Mon Dec 12 09:41:27 PST 2005 Kevin Karplus

We should do one-step calibrate and find hits, if hmmscore supports it now.
If not, calibrate on smaller fixed set.

Sun Dec 11 20:26:36 PST 2005 Kevin Karplus

Base the characters per row in the logos on the length of the protein,
so that <300 use 50,  <1000 use 100, longer use 200.

Note: there is a bug in the scripts for the gromacs optimization when
START_COL is not 1, in which the residues get renumbered to start at 1.

We may need to increase the weight for maybe_ssbond.


BELOW THIS LINE HAVE BEEN DONE
------------------------------------------------------------

Fri Nov 25 15:23:18 PST 2005 Kevin Karplus

When starting from a seed alignment the "all-align.a2m" file has lots
of copies of the non-guide sequences, which causes some problems for
undertaker, which assumes that they are supposed to be PDB files.

This could be fixed by creating a "guide.a2m.gz" file that is the
first sequence from ${TARGET}.a2m and usig it instead of ${TARGET}.a2m
in the hmmscore calls that make the pairwise alignments.

[FIXED 25 Nov 2005]