"Logic of Cladistics"

Mark Siddall mes at zoo.toronto.edu
Sun Jun 12 09:17:51 EST 1994


Erich Schwartz raises some interesting thoughts in art. 1779 regarding
what do we do about alternative methods of phylogeny reconstruction, that
I won't reiterate here for the sake of band-width.

In response, I hope I didn't leave the impression that one should ignore
different methods.  If only it were so easy...
In fact, quite the contrary, I would advocate as many perspectives on the data
as possible.
e.g., what does parsimony tell me?
      what does transversional parsimony tell me?
      what does the difference between using bases and using the implied
       amino acids tell me?
      what do distance methods tell me?
      to what extent are the various approaches defensible in my analyses?
      to what extent are they reliable?

Unfortunately, I do not think there's neccessarily a set answer or answers
to these queries.    I believe that it will have a lot to do with the 
type of data, and the taxonomic level of analysis.
For example, it may be that distance based methods are more reliable for
"deep-time" questions like Sogin's trees of life but that parsimony based 
approaches are more reliable for more localized investigations.
Undoubtedly, analyses should be viewed beyond the data at hand.

Using an example from Sogin (et al variously)... using distance-based 
methods, the diplomonads (incl. Giardia lamblia) fall out as the first
eukaryotic diverging lineage, but with parsimony the microsporidans
do.  Which is correct?
Going beyond the data at hand can help...
a) diplomonads have a basal-body oriented nuclear division while the
   microsporidans do not.
b) diplomonads have disjunct 5S and 24S rRNA genes, in microsporidans (like
    bacteria) they are in one larger gene that folds on itself after being
    transcribed.

This suggests to me that the diplomonads have synapomorphies with the 
other eukaryotes that the microsporidans do not and I would, thus, accept
the parsimony based solution over the distance based one.  That doesn't
mean that parsimony is going to be the best in all analyses, but the 
foregoing is an example of how one can seek some solution to the 
conflict of multiple methods by being scholarly, and in so doing, extracts
oneself from the trap of just being a gene-jockey pumping out trees and
throwing one's arms in the air when there's conflicting hypotheses.

Mark


-- 
Mark E. Siddall                "I don't mind a parasite...
mes at vims.edu                    I object to a cut-rate one" 
Virginia Inst. Marine Sci.                     - Rick
Gloucester Point, VA, 23062



More information about the Mol-evol mailing list