From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Mon Jun 03 2013 - 00:43:15 CDT
Hi Francesco,
maybe this has something to do with the hard coded water models in NAMD. Please remember that most water models doesn’t include H-O-H angle parameters, as not needed due the H-H bond and SHAKE, possibly some other parameters are missing too, or NAMD just doesn’t care about the settings from the psf/parm and just uses its hard coded water treatment as set: tip3? So NAMD, due your bonding, doesn’t recognize your bonded water molecule now(or it does but the psf hasn’t all parameters) and can’t apply all required parameters and treatments to it and so behaves weird and hangs and crashes or asks for parameters. What you are doing could be an unexpected/unhandled case in the NAMD code. Using a more detailed water model could help here, or at least defining all angles. Additionally: to prevent NAMD from recognizing the bonded water as solvent, it could be necessary to change the atom names, as I guess this is the way (+masses) the water molecules get recognized.
Also you could be interested in the “extrabonds” feature of namd. This would’t cause any changes to your residues and provides the ability to apply bonds,angels,torsions etc.
Hope that helps, good luck.
Norman Geist.
Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im Auftrag von Francesco Pietra
Gesendet: Sonntag, 2. Juni 2013 15:45
An: NAMD
Betreff: namd-l: Does MOLLY really retain bond motion for water?
Hello:
I have a problem with a water molecule bound to Zn(II) in a metal cluster of a protein, under parm7 amber ff. As I posted before (unanswered), by setting "rigidBonds water", NAMD 2.9-CUDA hangs, asking for bond parameters for the two H-atoms of the water molecule bound to Zn(II). Although the whole cluster, such water included, is called MOL.
By not freezing any bond, the simulation goes on smoothly, albeit slowly. Therefore, I have now tried to retain all bond motions by the use of MOLLY, of which I had no previous experience, with multiple timesteps. Got the same error message, asking for params as said above. In all cases, "timestep 1.00" The relevant portion of the conf file:
## colvars
colvars on # provide a colvarsConfig (and colvarsINput if restart)
colvarsConfig ./Zn-ACT_colvars.in
colvarsInput ./press-01.restart.colvars.state
## Output
restartfreq 1000
dcdfreq 1000
xstFreq 1000
outputEnergies 6000
outputPressure 1000
outputname ./press_molly-02
## SIMULATION PARAMETERS
## AMBER FF settings
amber on
# rigidBonds water
#useSettle on
# rigidTolerance 1.0e-8
molly on
mollytolerance 0.00001
mollyiterations 100
cutoff 9.0
pairlistdist 11.0
switching off
exclude scaled1-4
readexclusions yes
1-4scaling 0.83333333
scnb 2.0
zeromomentum on
ljcorrection on
watermodel tip3
## Integrator Parameters
timestep 1.00
nonbondedFreq 2 # ts between nonbonded eval
fullElectFrequency 6 # number ts between full electrostatic eval
stepspercycle 12
margin 0
Which error related to MOLLY in conf file? If none, I wonder whether MOLLY really retains bond motions for water, or the way MOLLY does that is different by allwing all bond motions by the traditional flags.
Thanks
francesco pietra
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:17 CST