Hi
I used SHAKE for the minimization step (restrained protein + ligand, free water and ions).
&cntrl
imin=1,
maxcyc=1000,
ntmin=3,
ntb=1,
ntr=1,
ntc=2,
ntf=2,
cut=10,
restraint_wt = 10.0,
restraintmask=':1-246.CA,C,O,N | .3834-3843',
ntpr=1, ntwx=1, ntwr=100
The job stopped at step 119 and i got the following stout:
ERROR in load_lbfgs(): YS=0.
What does it mean? Is it related to the use of SHAKE during the min. stage?
By the way, I ran a minimization and a MD for the ligand itself and the protein itself.
For the ligand i ran a min. of 500 steps in vacuum (XMIN min, no SHAKE) and the following MD:
&cntrl
imin=0,
irest=0,
ntx=1,
ig=-1,
ntb=0,
cut=999.0,
ntc=2,
ntf=2,
tempi=300.0,
temp0=300.0,
ntt=3,
gamma_ln=5.0,
nstlim=1000, dt=0.001,
ntpr=1, ntwx=1, ntwr=1000
The difference between the energy's components of the last min step and the first MD step is this:
min.
NSTEP ENERGY RMS GMAX NAME NUMBER
500 -1.9339E+02 6.1958E-03 1.5149E-02 C4 10
BOND = 1.9919 ANGLE = 10.0103 DIHED = 0.0000
VDWAALS = -0.8645 EEL = 49.4645 HBOND = 0.0000
1-4 VDW = 6.3659 1-4 EEL = -260.3614 RESTRAINT = 0.0000
MD
NSTEP = 0 TIME(PS) = 0.000 TEMP(K) = 260.00 PRESS = 0.0
Etot = -183.1726 EKtot = 10.3332 EPtot = -193.5059
BOND = 1.8793 ANGLE = 10.0103 DIHED = 0.0000
1-4 NB = 6.3659 1-4 EEL = -260.3614 VDWAALS = -0.8645
EELEC = 49.4645 EHBOND = 0.0000 RESTRAINT = 0.0000
In that case the BOND component does not change too much but the MD stopped at the step 2 with the following message:
vlimit exceeded for step 2; vmax = 129.5132
Coordinate resetting (SHAKE) cannot be accomplished,
deviation is too large
NITER, NIT, LL, I and J are : 0 1 4 6 14
Note: This is usually a symptom of some deeper
problem with the energetics of the system.
For the protein i added water and ions and then I ran a min of 2500(XMIN, no SHAKE) restraining the protein and letting free solvent molecules. The minimazed structure was then used for a MD to relax solvent molecules:
&cntrl
imin=0,
irest=0,
ntx=1,
ig=-1,
ntb=1,
ntr=1,
cut=10,
ntc=2,
ntf=2,
tempi=300.0,
temp0=300.0,
ntt=3,
gamma_ln=5.0,
nstlim=15000, dt=0.001,
ntpr=1, ntwx=1, ntwr=1000
restraint_wt = 10.0,
restraintmask=':1-246.CA,C,O,N'
/
In that case the MD finished without any error.
The difference between the components of energy of the last min step and the first MD step is this:
min
NSTEP ENERGY RMS GMAX NAME NUMBER
2500 -1.1289E+05 6.4623E-02 4.1885E+00 H2 24989
BOND = 8945.1185 ANGLE = 529.6783 DIHED = 2324.8005
VDWAALS = 23331.2721 EEL = -158202.8624 HBOND = 0.0000
1-4 VDW = 829.6991 1-4 EEL = 9210.5878 RESTRAINT = 143.7273
EAMBER = -113031.7061
MD
NSTEP = 0 TIME(PS) = 0.000 TEMP(K) = 435.14 PRESS = 0.0
Etot = -96461.5836 EKtot = 25401.8527 EPtot = -121863.4363
BOND = 113.3882 ANGLE = 529.6783 DIHED = 2324.8005
1-4 NB = 829.6991 1-4 EEL = 9210.5878 VDWAALS = 23331.2720
EELEC = -158202.8623 EHBOND = 0.0000 RESTRAINT = 0.0000
Ewald error estimate: 0.1479E-03
In that case (like in my previous job with both protein and ligand) there is a big difference for the BOND component between the last min. step and the first MD step. Nevertheless the MD finished without errors.
I am not sure that the problem is related to the SHAKE.
What do you think about?
Vale
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Valentina Romano | PhD Student | Biozentrum, University of Basel & SIB Swiss Institute of Bioinformatics
Klingelbergstrasse 61 | CH-4056 Basel |
Phone: +41 61 267 15 80
________________________________________
From: Jason Swails [jason.swails.gmail.com]
Sent: Thursday, May 15, 2014 5:46 PM
To: amber.ambermd.org
Subject: Re: [AMBER] Analysis of minimization stage
On Thu, 2014-05-15 at 09:35 -0600, Daniel Roe wrote:
> On Thu, May 15, 2014 at 7:07 AM, Valentina Romano
> <valentina.romano.unibas.ch> wrote:
> > BOND = 7906.5852 ANGLE = 529.6831 DIHED = 2310.5740
> > MD
> >
> > BOND = 120.7537 ANGLE = 529.6831 DIHED = 2310.5740
> >
>
> This appears to be your problem: your bond energy jumps from 7906.5852
> to 120.7537. All of your other energy components are the same. Are you
> sure you are using the same topology file?
This looks to me like a SHAKE problem. You are _not_ using SHAKE for
minimization (ntc=1, ntf=1 by default), but then you set ntf=2 and
ntc=2, which disables the energy calculation of the SHAKEn bonds.
The very high bond energy from the minimization suggests (to me!) that
some of the bonds that are constrained in the MD adopt very strained
configurations during the minimization (for whatever reason). This
would explain why you get SHAKE failures when you start dynamics.
You can try setting ntc=2 and ntf=2 for the minimization as well and see
if that helps fix things. Note that SHAKE is not strictly correct for
minimization, but if you are just looking to relax your structure to
start dynamics it is typically OK.
Good luck,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri May 16 2014 - 05:00:02 PDT
