Hello!
Is it possible in MODELLER to change the weight of certain parameters in the optimization process? I am doing homology modelling with incorporation of ligand information as special restraints. The restraints are expressed in terms of knowledge based distance dependent pair potentials between protein and lingand atoms. These pair potentials are expressed in terms of cubic splines.
I would like to try if I can improve my results by increasing the scaling weight of my special restraints compared to the predefined restraints which Modeller uses in the model building process.
my top file basically looks like this: -------------------------------- INCLUDE SET ALNFILE = '../alis/1dlw_2d.pir' SET KNOWNS = '1dlw' SET SEQUENCE = 'fake' SET ATOM_FILES_DIRECTORY = '../pdbs' SET HETATM_IO = on SET STARTING_MODEL = 1 SET ENDING_MODEL = 1 SET FINAL_MALIGN3D = 1 CALL ROUTINE = 'model' STOP
SUBROUTINE ROUTINE = 'special_restraints' SET DYNAMIC_MODELLER = on SET CONTACT_SHELL = 6.00 READ_ATOM_CLASSES ATOM_CLASSES_FILE = 'atmcls-special.lib' READ_PARAMETERS FILE = 'special-pairpot.lib', ADD_PARAMETERS = on END_SUBROUTINE ---------------------------------
The file atmcls-special.lib contains the definitions of the atom groups I use, the file special-pairpot.lib contains the restraints for each protein - ligand atom pair, e.g.:
--------------------------------- MODELLER12 VERSION: MODELLER FORMAT R 10 60 1 31 2 66 0 L001 P001 1.00 0.00 5.90 0.10 -0.13 -0.05 10.00 9.99 9.95 9.89 9.80 9.68 9.55 9.39 9.21 9.02 8.80 8.57 8.33 8.07 7.80 7.53 7.25 6.96 6.67 6.38 6.09 5.80 5.51 5.23 4.96 4.69 4.44 4.06 3.95 3.59 3.07 2.62 2.26 1.86 1.40 0.95 0.61 0.38 0.17 0.01 -0.10 -0.18 -0.27 -0.36 -0.44 -0.49 -0.50 -0.51 -0.50 -0.50 -0.55 -0.62 -0.70 -0.75 -0.74 -0.73 -0.70 -0.66 -0.64 -0.64 . . . ------------------------------ ... defines the distance distribution between protein and ligand C.3 atoms.
Thank you in advance for your help.
Andreas
###################################################################### # Andreas Evers # # # # Research group for Drug Design & X-ray Crystallography # # Institute of Pharmaceutical Chemistry # # Philipps-University of Marburg # # Marbacher Weg 6 phone2 +49-6421-28-25072 # # D-35032 Marburg FAX: +49-6421-28-28994 # # email: Andreas.Evers@mailer.uni-marburg.de # # AG-Klebe-Home: http://www.agklebe.de # ######################################################################
Hi Andreas,
yes, you could play with the MAKE_SCHEDULE command, to set up a new shedule: to scale up or down certain type of restraints during optimization. See the explanation in the manual. Also, for the existing schedules, see the $MODINSTALL6a/modlib/sched.lib file
You can define your own, ligand related restraint in one of the 32 type of restraint groups (preferably choose such a group which is not used, so your restraints will be the only one in that group) so you can vary them independently. For these groups see manual pp.75, table 2.4 . Do not worry about the names, most groups are not used in the process (check your actual log file for a report about them) and the names and groupings are somewhat arbitrary, but you can go for groups 27-31, that is about arbitrary distance or NMR distance restraints and usually empty.
Andras
Evers Andreas wrote: > > Hello! > > Is it possible in MODELLER to change the weight of certain > parameters in the optimization process? I am doing homology modelling with > incorporation of ligand information as special restraints. The restraints > are expressed in terms of knowledge based distance dependent pair > potentials between protein and lingand atoms. These pair potentials are > expressed in terms of cubic splines. > > I would like to try if I can improve my results by increasing the scaling > weight of my special restraints compared to the predefined restraints > which Modeller uses in the model building process. > > my top file basically looks like this: > -------------------------------- > INCLUDE SET ALNFILE = '../alis/1dlw_2d.pir' > SET KNOWNS = '1dlw' > SET SEQUENCE = 'fake' > SET ATOM_FILES_DIRECTORY = '../pdbs' > SET HETATM_IO = on > SET STARTING_MODEL = 1 > SET ENDING_MODEL = 1 > SET FINAL_MALIGN3D = 1 > CALL ROUTINE = 'model' > STOP > > SUBROUTINE ROUTINE = 'special_restraints' > SET DYNAMIC_MODELLER = on > SET CONTACT_SHELL = 6.00 > READ_ATOM_CLASSES ATOM_CLASSES_FILE = 'atmcls-special.lib' > READ_PARAMETERS FILE = 'special-pairpot.lib', ADD_PARAMETERS = on > END_SUBROUTINE > --------------------------------- > > The file atmcls-special.lib contains the definitions of the atom groups I > use, the > file special-pairpot.lib contains the restraints for each protein - ligand > atom pair, e.g.: > > --------------------------------- > MODELLER12 VERSION: MODELLER FORMAT > R 10 60 1 31 2 66 0 L001 P001 1.00 0.00 5.90 0.10 -0.13 -0.05 10.00 9.99 > 9.95 9.89 9.80 9.68 9.55 9.39 9.21 9.02 8.80 8.57 8.33 8.07 7.80 7.53 7.25 > 6.96 6.67 6.38 6.09 5.80 5.51 5.23 4.96 4.69 4.44 4.06 3.95 3.59 3.07 2.62 > 2.26 1.86 1.40 0.95 0.61 0.38 0.17 0.01 -0.10 -0.18 -0.27 -0.36 -0.44 > -0.49 -0.50 -0.51 -0.50 -0.50 -0.55 -0.62 -0.70 -0.75 -0.74 -0.73 -0.70 > -0.66 -0.64 -0.64 > . > . > . > ------------------------------ > ... defines the distance distribution between protein and ligand C.3 > atoms. > > Thank you in advance for your help. > > Andreas > > ###################################################################### > # Andreas Evers # > # # > # Research group for Drug Design & X-ray Crystallography # > # Institute of Pharmaceutical Chemistry # > # Philipps-University of Marburg # > # Marbacher Weg 6 phone2 +49-6421-28-25072 # > # D-35032 Marburg FAX: +49-6421-28-28994 # > # email: Andreas.Evers@mailer.uni-marburg.de # > # AG-Klebe-Home: http://www.agklebe.de # > ######################################################################
participants (2)
-
Andras Fiser -
Evers Andreas