IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

Am I a pattern or transformed cladist?

fatherdes at hotmail.com fatherdes at hotmail.com
Wed Feb 3 11:51:16 EST 1999

Joe F. has enlightened us all with an exhaustive list of the various types of
hat that a cladist might wear.	In practice, there are many people who wear
cladist hats who are quite sane and normal and not at all as belligerent as I
indicated.  You find them in museums (and no I am not being snide (for once))
where they attempt to classify things and for whom, the work of Willi Hennig
was a significant breakthrough.  These people try to find monophyletic groups
(sometimes a hard thing to do) and look for synapomorphies to help do this
and try to avoid other types of misleading or uninformative characters. 
Hennig's work provided a framework for doing this and a pile of jargon to
describe the things to look for when you do it.  It is not always easy to use
and sometimes you end up just as confused as before Hennig but at least you
now have a logical framework to try and operate in.  Before that you kind of
hand waved.

Then the trouble started.  There was a group of systematists called
pheneticists (most famously Sokal and Sneath) who, since the late fifties,
tried to disentangle phylogeny from classification and who said you should
classify things based on some notion of overall similarity.  It turned out to
be quite complicated in practice but it stimulated a lot of discussion and
the invention of some useful new methods (including UPGMA).  These days, few
people use phenetics as such but the methods are still very useful sometimes
(e.g. as part of the multiple sequence alignment process) as long you see
them for what they are (and aren't).  Then a different group arose, lead by a
bloke called Farris, from the same institute as Sokal (at Stony Brook) who
agreed with the pheneticists that phylogenies were probably too difficult to
go for directly and that one should treat classification as an end in itself.
 Where they differed was with regard to methodology; they preferred to use
what they described as "character based" methods and especially maximum
parsimony.  This group also produced some very useful results and methods,
some of which are still in use today.  They became known as transformed or
pattern cladists.

Now, at this stage you might say that these groups differed only in some
methodological niceties but here followed a row of mammoth proportions which
still rumbles today.  The groups became mortal enemies and took to the
streets burning each others images (metaphorically) and questioning the
parentage and even taxonomic group of each other.  (I was a pheneticist (as
if you could not guess) and did not really care enough to get involved and I
was only a PhD student at the time and got sick of it.	It became very silly
and ended up as an esoteric personality war (with the good Joe felsenstein
cast in the role of a special horned one for daring to explain some of what I
have tried to explain here (he did it much better)).

I have to insert a comment here to point out that most people who use and
advocate parsimony are quite sane and can give very good reasons for liking
the method.  David Swofford is a good example.	Personally, I have sometimes
found parsimony to be the tool of choice.

I left systematics to do molecular evolution and sequence analysis, where I
thought I would leave all the silliness behind.  A new complication had
arisen which was the development of methods for generating large volumes of
DNA sequence data.  Many of the people making trees did not know what a
cladist was, did not care and just wanted phylogenies.	They used (thanks to
the Phylip package) UPGMA, NJ, ML, Parsimony and others with much
bootstrapping as they felt was appropriate.  The pattern cladists felt left
out but they had spent so long fighting the pneneticists (now defunct) that
they did not realise the new revolution was so far progressed and they were
suddenly isolated and ignored by the mainstream of tree makers.  Ironically,
I was able to use UPGMA to help make crude trees which were used to make
multiple alignments which were then used to make better trees.	This was a
vicious insult to the memory of all the pattern cladists who had gone insane
on the battlefield.

You still get lingering after shocks.  I have written papers where we did NJ,
ML and Parsimony trees but we showed the NJ one because the branch lengths
and bootstraps looked nice but said in the figure legend that the other trees
were identical in topology.  I have had referees say that I should still show
the Parsimony tree or that I should not have done NJ at all.  Farris, when I
last heard of him, was somewhere in Scandinavia, and his influence still
trickles out to this newsgroup to this day.

Des "sorry that was a bit long" Higgins

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

More information about the Mol-evol mailing list

Send comments to us at biosci-help [At] net.bio.net