pacaputo at CDS.UNINA.IT
Thu Nov 28 14:49:56 CST 1996
many apologies for cross-postings (taxacom & class-l).
I am presently involved in molecular systematic studies on
several plant groups. In particular, I am studying Internal
Transcribed Spacers (ITS) 1 and 2 of nuclear ribosomal DNA. The
reason of this message is the following: I use an alignment
software (ClustalV) which does not guarantee the best alignment.
Therefore, I usually run the software several times, varying the
parameters in a continuous fashion, and especially those related
to gap production. To choose among alignments, I run a parsimony
analysis on them and use homoplasy indicators (consistency and
retention indices) as optimality criteria. Now, I am afraid that
the optimal alignments ("homoplasy-wise") for ITS1 and ITS2
(analyzed separately), when combined may yield a suboptimal
solution (i.e., the lowest homoplasy in my ITS1 and ITS2 sets,
but not the lowest possible homoplasy in the "total" ITS set).
And here is the question: is it better to align separately
ITS1 and ITS2, choose the less homoplasious results (even if
obtained with different alignment parameters) and THEN combine
the matrices? Or should I align them together, with some 5.8S as
separator, delete it, and THEN choose the optimal alignment?
I can see "molecular evolutionary" reasons to do the former
and "phylogenetic" reasons to do the latter (obviously, I could
just align my sequences "by hand" and care not about homoplasy
Any suggestion (either to me - pacaputo at unina.it or
pacaputo at ds.unina.it - or to the list) will be gratefully
Dip. di biologia vegetale
Universita' di Napoli "Federico II"
Via Foria, 223
More information about the Taxacom