Symbol comparisons in PRETTY
vjongene at eliot.unil.ch
Mon Jul 10 09:03:41 EST 1995
I was trying to get PRETTY (GCG) to do a relatively simple thing, viz.,
printing an alignment from a .MSF file with a consensus at the bottom, and
each sequence in the alignment with '-' at positions where it matches the
consensus, or the amino acid code at positions where it doesn't. This is in
principle possible using the /CON and /DIF='-' command-line switches.
The output however is not quite what you would expect. It seems that the
consensus is calculated using a very permissive comparison table, and that
only one member of the permitted values in the consensus is displayed. For
example, at one position PRETTY decided that W and R both matched the
consensus, which was given as W, while the sequences containing R did just get
a '-' to indicate a match. This of course is unacceptable.
So the question(s): Is it the PrettyPep.Cmp file that does this? Is there a
relatively easy way to change this behavior? Why was this default matrix
chosen in the first place? (I can't think of a single good reason myself).
I'll post an example if the description above turns out to be insufficient.
Thanks to all for feedback/help.
|C. Victor Jongeneel, PhD, Associate Member |
|Ludwig Institute for Cancer Research |
|CH-1066 EPALINGES Switzerland |
|Tel: +41-21-692-5994 FAX: +41-21-653-4474 |
|E-mail: CornelisVictor.Jongeneel at isrec.unil.ch |
More information about the Bio-soft