AFLP reliable for study of evolution?

[baumel]

 I'm agree with David Baum on the difficulties encountered using any
multilocus DNA markers (AFLP, RAPD, ISSR ...) for phylogenetic studies.  At
the interspecific level the problem is the homologies of comigrating bands,
and at the intraspecific level another problem is the homologies of missing
(two different mutations can lead to the same absence state).  So the
advantages of these markers is reduced because you have to chek homology,
if you are able to do it !  (How can you chek homology of two missing bands
?)

        I'm also agree with David for the danger of mixing genes trees and
OTU trees, but this is the problem of all the nuclear markers when you are
studying reticulate evolution of population or species (polyploid complex
for example).  However, I think that these markers are still useful for
study of closed species or population with the caution that you have to
increase the number of primers (so the number of locus) and to use a
phenetic analysis.  Because these markers are scattered in all the genomes,
I think (or I hope) that they are a good complementary tool when used with
sequences analysis of nuclear or cytoplasmic DNA.
        In the case of a work dealing with a large number of OTU where
reticulate evolution is possible (a  plant genus for example), PCR based
DNA markers are an important step to clear up the relations of closed
species and hybridization patterns.


        Many thanks for all that help me to clarify my troubles with PAUP
and Sequap, I'm going to test the huge number of tips that you send me.







        Alex BAUMEL
        Service de Botanique
        UMR 6553 "Ecobio"
        Campus de Beaulieu, Bat. 14, 35042 RENNES CEDEX
        FRANCE