Apologies for troubling the list twice in one week with questions
about the python interface - but John K's last answer was so
helpful ...
Anyway, the short version of the question is: does anyone know how to set a
parameter for a constraint type when the python interface doesn't seem to
provide access to it? (here, for example the "ceiling" for NOE constraints)
In more detail: we're having problems with a script based on
eginput/gb1_rdc/anneal.py (xplor-nih 2.11) not leading to models that
satisfy long range constraints. The local structure is good, satisfying
the short-medium range constraints fairly well. The long range
constraints (admittedly sparse, initially) appear to be read in,
and the violations (20 - 35 Angstroms!) are listed in the analysis
section of the log-file, eg:
596 (1087 73 TYR HB1) (161 10 TYR HE1) 5 40.42 35.42 0
with a summed NOE energy for all ~1400 noe constraints of ~175
We've increased the scale factors (ultimately to ridiculous levels),
with no effect. The next guess was that the ceiling might be set too
low? A fairly analogous script in the xplor interface, with a ceiling
of 1000 brings the distant parts of the structure together, satisfying
the l.r. constraints. So, we tried setting the ceiling, with things
like:
noe.ceiling = 1000
or
xplor.command('''
noe class %s
ceiling = 1000
end''' % noe_table)
with no obvious effect. Will next try setting up the NOEs entirely with
xplor.command as John K described earlier in the week, or revert to
anneal.inp, but wondered if there was a way to do it that would retain
the benefits of the python module handling the constraints?
Thanks,
- Mark
