I notice you started with a GBSA minimization. You colud try a gas phase
minimization first, to remove the worse bad contacts. If atoms clashes badly
even the calculation of the surface area may meet problems. It is usual that
in case of clashes, even in gas phase, the energy in the first step is
positive and large, but it may recover, if the situation is not too
desperate.
Best Regards
Stefano Pieraccini
----- Original Message -----
From: <wang.cbs.cnrs.fr>
To: <amber.scripps.edu>
Sent: Tuesday, September 19, 2006 4:42 PM
Subject: AMBER: resolving a bad initial model
> Hi
>
> I am using AMBER8 on a Linux system. I have been trying to resolve a poor
> initial model (with clashes) by steepest descent energy minimisation. I
> use the input below.
>
> &cntrl
> imin=1, nmropt=0,
> ntx=1, irest=0,
> ntpr=1,
> ntf=1, ntb=0,
> igb=2, gbsa=1,
> ibelly=0,
> maxcyc=1000, ntmin=2,
> ntc=1, tol=0.0005, jfastw=0,
> ivcap=0, fcap=0.5,
> cut=99.0,
> &end
>
> In the output file:
>
> NSTEP ENERGY RMS GMAX NAME NUMBER
> 1 NaN NaN 6.2047E+12 HG13 160
>
> BOND = 859.0824 ANGLE = 840.4581 DIHED =
302.9558
> VDWAALS = ************* EEL = -1870.7016 EGB
= -259.3464
> 1-4 VDW = 8778.4861 1-4 EEL = 1433.5810 RESTRAINT =
0.0000
> ESURF = NaN
>
> NSTEP ENERGY RMS GMAX NAME NUMBER
> 2 NaN NaN 0.0000E+00 N 1
>
> BOND = NaN ANGLE = 52129.9369 DIHED =
189.3636
> VDWAALS = NaN EEL = NaN EGB = NaN
> 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
0.0000
> ESURF = 0.0000
>
>
> I've tried weighting the energy terms (ALL, RSTAR, INTERN, IMPROP) using
> '&wt' to get low values for the energy but minimisation still fails. I'm
> wondering if there is anything else I should try? Or must I start with a
> good initial model?
>
> Thankyou for your help.
> Conan
>
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Sep 20 2006 - 06:07:19 PDT
