[gmx-users] vdW parameters and atoms moving too fast

[Nicolas SAPAY]

<quote who="Mark Abraham">
> Nicolas SAPAY wrote:
>> good remark... here is the files! I precise that modifying the
 inter-atomic distance doesn't solve the problem and that nb parameters
 are
>> defined in the ffcharm27nb.itp file as follow :
>>
>> [ atomtypes ]
>> OT        15.9994   0.000      A  2.489540e-03  2.434767e-06
>> HT         1.0080   0.000      A  3.153970e-09  1.292128e-17
>>
>> [ nonbond_params ]
>> OT     HT      1     4.382447e-05  1.371950e-09
>
> Well the good news is that my CHARMM27 implementation agrees with these
 C6 and C12 values... for reference, I use these with
>
> [ defaults ]
> ; nbfunc        comb-rule       gen-pairs       fudgeLJ fudgeQQ
>     1            1               yes             1.0     1.0
>
> and see no problems.
>
>> [ pairtypes ]
>> OT     HT      1     4.382447e-05  1.371950e-09
>
> This shouldn't be needed, but won't be causing problems.
>
> I can't see anything wrong with your .top or .mdp files.
>
> The V and W values in the [ atomtypes ] section won't be being used,
 because of the existence of this defined interaction in [
nonbond_params
> ], so you can test that that's correct by zeroing these values and
 looking for no change in the output. They look wrong, though, and
 certainly don't combine using comb-rule 1 to give the correct values.

> Mark
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php

>
>

I'm absolutely agree to say that LJ parameters given in the [ atomtypes ]
section cannot give the correct results when using the combination rule 1.
The Gromacs manual say for comb-rule 1:
  Vii = Ci(6) = 4 * ei * si^6
  Wii = Ci(12) = 4 * ei * si^12

and then:
Cij(6) = (Ci(6) * Ci(6))^1/2
Cij(12) = (Ci(12) * Ci(12))^1/2

This assumes that sij = (si * sj)^1/2 while in CHARMM sij = si/2 + sj/2. I
have tried to used the comb-rule 2 of Gromacs to fix that. Unfortunately,
I obtain an error each time I used grompp when I define a [ nonbond_params
]("Trying to add Buck.ham(SR) while the default nonbond type is LJ (SR)").
Additionally, I obtain wrong energy values. So, I have switched back to
comb-rule 1 and [ nonbond_params ].

I have also tried to nullify OT and HT parameters in [ atomtypes ] (with
comb-rule 1 and defined [ nonbond_params ]). No modification happened, as
expected. However, I frequently obtain this kind of message when working
on simple systems made of 2 atoms:

"Can not invert matrix, determinant =   "

Together with the fact that I obtain a kinetic energy of 0.00 make me
think that a variable is not initialized or something like that... I'm
still puzzled.

Nicolas

-- 
[ Nicolas Sapay Ph.D. ]
University of Calgary, Dept. of Biological Sciences
2500 University Dr. NW, Calgary AB, T2N 1N4, Canada
Tel: (403) 220-6869
Fax: (403) 289-9311