RE: AMBER: Ewald error estimate

From: Thomas Cheatham III <tec3.utah.edu>
Date: Thu, 8 May 2008 11:21:38 -0600 (Mountain Standard Time)

> I think "VDWAALS = 28137.6817 , 1-4 VDW = 343.3582" it is high which
> represents the bad contacts in water and RNA and cannot drop when run
> minimisation stage 2. Please see the the minimisation stage 2 results:

No, this likely does not represent bad contacts. If you would like to
check this, use the "checkoverlap" command of ptraj.

ptraj prmtop << EOF
trajin restrt
checkoverlap
EOF

> NSTEP ENERGY RMS GMAX NAME NUMBER
> 2500 -1.5472E+05 5.3890E-01 3.1856E+01 H1 26328
>
> BOND = 11197.6011 ANGLE = 293.1576 DIHED = 878.3573
> VDWAALS = 28137.6817 EEL = -191368.2955 HBOND = 0.0000
> 1-4 VDW = 343.3582 1-4 EEL = -4205.1036 RESTRAINT = 0.0000
>
> Maybe my initial structure has serious problem. I found there are
> warning "---close contact---" when I check the input file in xleap.
> However, I build it as the tutorials and donot know how to deal with it.

If you are worried about the reliability of the initial model, do a
restrained minimization with implicit solvent first. For example,
restrain the phosphate atoms. However, this is likely not necessary.

As remarked in previous e-mail to you by both DAC and myself, positive VDW
is not necessarily bad. It is likely water-water overlap where minor
overlap occurs as the molecules due to the favorable electrostatics.

For reference, in one of my large RNA simulations after ~200 ns, the
energies look like:

 NSTEP = 1000000 TIME(PS) = 203698.000 TEMP(K) = 299.30 PRESS = -10.1
 Etot = -236513.7298 EKtot = 46519.6974 EPtot = -283033.4272
 BOND = 748.8510 ANGLE = 1441.4921 DIHED = 2136.2434
 1-4 NB = 781.9526 1-4 EEL = -8454.9474 VDWAALS = 34320.9522
 EELEC = -314007.9712 EHBOND = 0.0000 RESTRAINT = 0.0000
 EKCMT = 22322.4293 VIRIAL = 22492.4526 VOLUME = 776128.8168
                                                    Density = 1.0280

See large +vdw but much larger negative electrostatics...

-- tec3
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun May 11 2008 - 06:07:32 PDT
Custom Search