Why is Sigma value of Hydrogen in OpenMM's force field file (.xml) equal to 1 ?

[arjunwadhawan]

Hi,

I am

I am comparing the gromacs and openMM force field value for tip3p-fb.

The sigma for Hydrogen is 0 in gromacs force field parameters (upper attached file) where as the sigma for Hydrogen is 1 in OpenMM force field parameters (lower attached file 2 ).

Am I wrong in understanding or missing something ?

[leeping]

Hi there,

Good eye - there is no issue with the differences in this parameter. There are no vdW interactions that involve this hydrogen, so changing the sigma parameter in OpenMM makes no difference. Another way to say this is, because epsilon is set to zero and we're using the Lorentz-Berthelot rules, any pairwise LJ interaction involving HW_tip3pfb is zero everywhere.

According to Peter Eastman, the OpenMM developer, "There are situations where [sigma=0] can lead to a nan. Better to set only epsilon to 0 and use some arbitrary positive value (like 1.0) for sigma." So the "1" is there only for numerical stability of OpenMM and doesn't affect any real simulation results.

Thanks,

• Lee-Ping

[arjunwadhawan]

Cool. Thanks. I was making a file for tip4p-ice.xml which does not exist in the default version of OpenMM and then I noticed this while coping from gromacs

[arjunwadhawan]

Hi , Can you help me with another small issue (kindly ignore if not of your domain):

What is the meaning and value of "wight1", "wight2", and "wight3" for Tip4Pice water model's virtual site ?? I am using the above values ( "wight2", and "wight3" taken from gromacs's file) but my simulation is not running as it should be.

[leeping]

The weights in OpenMM are numerical averages of the positions of the three atoms, so they should add up to one.

[arjunwadhawan]

Hey Lee, i have another question on similar line. Sorry if it bothers you but I have been getting different results from Gromacs vs OpenMM and have been trying to figure out the issue.

I want to ask about the "HarmonicAngleForce" 's k (spring constant ). For Tip4pEw.xml, the value of k="836.8". It has units as kJ/mol/radian^2. (first image)

For same water model, the value in gromacs is k="628.02" and their manual says the units are same as kJ/mol/radian^2. (page 71 of ftp://ftp.gromacs.org/pub/manual/manual-4.5.6.pdf).

I am a bit confused with what is right or what I am missing. Best case , can you please share tip4p_ice.xml with right parameters as that is what I am trying to construct rightly

[leeping]

The value of the angle parameter should not affect the simulation results at all, as these are rigid water models. The parameters are there as "dummy" values, but they should never be used.

I'm not the developer of the TIP4P-Ice force field and cannot provide a parameter file that I know to be correct. Because this is your research project and you're the first person to attempt running TIP4P-Ice simulations in OpenMM, you should be the one to do this. :)