There are an estimated 10 15 millions species on Earth Hammond 1992 of which only about 1,8 millions

1
Character-based DNA barcoding Accurate
identification of conservation units in
odonates J. Rach1,2, B. Schierwater1,2, R.
DeSalle3, I.N. Sarkar3 H. Hadrys1,2 1ITZ,
Ecology Evolution, TiHo Hannover, Bünteweg 17d,
D-30559 Hannover, Germany 2Ökologische
Forschungsstätte Bahnhof Schapen, Lindenallee 20,
38104 Braunschweig 3American Museum of Natural
History, Division of Invertebrate Zoology, New
York, NY, USA
Introduction There are an estimated 10 15
millions species on Earth (Hammond 1992) of which
only about 1,8 millions have yet been described.
In addition, human impact - e.g. defragmentation
or destruction of habitats - results in a steady
increasing number of endangered species.
Evidently, many species become extinct without
having been described. Consequently, the
development of tools for the fast identification
of organisms becomes a sine qua non. DNA
barcoding has become popular as a rapid and
reliable technique for species identification. In
addition to morphological keys, a public
reference library containing short DNA sequences
of all known species should enable scientist to
identify specimens quickly and accurately and to
discover new species. Dragonflies are
excellent indicators for all kind of freshwater
ecosystems due to substantial differences in
habitat specificity and their complex life cycles
(Hadrys et al.2006). Here we present a new
approach to achieve the accurate identification
of dragonflies the character-based DNA
barcoding. By means of the CAOS algorithm we
find combinations of character states within a
short DNA fragment that enable the identification
and differentiation of conservation units in
dragonflies (Figure 2-3). Since these
character-based DNA barcodes are valid for all
life stages, we suggest that this technique will
be highly valuable for conservation biology.
(a)
(b)
(c)
(d)
Figure 1 Identification of dragonflies through
phenotypic traits is difficult (a) Females of
the same genus show a high level of ressamblance
(b) Wing veneation as identification key requires
a lot of experience (c) Identification through
ecological and behavioural patterns is difficult
and time-consuming (d) Morphological keys mostly
do not consider different life stages (Hebert et
al. 2003)
Methods PCR was performed with gene specific
primers to amplify an approximately 560 bp long
fragment of the mitochondrial ND1 (NADH
dehydrogenase 1) gene region for 833 odonate
samples. A Neighbor-Joining (NJ) tree was
generated using PAUP (version 4.0b10) as basis
for the application of the P-Gnome software that
is driven by a Columbia University Patent Pending
Algorithm, the Characteristic Attribute
Organization System (CAOS) (Sarkar et al. 2002).
By means of the CAOS algorithm a search for
diagnostic characters within the ND1 sequences
for every clade at each node within the NJ tree
was performed. Further analyses were assisted by
a set of scripts written in perl. Finally,
combinations of diagnostic characters found for
taxonomic groups were used for the establishment
of character-based DNA barcodes.  
Figure 2 Character-based DNA barcodes for
species shown are unique combinations of
character states within short DNA sequences (ND1
gene region) for the discrimination of odonate
species
Results Discussion By means of the CAOS
algorithm we obtained reliable character-based
DNA barcodes for most of the analysed genera and
species. Within many species high genetic
distances between geographical groups were
detected. In these cases, we also found
diagnostic characters for the differentiation of
single populations (Figure 3). The latter is
particularly important for conservation
management. Information about the locality and
the number of diagnostic characters found for
single populations give important hints about the
geographical distribution of a species and the
level of gene flow between populations.
Character-based DNA barcodes are combinations of
character states within a short DNA sequence that
are unique for taxonomic groups. Thus, the
accurate identification of specimens can be
achieved easily through regarding only a few
nucleotide positions of an appropriate barcode
marker.
(b)
Figure 3 Character-based DNA barcodes are able
to identify conservation units (a) All four
populations of Pseudagrion bicoerulans formed
separate groups in the Neighbor-Joining tree (b)
Numerous diagnostic characters for each
population were found through analysing the NJ
tree with the CAOS algorithm Unique
character-based DNA barcodes were obtained by
combining the character states at nucleotide
positions at which pure diagnostic characters for
at least one population was found
References Hammond, P. 1992 Species inventory.
In Global biodiversity Status of the earth's
living resources. pp. 17-39. Hebert, P. D.,
Cywinska, A., Ball, S. L. deWaard, J. R. 2003
Biological identifications through DNA barcodes.
Proc Biol Sci 270, 313-21. Rach, J., DeSalle, R.,
Sarkar, I.N., Schierwater, B. Hadrys, H. 2006
Character-based DNA barcoding Discrimination of
genera, species and populations in Odonata by
means of the CAOS algorithm. Submitted Hadrys,
H., Clausnitzer, V. Groeneveld, L. V. 2005 The
present role and future promise of conservation
genetics for forest Odonates. In Forest and
Dragonflies (ed. A. Rivera). Pontevedra, Spain
Pensoft Publishers Sofia-Moscow. Sarkar, I. N.,
Thornton, J. W., Planet, P. J., Figurski, D. H.,
Schierwater, B. DeSalle, R. 2002 An automated
phylogenetic key for classifying homeoboxes. Mol
Phylogenet Evol 24, 388-99. Samples were
collected in southern France (Crau/Carmargue) and
Germany/ Lower Saxony (Schapen).