This program allows you to tidy up gaps that are not meaningful in the context
of a multiple sequence alignment derived from structure. In other words, regions
that are not similar across all members of a structural family can be `cleaned' to
remove isolated residues aligned in the middle of nowhere. Note that one doesn't always
want to do this (since the sub-alignments can be meaningful).
The format is:
stamp_clean <stamp alignment file> <minimum segment length> > <output file>
The 
minimum segment length
is the minimum number of residues that is to
be considered significant. I always use 3, since this means that short stretches of
1-2 residues that are surround by gaps (i.e. in any sequence) are `cleaned'. Try it
and see what I mean.