Time to update taxonomy? Article from BioMedNet [Snip w/link]

[John Noyes]

Dear All,

The current discussion on the use of sequencing to "speed up" taxonomy and
inventorying of the World's biota has been really interesting. However,
like most "traditional" taxonomists I have a few problems because I cannot
see for the life of me how "bar-code" taxonomy can possibly work. Surely
there can no rule or set of rules for separating species using DNA
sequences that are any better than those that we use for more "traditional"
approaches (morphology, behaviour, enzymes, etc.). Sequencing is just
another tool that can help us ikmprove how we define species and can be no
better than any other approach taken in isolation. Given this bar-code
approach I can only see that we will get ourselves very, very unstuck with
the result that we will eventually have to return to the more traditional
approaches. What do we do with sequences that are not an exact match
(pretty nearly 100%)? Do we have an arbitrary figure that says if the
sequence is X% different then it must be a different species? Some species
of fungi can be separated on a single base pair difference in a particular
gene, whilst in many insects even a 5% difference in sequence may be
insufficient to separate the species with any certainty. Do we have an
arbitraty assumption that if there are differences at certain codon
positions (or any other position for that matter) then it is a different
species? I can see that it might work if we have a fully refined phylogeny
which is incredibly robust (i.e. approaching total stability whatever we
throw at it). Yes, it might work with bacteria or other "lower" organisms
because in many cases there is precious little that is any better. Maybe
someone out there can explain to us traditionalists in a few sentences how
this method WILL work.

In an article published in a recent issue of TREE it was said that 1000
samples could easily be processed in a one day. Maybe that can done using
an automatic sequencer that can handle multiple samples at a time. However,
how long would it take to prepare these samples before they are placed in
the sequencer and how long would it take to check and correct the output (I
suppose pretty soon that will be totally automated as well). More to the
point, who would do it? It looks pretty brain numbing work to me. I am not
sure that this method would be more cost effective or faster than the more
tradional approaches in many groups. If we go down this automated route
then we risk losing those with the taxonomic knowledge of a group. If we
take Lepidoptera for instance. Anyone who is half decent can sit down at a
light trap in the UK and sort out several hundred specimens to species (and
record numbers of each species present) within a half hour (including
microleps) or so without the need to kill any specimens. Of course there
would be a very small proportion of difficult specimens that would have to
be examined in more detial and, heaven forbid, here it might be quicker and
possibly cheaper to use sequencing, although I doubt it. If all this was
done using the approach proposed in TREE then it would take several days to
analyse the one sample (from start to finish) and it would mean that every
specimen would have to be killed. I suppose that if this were the case then
maybe the whole system could become totally automated from start to finish.
But how much would a machine like this cost? How much would it cost to run?
Compare this with one light trap and maximum of 1.5 hours a day for the
"tradional" lepidopterist. No contest I think.

My final point is this figure of a complete inventory of life in 20 years.
That is completely idiotic. It would not be possible even with completely
automated DNA taxonomy and totally unlimited funds. We have been surveying
the Chalcidoidea (Insecta: Hymenoptera) of Costa Rica intensively for the
past 18 years. In that time we may have discovered no more than about 20%
of the species that occur there (optimistically not more than 50%). Costa
Rica is a very small country.

This has become a real waffle. Sorry I shall stop there.

John

At 02:14 AM 1/31/2003 -0500, Jim Bass wrote:
>The summary article below may be a bit more accessible for
>some than the Vol. 18, No. 2, February 2003  issue of TREE
>suggested by Rich Pyle on 20 January.
>Jim
>_________________________________________________________
>Time to update taxonomy?
>BioMedNet
>by Henry Nicholls
>
>Taxonomy is stuck in the 18th Century, claims Canadian zoologist
>Paul Hebert, who calls for a molecular revolution. "There is no more
>likely death of a discipline than the failure to innovate," he warns.
>
>With the right funding and support from the taxonomic community,
>Hebert, professor of zoology at the University of Guelph, predicts
>that using DNA to identify and help categorize new species would
>make a complete inventory of life a feasible goal in as little as 20 years.
>
>Traditionally, species have been identified and categorized on the basis
>of morphological differences, but parts of the academic community are
>becoming increasingly concerned that "morphological taxonomy",
>pioneered by Linnaeus in the 1750s, is no longer equipped to account for
>the extent of earth's biodiversity, frequently estimated at between 10- and
>15-million species.
>
>"Let's face it, the morphological approach has had 250 years to advance
>the task, and we're only 10% of the way towards the goalposts," Hebert
>told BioMedNet News. His solution, to which many taxonomists are
>strongly opposed, is to distinguish between species on the basis of
>similarities and differences in their DNA.
>
>A DNA-based taxonomy would vastly accelerate the final inventory of life,
>says Hebert. Each novel organism would be described rapidly and simply
>by a "barcode" - a number that corresponds to a sequence of its DNA. Hebert
>envisages this "gene species" as a first, mandatory step towards describing
>a
>real species. At a later stage, traditional taxonomists could make the
>formal
>morphological description of the specimen, which would then become
>associated
>with its DNA barcode, he suggests.
>
>Although most taxonomists support a role for DNA in taxonomic research, some
>have expressed strong concerns about any move that would make DNA sequencing
>a mandatory step in the description of a new species.
>
>Many think that describing a species using DNA is expensive and unnecessary.
>"In most instances, a quick survey of morphology will serve the same
>purpose," write
>Ole Seberg, professor of evolutionary botany at the University of Copenhagen
>in
>Denmark, and colleagues in a letter to be published in the February issue of
>Trends in
>Ecology and Evolution (TREE).
>[Snip]
>http://news.bmn.com/magazine/specialreport?uid=SREP.030109-1
>http://news.bmn.com/magazine/specialreport?uid=SREP.030109-1
>
>

Entomology Department, The Natural History Museum, Cromwell Road, South
Kensington, London, SW7 5BD, UK

Tel. +44 (0)207-942-5594  Fax: +44 (0)207-942-5229

INTERNET: jsn at NHM.ac.uk

Interactive catalogue and biological database of World Chalcidoidea on CD:

Contact the publisher: DickyS._Yu at telus.net, or Dicky S. Yu, P.O.Box 48205,
Bentall Centre, Vancouver B.C., V7X 1N8, CANADA;
see also www.taxapad.com

Encyrtidae of Costa Rica:
http://www.nhm.ac.uk/science/intro/entom/project2/index.html

Course on taxonomy and biology of parasitic Hymenoptera:
http://www.nhm.ac.uk/entomology/hymcours