Hi,
Is there any way to force Modeller to generate restraints for consecutive CA atoms instead for every other one? What is the reason for excluding i, i+1 restraints for CA atoms?
Thanks,
Mensur
========================================================================== | Mensur Dlakic, PhD | Tel: (406) 994-6576 | | Department of Microbiology | Fax: (406) 994-4926 | | Montana State University | | | 109 Lewis Hall, P.O. Box 173520 | http://myprofile.cos.com/mensur | | Bozeman, MT 59717-3520 | E-mail: mdlakic@montana.edu | ==========================================================================
Mensur Dlakic wrote: > Is there any way to force Modeller to generate restraints for > consecutive CA atoms instead for every other one? What is the reason for > excluding i, i+1 restraints for CA atoms?
The i, i+1 CA-CA restraint (also the N-O and sidechain-sidechain restraint) is excluded because it would conflict with the stereochemical backbone restraints, but you can certainly turn it back on. Just modify or override the mkhomcsr method in the automodel class (modlib/modeller/automodel/automodel.py). The ranges for all three of these restraint types are currently set to (2, 99999). Just change them to (1, 99999).
Ben Webb, Modeller Caretaker
Thanks for your prompt response. I have already played with automodel.py, although by changing different parameters, and could not get this to work. In automodel.py, when I change
for (dmodel, maxdis, rsrrng, rsrsgn, rsrgrp, typ2, typ3, stdev) in \ ((5, self.max_ca_ca_distance, (2, 99999), True, 9, 'CA', 'CA', (0, 1.0)),
to
for (dmodel, maxdis, rsrrng, rsrsgn, rsrgrp, typ2, typ3, stdev) in \ ((5, self.max_ca_ca_distance, (1, 99999), True, 9, 'CA', 'CA', (0, 1.0)),
the program still makes the restraints only for i, i+2 CA atoms. Is there an additional over-ride for this setting in a different script? Or is it by now hard-coded into the program regardless of the script setting?
I just tried changing these numbers in Modeller 6v2, still makes the restraints only for i, i+2 CA atoms.
Thanks,
Mensur
At 11:40 AM 5/9/2006, you wrote: >Mensur Dlakic wrote: >>Is there any way to force Modeller to generate restraints for consecutive >>CA atoms instead for every other one? What is the reason for excluding i, >>i+1 restraints for CA atoms? > >The i, i+1 CA-CA restraint (also the N-O and sidechain-sidechain >restraint) is excluded because it would conflict with the stereochemical >backbone restraints, but you can certainly turn it back on. Just modify or >override the mkhomcsr method in the automodel class >(modlib/modeller/automodel/automodel.py). The ranges for all three of >these restraint types are currently set to (2, 99999). Just change them to >(1, 99999). > > Ben Webb, Modeller Caretaker
========================================================================== | Mensur Dlakic, PhD | Tel: (406) 994-6576 | | Department of Microbiology | Fax: (406) 994-4926 | | Montana State University | | | 109 Lewis Hall, P.O. Box 173520 | http://myprofile.cos.com/mensur | | Bozeman, MT 59717-3520 | E-mail: mdlakic@montana.edu | ==========================================================================
Mensur Dlakic wrote: > Thanks for your prompt response. I have already played with > automodel.py, although by changing different parameters, and could not > get this to work. In automodel.py, when I change > > for (dmodel, maxdis, rsrrng, rsrsgn, rsrgrp, typ2, typ3, stdev) > in \ > ((5, self.max_ca_ca_distance, (2, 99999), True, 9, 'CA', 'CA', > (0, 1.0)), > > to > > for (dmodel, maxdis, rsrrng, rsrsgn, rsrgrp, typ2, typ3, stdev) > in \ > ((5, self.max_ca_ca_distance, (1, 99999), True, 9, 'CA', 'CA', > (0, 1.0)), > > the program still makes the restraints only for i, i+2 CA atoms. Is > there an additional over-ride for this setting in a different script?
Yes, like I said, these i,i+1 restraints are excluded because they would conflict with the stereochemical backbone restraints. See the description of 'DISTANCE' restraints at http://salilab.org/modeller/manual/node160.html and follow the link to energy_data.excl_local, which excludes restraints which conflict with bonded atoms.
In order to generate these restraints, you would have to set the third element (dihedrals) of energy_data.excl_local to False (by default all four elements are True). But I wouldn't recommend that, as unless you are very careful this will also change the behavior of nonbond interactions.
Ben Webb, Modeller Caretaker
-
Mensur Dlakic -
Modeller Caretaker