[Taxacom] FW: Fwd: Re: base alignment

John Grehan jgrehan at sciencebuff.org
Tue Aug 29 08:02:52 CDT 2006



> -----Original Message-----
> From: John Grehan
> Sent: Tuesday, August 29, 2006 9:02 AM
> To: 'pierre deleporte'
> Subject: RE: [Taxacom] Fwd: Re: base alignment
> 
> > -----Original Message-----
> > From: pierre deleporte [mailto:pierre.deleporte at univ-rennes1.fr]
> >
> > This is exactly what I mean, and this can only come from a previous
> > analysis at a larger level (check "outside" for the presence
/absence
> for
> > these characters), without entering into the details of the ingroup.
And
> > this can be done by hand, or with the help of compatibility analysis
in
> > computer programs ("unique to ingroup" exactly means "no homoplasy
> between
> > the ingroup as a whole and some outgroups" - just check by yourself)
> 
> I understand homoplasy to refer only to characters incongruent with a
> selected phylogeny. So if one has not done the phylogeny of
relationships
> within the ingroup yet, it is not possible to identify homoplasy for
> characters involved with ingroup relationships. Of course selecting
> derived character states requires a previous analysis of some kind to
> establish the outgroup and the monophyly of the ingroup.
> 
> > >hence the need for an analytical procedure to then
> > >determine which cluster of putatively derived characters is best
> > >supported (whether through parsimony or any other tree building
> > >procedure one might use).
> >
> > OK, but it would simply seem illogical that you impose "unique to
the
> > ingroup" at the beginning (constituting the data matrix of
compatible
> > characters considering the ingroup as a whole versus outgropups),
and
> that
> > you relax this condition when dealing with "inside the group"
analysis.
> > Do you really propose such a self-contradictory approach? Or will
you
> also
> > use compatibility analysis of the ingroup in order to fiond the best
> > supported clade on the basis of the larger clique of completely
> compatible
> > characters inside the group at stake?
> 
> Nothing illogical about restricting characters to those unique to the
> ingroup to study relationships between members of the ingroup. Of
course
> that rests on the assumption that the ingroup is monophyletic. Of
course
> that assumption rests on accepting evidence and it could be wrong. In
the
> case of the orangutan, this is an example of a genus that has widely
been
> accepted as the outgroup to humans and African apes. By analyzing the
> derived character states of orangutans, humans, and African apes
within a
> larger monophyletic clades (great apes), one finds that the orangutan
is
> not the outgroup to humans and African apes at all.
> 
> > Your not using a computer program does not help to understand
exactly
> what
> > is your procedure. But you an describe it as well: how do you
proceed
> for
> > the ingroup?
> 
> I don't need a computer program to recognize that there are 26
> substantiated derived features for humans and orangutans and only 8
for
> humans and both Africana apes, and apparently ZERO for humans and
> chimpanzees. However, using the same parsimony analysis used by those
> supporting the chimpanzee theory one can corroborate what one can
manually
> count. Its not published yet, but will be.
> 
> > Don't you think it could help that you use standard acception of the
> terms
> > instead? Like "compatibility analysis" instead of "cladistic
characters"
> > which makes no sense for any phylogenetician I can think of?
> 
> You make a lot of claims about what is "standard acceptance". Perhaps
you
> are right in such claims, perhaps not. Either way, my approach seems
to be
> ok with at least some cladists (or at least they have not indicated
the
> kind of confusion or objections that Pierre raises). Perhaps it is in
my
> nature, since I had no problem accepting Croizat's deviation from
standard
> (Darwinian) practice and terminology in biogeography.
> 
> > To the contrary, nobody in the Albert et al. book talks of
"cladistic
> > characters"... and nobody in "Cladistics" or "Systematic biology" as
> > well...
> 
> OK - then I am the exception. I am unique. Pierre will just have to
get
> used to it.
> 
> > >I don't have any lasting difficulties understanding cladistic
> literature
> > >or discussing with colleagues. People seem to understand my
position
> > >well enough, and visa versa. Whether or not we might agree or not
is
> > >another matter.
> >
> > I really don't think so.
> 
> Pierre and anyone else is welcome to disagree. I agree to disagree.
> 
> If so, this is clearly
> > inconsistent, because your results will possibly be different
according
> to
> > the level at which you are delineating the "ingroup" versus
"outgroups":
> > compatibility analysis outside, and standard cladistic inside.
> 
> Yes that is a possibility. However, it's pretty remote in the
orangutan
> case. I can appreciate that computationally things may get pretty
> complicated if one gets out a few taxa and the same character states
start
> turning up. This has been a problem where the outgroup for the human-
> African ape relationship has been limited to orangutans or gibbons. By
> taking in at least ALL the OW monkeys I have quite an extensive
outgroup.
> If I extend to all monkey, or all primates one has an even better
> representation. There are a few characters that turn up in the same
state
> once all primates are considered. For example, ischial callosities are
> absent from NW monkeys and prosimians and so this might be considered
a
> derived character along with humans and orangutans. I have chosen to
> regard the human and orangutan feature as derived. I suppose an
analytical
> theorist would not take that approach and use outgroup as well as
ingroup
> analysis at the same time. I admit that I do not get into such
> computational issues and will leave that for others.
> 
> > Otherwise you'd be logically consistent, isn't it?
> 
> In the end all methods are logically inconsistent. In the end the
issue
> for me is whether or not orangutans and humans are most closely
related or
> not, and whether the widespread assumption that DNA base similarities
are
> the last word on evolutionary replationships.
> 
> John Grehan


_______________________________________________
Taxacom mailing list
Taxacom at mailman.nhm.ku.edu
http://mailman.nhm.ku.edu/mailman/listinfo/taxacom



More information about the Taxacom mailing list