[gmx-users] Larger number of decimal places for coordinates with velocities

[Mark Abraham]

On Fri, Jul 5, 2013 at 5:16 PM, C.M.Sampson <cms1g11 at soton.ac.uk> wrote:
> Dear all,
>
> I use Gromacs 4.5.5 with the AMBER ff99SB force field.
>
> I'm working on a method that requires me to run short NVE simulations
> one after the other, but randomly generating velocities at the start of
> each simulation. i.e. I would run 10 ps, use the final structure with
> new random velocities and run another 10 ps.
>
> I had issues with the final potential energy of the previous simulation
> not matching the first potential energy of the current simulation, but
> managed to fix that by converting the coordinates from the .xtc file to
> a .gro file.

Seems like you're not using the unconstrained-start .mdp option correctly.

> My problem is that when I then put the velocities into my new .gro file
> the simulation breaks after the first step and the kinetic energy is way
> too high:
>
> "
>            Step           Time         Lambda
>               0        0.00000        0.00000
>
>    Energies (kJ/mol)
>      Bond          Angle Ryckaert-Bell.          LJ-14     Coulomb-14
> 5.92468e+03    1.01216e+03    7.68471e-01    6.38163e-01    3.95610e+00
>      LJ (SR)   Coulomb (SR)   Coul. recip.      Potential    Kinetic En.
>  8.66770e+03   -2.27543e+03    5.12418e+04    6.45763e+04    2.11278e+07
>    Total Energy    Temperature Pressure (bar)
>     2.11924e+07    7.81749e+05    1.08075e+07
> "
>
> The temperature should be 300K and if I use a .gro file with less
> decimal places these same velocities work.

The .gro format is fixed-width, and that's life. The above results are
an immediate consequence of breaking the rules.

> I thought it had to be the way I had formatted my .gro file, but found
> the following:
>
> "
> This format is fixed, ie. all columns are in a fixed position.
> Optionally (for now only yet with trjconv) you can write gro files with
> any number of decimal places, the format will then be n+5 positions with
> n decimal places (n+1 for velocities) in stead of 8 with 3 (with 4 for
> velocities). Upon reading, the precision will be inferred from the
> distance between the decimal points (which will be n+5). Columns contain
> the following information (from left to right):
>       * residue number (5 positions, integer)
>       * residue name (5 characters)
>       * atom name (5 characters)
>       * atom number (5 positions, integer)
>       * position (in nm, x y z in 3 columns, each 8 positions with 3
>         decimal places)
>       * velocity (in nm/ps (or km/s), x y z in 3 columns, each 8
>         positions with 4 decimal places)
> "
>
> within http://manual.gromacs.org/online/gro.html .
>
> An extract of my .gro file can be seen below:
>
> "
> Generated by trjconv : 2168 system t=  15.00000
>  2168
>     1ETH     C1    1   2.735383   2.672010   1.450194  0.2345 -0.1622
> 0.2097
>     1ETH    H11    2   0.015804   2.716597   1.460588  0.8528 -0.7984
> 0.6605
>     1ETH    H12    3   2.744822   2.565544   1.409227 -2.3812  2.8618
> 1.8101
> "
>
> I was wondering if there's a way to use a larger number of decimal
> places for the coordinates with velocities?

The .g96 file format has many digits and text representation. Use
mdrun -c file.g96 with otherwise normal inputs to get an example to
reproduce.

Mark

> Best Wishes
>
> Chris Sampson
>
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> * 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/Support/Mailing_Lists