information of sequence data derived from 30 different species of nematodes
CladeGenus and speciesESTs Gene (clusters)LibrariesUnique to species (%)Unique to clade (%)Unique to Nematoda (%)
Dorylaimia (clade I)Trichuris muris
Trichinella spiralis
Trichuris vulpis
2,713
10,384
2,958
1,577
3,680
1,257
1
7
1
489 (31)
1,369 (37.2)
461 (36.7)
561 (40.6)
1,428 (38.8)
606 (48.2)
775 (49.1)
1,819 (49.4)
690 (54.9)
Spiruria (clade III)
Ascaridomorpha Ascaris lumbricoides
Ascaris suum
Toxocara canis
Brugia malayi
Dirofiliaria immitis
Onchocerca volvulus
1,822
38,944
4,206
25,067
3,585
14,922
853
8,482
1,447
9,511
1,747
5,097
1206 (24.2)
3,365 (39.7)
338 (23.4)
4,465 (46.9)
656 (37.6)
1,914 (37.6)
464 (54.6)
3,797 (44.8)
402 (27.8)
5,010 (52.6)
839 (48.1)
2,211 (43.4)
611 (71.6)
5,235 (61.7)
768 (53.1)
6,572 (69.1)
1,070 (61.2)
3,242 (63.6)
24
5
24
2
9
Spiruromorpha
Tylenchina (clade IV)
Panagrolaimomorpha Parastrongyloides trichosuri
Strongyloides ratti
Strongyloides stercoralis
Globodera pallida
Globodera rostochiensis
Heterodera glycines
Meloidogyne arenaria
Meloidogyne chitwoodi
Meloidogyne hapla
Meloidogyne incognita
Meloidogyne javanica
Pratylenchus penetrans
Zeldia punctata
7,712
9,932
11,236
1,317
5,905
18,524
3,251
7,036
13,462
12,394
5,282
1,908
388
3,086
3,264
3,635
977
2,851
7,198
1,892
2,409
4,479
4,408
2,609
408
208
6746 (24.2)
748 (22.9)
556 (15.3)
425 (43.5)
694 (24.3)
2,195 (30.5)
308 (16.3)
700 (29.1)
1,141 (25.5)
1,049 (23.8)
641 (24.6)
88 (21.6)
15 (7.2)
860 (27.9)
920 (28.2)
699 (19.2)
509 (52.1)
979 (34.3)
2,614 (36.3)
968 (35.3)
1,037 (43.1)
1,759 (39.3)
1,814 (41.2)
1,169 (44.8)
114 (28)
15 (7.2)
1,564 (50.7)
1,548 (47.4)
1,401 (38.5)
630 (64.5)
1,387 (48.6)
3,594 (49.9)
946 (50)
1,346 (55.9)
2,388 (53.3)
2,405 (54.6)
1,531 (58.7)
196 (48)
43 (20.7)
10
2
3
2
10
1
2
4
4
5
1
1
Tylenchomorpha
Cephalobomorpha
Rhabditina (clade V)
StrongyloideaAncylostoma caninum
Ancylostoma ceylanicum
Haemonchus contortus
Necator americanus
Nippostrongylus brasiliensis
Ostertagia ostertagi
Teladorsagia circumcincta
Pristiochus pacificus
9,079
10,544
17,268
4,766
1,234
6,670
4,313
8,672
265,494
4,203
3,485
4,146
2,294
742
2,355
1,655
3,690
93,645
3
9
1,453 (34.6)
730 (20.9)
797 (19.2)
689 (30)
145 (19.5)
500 (21.2)
362 (21.9)
1,108 (30)
28,353
1,910 (45.5)
1,037 (29.7)
1,039 (25)
920 (40.1)
174 (23.4)
681 (28.9)
548 (33.1)
1,128 (30.5)
2,648 (63)
1,696 (48.7)
2,012 (48.5)
1,476 (64.3)
382 (51.5)
1,369 (58.1)
1,008 (60.9)
1,915 (51.9)
52,267
12
3
3
10
4
3
172
Diplogasteromorpha
Total
Species are grouped into major taxonomic groups1,4(Fig. 1). The number of putative genes (size of partial genome) associated with each species is given by the number of clusters
(derived from the ESTs). ‘Unique to species’ indicates genes with no significant (BLAST score o50) sequence similarity to a gene outside that species. ‘Unique to clade’ indicates
genes that share significant sequence similarity to at least one other member of the same clade but do not have any similarity to any gene from out with the clade in question.
‘Unique to Nematoda’ indicates genes that do not share significant sequence similarity with any non-nematode protein.
NATURE GENETICS ADVANCE ONLINE PUBLICATION3
ARTICLES

Page 4

© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
0246810 12 14Mb
02468 101214Mb
0246810 12 Mb
C. elegans genes (3,059)
RNAi phenotypes (301)

Dorylaimia (clade I) (364)
Spirurina (clade III) (712)
Tylenchomorpha (clade IV) (657)
Panagrolaimomorpha (clade IV) (968)
Rhabditina (clade V) (894)

Ascaris suum (all) (354)
Ascaris suum (intestine) (73)
Ascaris suum (embryo) (58)
All nematode clades and elsewhere (401)
All nematode clades but unique to nematodes (25)
C. elegans genes (3,654)
RNAi phenotypes (299)

Dorylaimia (clade I) (293)
Spirurina (clade III) (643)
Tylenchomorpha (clade IV) (511)
Panagrolaimomorpha (clade IV) (830)
Rhabditina (clade V) (822)

Ascaris suum (all) (281)
Ascaris suum (intestine) (72)
Ascaris suum (embryo) (33)
All nematode clades and elsewhere (322)
All nematode clades but unique to nematodes (34)
C. elegans genes (2,946)
RNAi phenotypes (337)

Dorylaimia (clade I) (405)
Spirurina (clade III) (681)
Tylenchomorpha (clade IV) (625)
Panagrolaimomorpha (clade IV) (947)
Rhabditina (clade V) (891)

Ascaris suum (all) (321)
Ascaris suum (intestine) (83)
Ascaris suum (embryo) (42)
All nematode clades and elsewhere (398)
All nematode clades but unique to nematodes (36)
Chromosome I
Chromosome II
Chromosome III
0246810 12 Mb
C. elegans genes (2,977)
1416
024681012 Mb
C. elegans genes (5,032)
14161820
024681012
1416
Mb
C. elegans genes (3,348)
RNAi phenotypes (245)

Dorylaimia (clade I) (312)
Spirurina (clade III) (643)
Tylenchomorpha (clade IV) (501)
Panagrolaimomorpha (clade IV) (820)
Rhabditina (clade V) (837)

Ascaris suum (all) (284)
Ascaris suum (intestine) (65)
Ascaris suum (embryo) (35)
All nematode clades and elsewhere (366)
All nematode clades but unique to nematodes (47)
Chromosome IV
Chromosome V
Chromosome X
RNAi phenotypes (180)

Dorylaimia (clade I) (283)
Spirurina (clade III) (648)
Tylenchomorpha (clade IV) (557)
Panagrolaimomorpha (clade IV) (825)
Rhabditina (clade V) (816)

Ascaris suum (all) (284)
Ascaris suum (intestine) (67)
Ascaris suum (embryo) (42)
All nematode clades and elsewhere (343)
All nematode clades but unique
to nematodes (42)
RNAi phenotypes (112)

Dorylaimia (clade I) (245)
Spirurina (clade III) (511)
Tylenchomorpha (clade IV) (450)
Panagrolaimomorpha (clade IV) (685)
Rhabditina (clade V) (620)

Ascaris suum (all) (284)
Ascaris suum (intestine) (54)
Ascaris suum (embryo) (28)
All nematode clades and elsewhere (295)
All nematode clades but unique to nematodes (51)
Figure 3 Chromosomal location of C. elegans homologs of other nematode genes. Each panel represents a different C. elegans chromosome (autosomes I–V
and the X sex chromosome). Track 1: the average gene density per 100-kb division (brightest red, 440 genes per 100 kb; blue, o10 genes per 100 kb).
Track 2: the relative abundance of genes with visible RNAi phenotypes (red, highest abundance; blue, no genes represented). Tracks 3–7: the abundance of
C. elegans genes with homologs in the pooled partial genome data sets of the five major clades of the Nematoda. Tracks 8–10: the abundance of C. elegans
genes with homologs in the complete partial genome, in tissue (intestine) and in stage-specific (embryo) subsets from A. suum. Track 11: the positions of
individual C. elegans genes that have homologs in all five major clades. Track 12: the positions of individual C. elegans genes that have homologs in all five
major nematode clades but are not significantly similar to any non-nematode gene (a subset of the genes plotted in track 11). The number of genes
contributing to each track plot is given in brackets.
4ADVANCE ONLINE PUBLICATION NATURE GENETICS
ARTICLES

Page 5

similar features, such as tertiary structure, that enable them to use
similar molecular mechanisms.
Genomic conservation across the phylum Nematoda
Since the last common ancestor of Nematoda, B750–650 million
years ago25,26, nematodes have evolved to occupy many niches. The
success of nematodes today reflects the expression of successful
complexes of genetic traits, some of which are derived from the
common ancestor. A contrasting tendency towards evolution of new
genetic function underpinning particular fitness advantage in parti-
cular habitats will have resulted in divergence in gene complement. We
examined the patterns of gene gain, retention and loss across the
phylum by comparing partial genomes both within and between
major clades.
The complete genome of C. elegans yields a predicted proteome of
more than 22,000 polypeptides, some of which derive from alternative
splicing and more than 75% of which have some experimental
verification27,28. We carried out extensive comparisons of the 93,645
new nematode genes with this data set; comparisons with the genome
of C. briggsae10yielded similar results (data not shown). For each C.
elegans chromosome, we examined the patterns of gene density and
the density of genes with known RNA interference (RNAi) phenotypes
(Fig. 3). As described previously, each autosome has a greater density
of genes in the autosomal centers7,29, and RNAi phenotypes cluster to
the centers28,30. The autosomal arms are enriched in C. elegans–
specific genes, in tandemly duplicated gene families and in repetitive
and transposable elements7. The sex chromosome (X) has a more even
distribution of genes and RNAi phenotypes along its length.
We compared the nematode partial genomes to the C. elegans
proteome in this genomic context (Fig. 3). Although we cannot make
definitive statements concerning the absence of homologs, matches,
when summed over major clades, should identify overall themes.
Chromosomal centers were enriched, relative to arms, in the density of
similarity matches to all the major taxonomic groups. The pattern of
match density faithfully reflected the distribution of genes with RNAi
phenotypes28,30rather than overall gene density. Thus, taking chro-
mosome (chr) II as an example, each of the five major clades had a
high density of matches from B4 Mb to 12 Mb (the chromosomal
center), with additional peaks of RNAi phenotype genes at 0–0.5 Mb,
3.5 Mb, 13 Mb and 14–15 Mb mirrored by matches in the nematode
partial genomes (Fig. 3).
There were additional regions of high density of matches that did
not coincide with RNAi phenotype genes. On chr II, we observed high
match densities in all major clades at 1.0–1.2 Mb and 13.6 Mb, where
RNAi phenotype genes were rare. We noted similar regions on chr I
(1.2–2.0 Mb), chr III (0.8–1.0 Mb), chr IV (1.8 Mb) and chr V (B1
Mb and B20 Mb; Fig. 3). On the X chromosome there was a cluster
of conserved genes at B17–18.5 Mb that did not have a high
frequency of RNAi phenotypes. These conserved genes may not
yield RNAi phenotypes in C. elegans because of gene family redun-
dancy, or because they are involved in nematode-specific phenotypes
that are not assessed by currently applied assays under select labora-
tory conditions. Conversely, there were a few chromosomal regions
where a high density of RNAi phenotype genes was not matched by a
high density of matches to the partial genomes: one example is on chr
II at 2.8 Mb (Fig. 3).
Comparison between autosomes showed that chr V had B60%
fewer matches per C. elegans gene than the other autosomes. This adds
to the known peculiarities of chr V (ref. 31), which also has more
C. elegans–specific genes, has fewer RNAi phenotypes per gene, and is
significantly less likely to match genome survey sequences from the
filarial nematode B. malayi21.
Mapping genomic divergence within the phylum Nematoda
C. elegans is a rhabditine nematode1, closely related to the Diplogas-
teromorpha (represented here by Pristionchus pacificus) and the
vertebrate-parasitic Strongyloidea (represented here by seven species;
Fig. 1 and Table 1)4. The proportion of the partial genome of each
sampled species that contained C. elegans homologs roughly followed
the species’ relationship to C. elegans, with rhabditine species having
the highest mean proportion of homologs (Figs. 1 and 4). Compara-
tive analyses of the full partial genome data sets yield a set of similarity
relationships congruent with the SSU rRNA phylogeny4. We used
SimiTri analyses, which plot the relative similarity of entire gene data
sets against three data sets of interest32, to examine these relationships
further. Comparison of the partial genomes of the hookworm Ancy-
lostoma caninum and Haemonchus contortus (another strongyloid
parasite), P. pacificus and C. elegans (Fig. 4a) showed that although
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
Caenorhabditis elegans
Other rhabditid nematodes
466
Dorylaim, spirurine and
tylenchine nematodes
Spirurine nematodes
Dorylaim nematodes
116
Tylenchine nematodes
ac
Haemonchus contortus
114
Pristionchus pacificusCaenorhabditis elegans
b
34 831
34369
216
32 223
109223
245
8591,924
83477
3,631
Color scale of maximal
BLAST scores for tiles
695
1,420
4,583
Figure 4 Comparing partial genomes across the Nematoda. SimiTri plots provide a two-dimensional representation of the degree of similarity of an entire
data set of sequences with those of three different organisms32. The plots allow estimation of relationships of whole sequence data sets and highlight genes
with patterns of conservation that differ from the main trend in a data set (see Supplementary Methods for a more detailed description). (a) A. caninum
compared with H. contortus, P. pacificus and C. elegans. This plot shows that although A. caninum has more matches to C. elegans (because its complete
genome is available), overall, the sampled transcriptome from A. caninum is closer to that of H. contortus. (b) A. caninum compared with other rhabditine
nematodes (excluding C. elegans), nematodes from other clades and C. elegans. (c) All partial genomes from rhabditid nematodes (excluding C. elegans
species) compared to dorylaim, spirurine and tylenchine nematode partial genomes. The numbers at each vertex indicate the number of genes matching only
that database. The numbers on the edges indicate the number of genes matching the two linked databases. The number in each triangle indicates the
number of genes with matches to all three databases.
NATURE GENETICS ADVANCE ONLINE PUBLICATION5
ARTICLES

Page 6

genes with matches to C. elegans comprised the largest subset, most
genes were more similar to H. contortus than to P. pacificus. A number
of genes had high-scoring matches in P. pacificus and H. contortus but
were lacking from the complete C. elegans proteome, probably
indicative of gene loss in the C. elegans lineage. Extending these
analyses to span the major nematode clades showed the expected
global trend for A. caninum genes to be most similar to homologs
from other rhabditids (including other species from Strongyloidea;
Fig. 4b) and least similar to homologs from the dorylaims Trichinella
spiralis and Trichuris muris (Fig. 4c)19.
Tylenchomorpha, Cephalobomorpha and Panagrolaimomorpha
species also had a high proportion of matches to rhabditine sequences,
but as the predicted phylogenetic distance from C. elegans increased,
there was a corresponding decrease in the number of genes sharing
significant similarity (Fig. 1). The cephalobomorph Z. punctata seems
to be an exception, but this anomaly is probably due to sampling from
a single spliced leader-PCR–based library that biases towards short,
conserved transcripts16. Only B45% of the genes from the dorylaims
T. muris and T. spiralis had significant similarity to genes from
C. elegans, but a similar percentage of their genes shared similarity
with Drosophila melanogaster (data not shown). Thus T. spiralis and
T. muris may be good choices for deeper genomic analysis of the
relationships of Nematoda to other metazoan phyla. Overall, these
results suggest that C. elegans will be an effective genomic model for
other rhabditid nematodes, but that accumulated differences will
make extrapolation to distantly related nematodes, such as dorylaims,
more challenging. There are many nematode genes, found in species
across the phylum, that are absent from C. elegans. Gene loss has
therefore been an important part of C. elegans genome evolution, as
was suggested by the finding that C. elegans’ depleted HOX gene
complement is a result of lineage-specific losses33.
It has previously been suggested that there is a high-level ordering
of genes within C. elegans chromosomes, with, for example, muscle-
expressed genes being located in close proximity to each other more
often than would be expected by chance34, and RNAi phenotyping
suggesting that large chromosomal domains of genes have similar
biological roles28. We investigated whether nematode-specific or stage-
specific genes were clustered at a megabase level but did not find
evidence for linkage of these classes of genes. For example, we mapped
homologs of A. suum genes that had stage- or tissue-specific expres-
sion patterns to the C. elegans genome (Fig. 3). The tissue-specific
(intestine) or stage-specific (embryo, L3, L4) genes showed the same
general pattern of distribution as all A. suum genes.
Nematode-specific genes and gene families
Putative nematode-specific targets for drugs with diminished risk of
toxicity to hosts or other nontarget organisms may be found in the
class of nematode-specific genes. We found that 30–50% of each of
our chosen species’ partial genomes was unique. Of the 52,267 genes
for which no homolog was identified outside the phylum (Table 1),
21,640 had significant similarity with a sequence from another
nematode species. Mapping these nematode-specific genes onto the
phylogeny showed an incremental evolution of novelty (Fig. 5). Most
unique genes were associated with shallow-level taxonomic groups,
but a considerable proportion had a deeper origin. Some deep splits in
Nematoda were associated with few unique genes (e.g., Panagrolai-
momorpha plus Tylenchomorpha/Cephalobomorpha has only 198),
perhaps reflecting relatively rapid divergence of these daughter clades
shortly after the origin of the ancestral tylenchine.
We clustered the predicted polypeptides associated with these
nematode-specific genes using Tribe-MCL35into putative gene
families and mapped the latest possible origin of these families onto
the phylogeny (Fig. 5). In general, the number of new gene families at
each node of the tree reflected the number of genes associated with the
smallest daughter clade. The two largest groups of unique gene
families occur at the base of Rhabditida and at the node connecting
Spirurina with Tylenchina plus Rhabditina. This possibly reflects both
the relatively large number of taxa and the number of ESTs generated
for these three clades (e.g., the Tylenchina data set included several
closely related Meloidogyne species17). Most unique gene–origin events
seemed to occur relatively early in the nematode lineage. For example,
more than 6,500 genes had homologs in each of the three taxonomic
groups in the Rhabditida, and these included 4,330 genes in 1,262
nematode-specific families (Fig. 5).
We examined some nematode-specific gene families in more
detail. For many we identified C. elegans members, permitting
exploration of possible function through published RNAi data28.
For example, a family with ten members from all major clades
in our data set had a single C. elegans homolog identified by RNAi
to be essential for postembryonic larval development (Supple-
mentary Fig. 1 online). Another was limited to Rhabditida and
had a C. elegans member with an RNAi phenotype of inhibition
of postembryonic growth (Supplementary Fig. 1 online). In both
cases, the degree of sequence conservation suggests that the
RNAi function may be ascribed to the other genes: the C. elegans
phenotype recommends these for further investigation as targets
for nematicides.
Domain and functional analysis of nematode proteins
We used InterPro36to identify known protein domains in the partial
genomes. Because the C. elegans proteome has been extensively investi-
gated for protein domains7,37, many domains of unknown function
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
Ascaridomorpha
Spiruromorpha
Strongyloidea
Diplogasteromorpha
Panagrolaimomorpha
Cephalobomorpha
Trichinellida
Tylenchomorpha
Rhabditoidea
Total unique genes
and gene families
21,640
4,228 (13,903)
Spirurina (clade III)
Tylenchina (clade IV)
Rhabditina (clade V)
Dorylaimia (clade I)
6,782
349 (1,096)
1,515
327 (976)
338
55 (178)
198
40 (139)
3,021
104 (317)
10,779
687 (2,114)
4,741
109 (457)
8,619
166 (541)
3,012
72 (255)
354
79 (247)
3,055
677 (1,948)
6,582
1,262 (4,330)
1,927
301 (1,305)
Rhabditida
Nematoda
Figure 5 Evolutionary origins of unique genes and gene families in the
phylum Nematoda. The inferred positions of origin for the nematode-specific
genes and gene families were mapped across the robust SSU rRNA
phylogeny1,4. For each node, the upper number shows the number of genes
unique to each clade, and the lower number shows the predicted number of
unique gene families (with the number of individual predicted genes
included in these families in brackets). In the absence of complete genome
sequences from most of the Nematoda, this mapping places the origin of
each gene or gene family at the highest possible node: adding complete
genome sequences will tend to move the node of origin of some genes
lower in the tree.
6 ADVANCE ONLINE PUBLICATION NATURE GENETICS
ARTICLES

Page 7

have been defined that are exclusive to C. elegans and C. briggsae.
Many of the matches we discovered were to these nematode-restricted
domains. For each species, 30–50% of the polypeptides were predicted
to contain at least one previously identified domain (Supplementary
Table 1 online). Fewer polypeptides from both spirurine and dorylaim
species than from tylenchine and strongyloid species contained a
domain, reflecting the C. elegans bias in InterPro. The number of
unique domains associated with each species increased with size of its
partial genome (Supplementary Table 1 online).
Comparison of the most abundant domains associated with each
clade showed that, with the exception of the protein kinase domain,
the abundant domains did not correlate well with those previously
identified in the complete proteomes of C. elegans and C. briggsae
(Supplementary Table 2 online)7,10. These differences may have arisen
from the unavoidable bias in the types of genes sampled by ESTs. We
minimized this bias by grouping the partial genomes and their domain
contents by major clade (Supplementary Table 2 online). Cuticle
collagens (IPR008160 and IPR002048) are abundant in C. elegans and
C. briggsae ( B170 in each) but were poorly represented in the
dorylaim partial genomes, possibly reflecting the derivation of these
data sets from nonmoulting stages. Collagens have a temporally
restricted expression pattern in C. elegans, with most expression in
larval stages38. The strongyloid partial genomes were enriched for
peptidases (IPR000169, IPR001254, IPR001353 and IPR00668), and
dorylaim sequences were enriched for potential proteinase inhibitors
(IPR008197 and IPR008198) and for chymotrypsin (IPR001254). This
may reflect the parasitic niche of the sampled species (the host
intestine), where peptidases and inhibitors may be required for feeding
and survival in such a hostile environment. Also in Strongyloidea, the
ShK metridin-like toxin domain (IPR003582) was highly represented,
perhaps reflecting involvement in parasitic interactions15. EGF-like
domains (IPR006209) are one of the more common domains in
C. elegans and C. briggsae7,10. Although EGF-like domains were found
in other Rhabditina, dorylaim and panagrolaimomorph organisms
had the highest relative abundance. EGF-like domains are associated
with membrane-bound or secreted proteins involved in signaling
and recognition and may therefore be involved in manipulation of
host responses.
Analysis of InterPro domain representation in nematode-specific
genes identified a set of domains that may be important in parasitism.
In the Strongyloidea, the thirteenth-most abundant InterPro
domain is the allergen V5/Tpx-1 related domain IPR001283, found in
many secreted proteins, most notably in the Ancylostoma secreted
proteins (asp) that have immunomodulatory activity39. Additional
asp-like proteins are present in other clades15. Although IPR001283 is
found in organisms other than nematodes, genes containing this
domain have undergone lineage-specific amplification and divergence
in Strongyloidea. An abundant domain found in nematode-specific
gene families of particular prominence is the ‘transthyretin-like’
IPR001534: this family had 394 members from all species
(only lacking in Nippostrongylus brasiliensis), of which 377 had no
significant BLASTsimilarity outside the Nematoda. It is prominent in
C. elegans and C. briggsae. Mammalian transthyretins transport
thyroid hormones, and many of the nematode genes have secretory
leader polypeptides, suggestive of a role in hormonal signaling in
nematodes also.
To compare the biological functions of the genes associated with
each nematode, we used InterPro matches to assign Gene Ontology
terms40. The high-level Gene Ontology profiles for each clade were
very similar (Fig. 6), but we noted some differences between clades.
Dorylaim and panagrolaimomorph nematodes had a lower incidence
of structural proteins (Fig. 6a). Rhabditine nematodes had an elevated
number of predicted extracellular proteins, and pirurine, tylenchine
and rhabditine nematodes had an increased proportion of ribonuclear
proteins (Fig. 6c)20. The two groups with the highest proportion
of nuclear-localized predicted polypeptides were Tylenchina and
Dorylaimia. In both, parasite-secreted proteins are known to localize
to the host nucleus (Fig. 6c).
Metabolic pathway analyses
There is a general perception that parasites have lost function (under-
gone reductive evolution) as they came to rely on the metabolic
capacity and homeostatic buffering of their hosts. But many parasitic
nematodes spend part of their life cycle outside any host, or have
multiple phylogenetically and metabolically different hosts, and there-
fore may experience evolutionary pressure to maintain or even expand
metabolic and regulatory functions41. We compared the partial
genome of each species with the KEGG database42to determine the
extent of metabolic pathway representation (Supplementary Table 3
online). For most pathways, the number of enzymes associated with
each major clade correlated with the number of sequences generated
(Table 1 and Supplementary Table 3 online). The general congruence
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
c
Percentage
Extracellular
Membrane
Nucleus
Ribonucleoprotein complex
Chromosome
Ribosome
Mitochondrion
Endoplasmic reticulum
Cytosol
Cytoskeleton
Cytoplasmic vesicle
30
25
20
15
10
5
0
b
Metabolism
Transport
Development
Cellular process
Cell organization and biogenesis
Cell proliferation
Response to external stimulus
Cell communication
Response to stress
Cell growth and/or maintenance
55
50
45
40
35
30
25
20
15
10
5
0
Percentage
Spirurina (clade III)
Trichinellida (clade I)
Panagrolaimomorpha (clade IV)
Tylenchomorpha (clade IV)
Rhabditina (clade V)
a
Percentage
Nucleic acid binding
Motor activity
Metal ion binding
Nucleotide binding
Signal transducer activity
Binding
Lipid binding
Enzyme regulator activity
Transporter activity
Carbohydrate binding
Protein binding
Transcription regulator
Catalytic activity
Chaperone activity
Structural molecule activity
Molecular function unknown
30
25
20
15
10
5
0
Figure 6 Functional annotation of genes using Gene Ontology terms. Each sequence was compared with the InterPro database of domains and
these matches were used to assign high-level Gene Ontology terms. The data is summarized by major clade (see Fig. 1). The x axes show the
percentage of all the Gene Ontology terms for each assignment: (a) ‘molecular function’ assignments; (b) ‘biological process’ assignments; and (c) ‘cellular
component’ assignments.
NATURE GENETICS ADVANCE ONLINE PUBLICATION7
ARTICLES

Page 8

between the major clades suggested that many pathways are conserved
within the nematodes despite their diversity. Some differences were
noted, however. Spiruria and Dorylaimia had 17 enzymes (34 clusters)
from fatty acid biosynthesis pathway 1 (using acyl carrier protein-
bound precursors) but lacked pathway-2 enzymes (using coenzyme A-
bound precursors) completely, whereas Tylenchina had only pathway-
2 enzymes (44 clusters mapping to three enzyme types), and Rhabdi-
tina had both. No valine or methionine biosynthesis enzymes were
identified in the animal-parasitic Spiruria, suggesting that these may
be essential amino acids. N-glycan degradation enzymes were notably
abundant in Tylenchina, but less evident elsewhere. As the complete
genome of C. elegans encodes many N-glycan degradation enzymes,
this suggests that this pathway is particularly highly expressed in these
plant parasites. Enzymes involved in inositol metabolism were
also prominent in Tylenchina (and in the complete C. elegans
proteome) but absent in other sampled species. These predictions of
taxonomically restricted biochemical pathways may serve to direct
drug target definition.
DISCUSSION
We used B250,000 ESTs to predict more than 90,000 genes for a
suite of important human, animal and plant parasitic, and two free-
living, nematodes. Comparison of each species’ partial genome
with the complete genomes of C. briggsae and C. elegans and
with genome data from other phyla identified a spectrum of genes
and gene families, some of which were deeply conserved, others
were pan-nematode but nematode-unique, and others were taxono-
mically restricted. This data set aids the annotation of the C. elegans
genome in confirming gene predictions and identification by align-
ment of homologs of conserved and thus functionally important
residues. Highly conserved genes discovered in species across the
phylum may have important function in C. elegans, but some
such genes currently have no known RNAi phenotypes, perhaps
showing the limitation of on-plate assays. We identified tens of
thousands of potential targets for drug and vaccine development.
Many of these are nematode-specific but conserved across the phylum,
offering the prospect of new pan-nematode treatments. We have
deposited our sequence data in public databases as it is generated
and offered our analyses openly over the Internet since the inception
of the project, and so many of the genes we identified have already
been selected by other researchers in parasitology and C. elegans
biology for further study.
We look forward to further expansions of the nematode genome
data sets. We are still sequencing additional ESTs from target species,
and other projects, including enoplid and chromadorid taxa, are also
underway or planned. The genus Caenorhabditis will soon have five
nearly complete genome sequences, and the B. malayi genome is
nearing completion at The Institute for Genomic Research43. Genome
sequencing is planned for H. contortus, Meloidogyne hapla, P. pacificus
and T. spiralis. Our survey indicates that model species cannot show
the genetic and genomic diversity of even their own phylum, and that
continuing, phylogenetically informed genome sequencing is essential
for advances in genomics, evolution and infectious disease biology.
METHODS
Large-scale EST generation across the phylum Nematoda. We selected a
portfolio of nematode species based on criteria of phylogenetic spread,
availability of material and health, economic or scientific importance, after
consultation with the research and funding communities. The number of ESTs
sequenced for each species varied because of factors such as availability of
material, quality and source of libraries, and perceived importance of the
organism. For each species, we aimed to accumulate an EST data set spanning
multiple life cycle stage–specific and, where possible, tissue-specific cDNA
libraries. Previous experience with the filarial parasite B. malayi showed that
sampling from throughout the nematode life cycle was essential to maximize
rates of gene discovery44. To this end, we constructed 172 different cDNA
libraries, using various cDNA synthesis technologies and vectors (Supplemen-
tary Methods online). Members of the research community were generous in
providing both biological materials and libraries.
Sequencing and data processing. Sequencing and EST processing at the
Genome Sequencing Center was carried out as described17,18,45. Sequencing
and processing at the Wellcome Trust Sanger Institute was carried out as
described46. Before submitting them to dbEST, we processed the sequences to
assess quality, trim vector, remove contaminants and cloning artifacts, and
identify BLAST similarities using Genome Sequencing Center pipelines45and
the trace2dbEST pipeline47. All ESTs have been submitted to dbEST12.
Sequence clustering, annotation and database creation. For each species, we
downloaded sequences from dbEST in May 2003 and parsed them through
PartiGene, a software pipeline designed to analyze and organize EST data sets47.
We first checked sequences for contaminating vector sequence and trimmed
poly(dA) tails. We then clustered sequences into groups (putative genes) on the
basis of sequence similarity using CLOBB48. We assembled clusters to yield
consensus sequences using PHRAP (P. Green, unpublished data). We then
subjected each consensus sequence to a series of BLAST analyses49against a
suite of protein and nucleotide databases derived from public databases (see
Supplementary Methods online for details) and the thirty sets of consensus
sequences (partial genomes) for each nematode species analyzed here. We
defined significant matches as those having a raw BLAST score Z50 (this
corresponds to an expect value of 10?5to 10?6, depending on the size
and composition of the databases). Although this cutoff may miss some
homologous matches, it is sufficiently inclusive to identify most domain
matches. Results were processed and stored in a local installation of a
PostgreSQL database22,23.
Predicted proteome analysis: Gene Ontology, domains and metabolic path-
ways. For each consensus sequence, we obtained polypeptide predictions using
the prot4EST package in the PartiGene pipeline47. Predicted polypeptide
sequences were compared to InterPro (data version 7.0) to identify functional
domains using InterProScan36. An InterPro annotation was assigned to each
domain and translated into Gene Ontology40codes. These results were parsed
into a local installation of AmiGO40from which broader functional categories
were derived. Protein families were identified using TRIBE-MCL35using
default parameters.
To map predicted polypeptides to metabolic pathways, we compared them
using BLASTP to the KEGG database (version 29)42. We retained each match
meeting a cut-off of an expect statistic r1 ? 10?10. When one cluster matched
several closely related enzymes, we considered the top match and all the
matches within a range of 30% of the top score.
URLs. See http://www.nematode.net/ for more information on Genome
Sequencing Center trace files and clone ordering, and see http://www.nemato-
des.org/ for more information on Wellcome Trust Sanger Institute trace files
and clone ordering. PHRAP is available at http://www.phrap.org.
Note: Supplementary information is available on the Nature Genetics website.
ACKNOWLEDGMENTS
We thank all our nematology and parasitology colleagues for supplying materials
(see Supplementary Methods online) and for their enthusiasm for this project
and A. Anthony, J. Wasmuth and A. Hedley for trace2dbest, PartiGene and
prot4EST software. The UK arm of the project was funded by the Wellcome
Trust, and the US arm by the National Institutes of Health (National Institute of
Allergy and Infectious Diseases). J.P.M. was supported by a Helen Hay Whitney/
Merck fellowship. Sequencing at The Wellcome Trust Sanger Institute was carried
out by C. Churcher, T. Chillingworth, P. Cummings, Z. Hance, K. Jagels,
S. Moule and S. Whitehead. Most of the computational analyses were done
using facilities at the Center for Computational Biology, Hospital for Sick
Children, Toronto.
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
8ADVANCE ONLINE PUBLICATION NATURE GENETICS
ARTICLES

Page 9

COMPETING INTERESTS STATEMENT
The authors declare competing financial interests (see the Nature Genetics website
for details).
Received 19 August; accepted 15 October 2004
Published online at http://www.nature.com/naturegenetics/
1. De Ley, P. & Blaxter, M.L. Systematic position and phylogeny. in The Biology of
Nematodes (ed. Lee, D.) 1–30 (Taylor & Francis, London, 2002).
2. Platt, H.M. Foreword. in The Phylogenetic Systematics of Free-living Nematodes
(Lorenzen, S.) (The Ray Society, London, 1994).
3. Lambshead, P.J.D. Recent developments in marine benthic biodiversity research.
Oceanis 19, 5–24 (1993).
4. Blaxter, M.L. et al. A molecular evolutionary framework for the phylum Nematoda.
Nature 392, 71–75 (1998).
5. Chan, M.-S. The global burden of intestinal nematode infections - fifty years on.
Parasitol. Today 13, 438–443 (1997).
6. Barker, K.R., Hussey, R.S. & Krusberg, L.R. Plant and Soil Nematodes: Societal Impact
and Focus on the Future (Committee on National Needs and Priorities in Nematology,
Society of Nematologists, Marceline, Missouri, USA, 1994).
7. The C. elegans Genome Sequencing Consortium. Genome sequence of the nematode
C. elegans: a platform for investigating biology. Science 282, 2012–2018 ( 1998).
8. Wood, W.B. (ed.) The Nematode Caenorhabditis elegans 667 (Cold Spring Harbor
Laboratory Press, New York, 1988).
9. Riddle, D. Blumenthal, T., Meyer, B. & Priess, J. (eds.) C. elegans II 1222 (Cold Spring
Harbor Laboratory Press, New York, 1997).
10. Stein, L.D. et al. The genome sequence of Caenorhabditis briggsae: A platform for
comparative genomics. PLoS Biol. 1, E45 (2003).
11. Parkinson, J., Mitreva, M., Hall, N., Blaxter, M. & McCarter, J.P. 400000 nematode
ESTs on the Net. Trends Parasitol. 19, 283–286 (2003).
12. Boguski, M.S., Lowe, T.M. & Tolstoshev, C.M. dbEST - database for ‘‘expressed
sequence tags’’. Nat. Genet. 4, 332–333 (1993).
13. Lizotte-Waniewski, M. et al. Identification of potential vaccine and drug target
candidates by expressed sequence tag analysis and immunoscreening of Onchocerca
volvulus larval cDNA libraries. Infect. Immun. 68, 3491–3501 (2000).
14. Tetteh, K.K., Loukas, A., Tripp, C. & Maizels, R.M. Identification of abundantly
expressed novel and conserved genes from the infective larval stage of Toxocara
canis by an expressed sequence tag strategy. Infect. Immun. 67, 4771–4779 (1999).
15. Daub, J., Loukas, A., Pritchard, D.I. & Blaxter, M. A survey of genes expressed in
adults of the human hookworm, Necator americanus. Parasitology 120, 171–184
(2000).
16. Blaxter, M.L. et al. Genes expressed in Brugia malayi infective third stage larvae. Mol.
Biochem. Parasitol. 77, 77–96 (1996).
17. McCarter, J.P. et al. Analysis and functional classification of transcripts from the
nematode Meloidogyne incognita. Genome Biol. 4, R26 (2003).
18. Mitreva, M. et al. Comparative genomics of gene expression in the parasitic and free-
living nematodes Strongyloides stercoralis and Caenorhabditis elegans. Genome Res.
14, 209–220 (2004).
19. Mitreva, M. et al. Gene discovery in the adenophorean nematode Trichinella spiralis:
an analysis of transcription from three life cycle stages. Mol. Biochem. Parasitol. 137,
277–291 (2004).
20. Harcus, Y.M. et al. Signal sequence analysis of expressed sequence tags from the
nematode Nippostrongylus brasiliensis and the evolution of secreted proteins in
parasites. Genome Biol. 5, R39 (2004).
21. Whitton, C. et al. A genome sequence survey of the filarial nematode Brugia malayi:
repeats, gene discovery, and comparative genomics. Mol. Biochem. Parasitol. 137,
215–227 (2004).
22. Parkinson, J., Whitton, C., Schmid, R., Thomson, M. & Blaxter, M. NEMBASE: a
resource for parasitic nematode ESTs. Nucleic Acids Res. 32, D427–D430 (2004).
23. Wylie, T. et al. Nematode.net: a tool for navigating sequences from parasitic and free-
living nematodes. Nucleic Acids Res. 32, D423–D426 (2004).
24. Tatusov, R.L. et al. The COG database: new developments in phylogenetic classification
of proteins from complete genomes. Nucleic Acids Res. 29, 22–28 (2001).
25. Blaxter, M.L. Caenorhabditis elegans is a nematode. Science 282, 2041–2046
(1998).
26. Wang, D.Y., Kumar, S. & Hedges, S.B. Divergence time estimates for the early history
of animal phyla and the origin of plants, animals and fungi. Proc. R. Soc. Lond. B Biol.
Sci. 266, 163–171 (1999).
27. Reboul, J. et al. C. elegans ORFeome version 1.1: experimental verification of the
genome annotation and resource for proteome-scale protein expression. Nat. Genet.
34, 35–41 (2003).
28. Kamath, R.S. et al. Systematic functional analysis of the Caenorhabditis elegans
genome using RNAi. Nature 421, 231–237 (2003).
29. Barnes, T.M., Kohara, Y., Coulson, A. & Hekimi, S. Meiotic recombination, noncoding
DNA and genomic organization in Caenorhabditis elegans. Genetics 141, 159–179
(1995).
30. Simmer, F. et al. Genome-wide RNAi of C. elegans using the hypersensitive rrf-3 strain
reveals novel gene functions. PLoS Biol. 1, E12 (2003).
31. Gregory, W.F. & Parkinson, J. Caenorhabditis elegans – applications to nematode
genomics. Comp. Funct. Genomics 4, 194–202 (2003).
32. Parkinson, J. & Blaxter, M.L. SimiTri - visualising similarity relationships for large
groups of sequences. Bioinformatics 19, 390–395 (2002).
33. Aboobaker, A.A. & Blaxter, M.L. Hox gene loss during dynamic evolution of the
nematode cluster. Curr. Biol. 13, 37–40 (2003).
34. Roy, P.J., Stuart, J.M., Lund, J. & Kim, S.K. Chromosomal clustering of muscle-
expressed genes in Caenorhabditis elegans. Nature 418, 975–979 (2002).
35. Enright, A.J., Van Dongen, S. & Ouzounis, C.A. An efficient algorithm for large-scale
detection of protein families. Nucleic Acids Res. 30, 1575–1584 (2002).
36. Zdobnov, E.M. & Apweiler, R. InterProScan - an integration platform for the signature-
recognition methods in InterPro. Bioinformatics 17, 847–848 (2001).
37. Hutter, H. et al. Conservation and novelty in the evolution of cell adhesion and
extracellular matrix genes. Science 287, 989–994 (2000).
38. Johnstone, I.L. & Barry, J.D. Temporal reiteration of a precise gene expression pattern
during nematode development. EMBO J. 15, 3633–3639 (1996).
39. Hawdon, J.M., Jones, B.F., Hoffman, D.R. & Hotez, P.J. Cloning and characteriz-
ation of Ancylostoma-secreted protein. A novel protein associated with the
transition to parasitism by infective hookworm larvae. J. Biol. Chem. 271,
6672–6678 (1996).
40. The Gene Ontology Consortium. Creating the gene ontology resource: design and
implementation. Genome Res 11, 1425–1433 (2001).
41. Blaxter, M. Nematoda: Genes, genomes and the evolution of parasitism. Adv. Parasitol.
54, 102–195 (2003).
42. Kanehisa, M., Goto, S., Kawashima, S., Okuno, Y. & Hattori, M. The KEGG resource for
deciphering the genome. Nucleic Acids Res. 32, D277–D280 (2004).
43. Ghedin, E., Wang, S., Foster, J.M. & Slatko, B.E. First sequenced genome of a parasitic
nematode. Trends Parasitol. 20, 151–153 (2004).
44. Williams, S.A. et al. The filarial genome project: analysis of the nuclear, mitochondrial
and endosymbiont genomes of Brugia malayi. Int. J. Parasitol. 30, 411–419
(2000).
45. Hillier, L.D. et al. Generation and analysis of 280,000 human expressed sequence
tags. Genome Res. 6, 807–828 (1996).
46. Whitton, C., Daub, J., Thompson, M. & Blaxter, M. Expressed sequence tags: medium
throughput protocols. in Parasite Genomics (ed. Melville, S.E.) (Humana, New York, in
the press).
47. Parkinson, J. et al. PartiGene - constructing partial genomes. Bioinformatics 20,
1398–1404 (2004).
48. Parkinson, J., Guiliano, D. & Blaxter, M. Making sense of ESTsequences by CLOBBing
them. BMC Bioinf. 3, 31 (2002).
49. Altschul, S.F. et al. Gapped BLAST and PSI-BLAST: a new generation of protein
database search programs. Nucleic Acids Res. 25, 3389–3402 (1997).
© 2004 Nature Publishing Group http://www.nature.com/naturegenetics
NATURE GENETICS ADVANCE ONLINE PUBLICATION9
ARTICLES