A lot of ink is really spilled over maximum parsimony (MP) and maximum
likelihood (ML)... I am a simple user of the ML method and I do not feel
authorized to discuss the subtleties of the method. You have not to
understand how all things work together inside a computer for using it
(more or less efficiently). I would advice people interested in MP and ML
"philosophy" to direct more specific questions to specialists such as
Swofford, Felsenstein and others. However, there are some issues
addressed by K. Fitzhugh which deserve a short reply, within the limits of
my ability.
It is true that maximum parsimony is not a method but rather a philosophy.
Interestingly, I know that some people are prone to consider MP as a
special case of ML but I will not venture discussing what is more related to
"le sexe des anges" (to count how many angels can dance on the head of
a pin).
>It is also worthwhile noting that "statistical" probability is not
>applicable to historical events. To observe a set of shared similarities,
>as effects observed in the present that are due to causal events in the
>past, one is attempting to infer what causal events best explain those
>similarities. But, shared similarities are either due to common ancestry or
>they are not. The "statistical" probabilities are either 0 or 1. Either
>some causal event per some theory was the case or some other event
>occurred. The only form of probability that applies here is logical
>probability.
It is, of course, clear that the statistical probability for an event to occur in
the past is either 0 or 1. It happened or not. However, the ML method has
nothing to do with that. It computes the probability of a tree (the
hypothesized history) to give rise to the observed data according to the
evolutive model considered.
The problem of parsimony which can be misleading has been widely
discussed in the literature. It seems clear that when there are "long-
branched" lineages in a phylogenetic tree, parsimony is prone to
spuriously group them together, whatever their historical relationships. This
has been demonstrated many times and I would refer the reader to these
papers (a.o. Felsenstein, 1978; Kim, 1996; Poe & Swofford, 1999; ...)
Felsenstein, J. 1978. Cases in which parsimony or compatibility methods
will be positively misleading. Syst. Zool. 27: 401-410.
Kim, J. 1996. General inconsistency conditions for maximum parsimony:
effects of branch lengths and increasing numbers of taxa. Syst. Biol. 45:
363-374.
Poe, S. & Swofford, D. L. 1999. Taxon sampling revisited. Nature 398: 299-300.
A last remark about the "position 556":
>Let's say one has a set of specimens, from which they "observe" that
>position 556 of each specimen has an A. Our perceptions tell us we "see"
>the same nucleotide among these individuals. Most systematists then go so
>far as to say, "this A at position 556 indicates homology." If one asks the
>definition of homology, they might respond, "shared similarity due to
>common ancestry."
There is here a misunderstanding. In a sequence alignment, homology
concerns positions, not bases. But it is true that most alignment algorithms
make first assumptions about homology of bases in order to infer homology
of positions... Lastly, I must confess that all the discussion about difficulties
with invoking a rate-dependent model for inference of cladograms is not
clear at all for me. This unfortunately limits my ability to go further in the
discussion...
Patrick MARTIN
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Patrick MARTIN
Institut royal des Sciences naturelles de Belgique
Biologie des Eaux douces
29, rue Vautier
B-1000 Bruxelles
Belgium
Tel : +32/2/627.43.17
Fax : +32/2/627.41.13
Email : martin at kbinirsnb.behttp://www.kbinirsnb.behttp://eea.eionet.eu.int/ec-chm/cbro/country/Belgium.htm#127
-- ANNELIDA LIST
Discuss = <annelida at net.bio.net> = talk to all members
Server = <biosci-server at net.bio.net> = un/subscribes
Archives = http://www.bio.net/hypermail/annelida/
Resources = http://biodiversity.uno.edu/~worms/annelid.html
--
