[Taxacom] Barcoding and large samples

[Richard Zander]

Judging from offline queries, there is a question that barcoding is
demanded to demonstrate higher standards than plain old taxonomy when
involved in new species descriptions. Why should that be?  Good
question. Basically, using DNA implies in this day and age considerably
more reliability than that which comes from alpha taxonomy. 
 
To explain: Alpha taxonomy has over time generated many new taxa with a
range of reliability from maybe a little bit more likely than not, to
pretty certain. False positives are many judging from the synonymies in
the literature, say 50%. That's tolerable because we acknowledge it and
are proud of the good stuff and work like hell on revisions. New taxa
nowadays are based on what we've learned when checking for false
positives, and are fairly reliable if we have large samples to study, or
find something way different from related taxa.
 
With cladistics and a philosophical assertion that the results of
cladistic analysis converge on Truth, or that scientists always accept
the simplest most parsimonious hypothesis, we have encouragement to rely
on computerized analysis of relationships rather than seat-of-pants
judgments. Plausible.
 
Though cladistics has little to do with identifying new species,
systematics now expects more reliability in its results. With the use of
DNA data and development of the bootstrap in phylogenetics by
Felsenstein in 1985, branch arrangements of less than 50% probability
are generally discarded. 
 
With Bayesian methods, DNA-based results are commonly near to or at
statistical certainty. Could be, could be, and some of us are following
such results in our classifications.
 
The use of DNA in barcoding implies reliable results in distinguishing
new taxa, simply by association with other uses of DNA in systematics.
It is necessary for barcoders, however, to actually demonstrate that
false positives are very few. If a 95% confidence level is proposed with
expectation of an average of 19 correct barcode projections of new
species per 20 tests, then many more tests that 20 are needed to
establish this, because the expectation of 19 correct is only an
average. Even if more-than-half correct (not false positives) is the
goal, demonstrating this requires a fairly large sample with morphology
(or evolutionary ecology) checked to test for false positives. 
 
Thus, in a group analyzed by barcoding without a large-sample statistic
demonstrating the ability of barcoding to identify new species in that
group, a morphological analysis for each specimen barcode-flagged as new
to science is absolutely necessary.
 
 
******************************
Richard H. Zander 
Voice: 314-577-0276
Missouri Botanical Garden
PO Box 299
St. Louis, MO 63166-0299 USA
richard.zander at mobot.org
Web sites: http://www.mobot.org/plantscience/resbot/
and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
For FedEx and UPS use:
Missouri Botanical Garden
4344 Shaw Blvd.
St. Louis, MO 63110
******************************