[Taxacom] Is this the future?

JF Mate aphodiinaemate at gmail.com
Fri Jul 19 00:20:22 CDT 2013


When all you´ve got available to bioblitz is ecologists, a dystopian future
may appear to be your best hope. Having said that there are serious issues
with metagenomics in biodiversity studies (ecogenomics is really the same
but sounds cool and new, eminently more fundable), namely:

Sensitivity: unless they are planning to clone and sequence every copy of
sequenced marker they will miss a heck of a lot. Further, metagenomics
doesn´t pick up suff with frequencies below 0.01 and as anybody who has
gone light-trapping or sweeping knows, most species are rare. In their
sample of 300,000 specimens they can expect at least 3-5000 species, and I
expect a quarter will be singletons.

Genebank: it is not up t the task of matching taxa. Little taxonomic
coverage, many misidentifications, hardly at a stage where you can throw a
sequence and get a generic ID (forget specific).

New species: knowing the previous two problems, how will you know that
sequence XYZ corresponds to a new species? How do I know that this copy of
COX1 is sufficiently distinct from this other one to claim, based solely on
these 800bp, that it is? Even if in 30 years we have sequenced all
described taxa, how will the alphabet soup help accelerate the process?

"Ecogenomics" is the latest incarnation of morphospecies/parataxonomy,
something (you hope) any freshman can do over the summer break.
The sad thing is that metagenomics is truly useful for microbiologists
(most microbes cannot be easily grown) but they are working with fairly
simple communities and, at this stage, very coarsely grained information.
The whole purpose of the process is nicely summed up by this quote:

"The speed and accuracy at which we'll be able to assess a given site, even
a really remote one, will greatly improve the efficiency of evaluating
possible biodiversity impacts,"

It is definitely fast, efficient, but it will not be accurate. But then, I
don´t think there was the intention anyway. If your interest is only go as
far as counting beans, it really doesn´t matter what the beans are. Faba,
lima or chick peas are all the same to you as long as you know that site A
has 10% more beans than site B. In which case, old school parataxonomy will
work as well.

Best

Jason

P.D. maybe there is no cause for concern as most species have been
described anyway ;)



More information about the Taxacom mailing list