[gmx-users] ffgmx: CH2-CH2 Dihedral

[Erik Lindahl]

Forgot to say,

I wrote a paper in 2001 (J. Chem. Phys. 115, p. 4938) where we studied 
the dynamics of dihedral transitions in the hydrocarbon tails of lipids.

This is extremely sensitive not only to the trans/gauche states but also 
to the barrier. Figure 4 in that paper contains a comparison of the 
effective
potential from either normal dihedral + 1,4 interactions, the 
Ryckaert-Bellemans potential, and the Kuwajima potential. I actually 
like Kuwajima
a lot, although OPLS will probably be as good if you want an all-atom model.

(Tried to extract the plot as PDF without luck - sorry... the paper 
should be available online, though.)

Cheers,

Erik


Erik Lindahl wrote:

> Hi Dallas,
>
> A lot of MD force fields have this problems with aliphatic dihedrals. 
> I would guess that OPLS is slightly better than Gromos96, which in 
> turn should be better than Gromacs. (The "gromacs" force field is 
> really Gromos87 with some modifications mentioned in the manual - we 
> don't try to develop our own force field, it is too
> much religion in it :-)
>
> If you want to stick to united atom I would recommend either 
> Ryckaert-Bellemans potentials or The Kuwajima alternative. The R-B 
> potential is already defined
> in Gromacs; just change the dihedral type to "3" and make sure you 
> REMOVE the 1,4 interactions since these are included in this 
> functional form.
>
> Cheers,
>
> Erik
>
>
> Dallas Warren wrote:
>
>> I was just wondering if someone here knows the reason for the value 
>> used for the CH2-CH2 dihedral in the GROMACS force field?
>>
>> I have been doing some simulations containing bile salts, which have 
>> a short hydrocarbon chain attached to essentially cholesterol.  The 
>> chain was too straight, so I started to look into why this would be 
>> the case.  A simulation of 100 decane molecules showed chains that 
>> are very straight with only one, may be two gauche confirmations per 
>> chain.  Something isn't looking right here ;-)
>>
>> So, then looked for a paper on the energies of a CH2-CH2 dihedral, 
>> pulled Allinger et al. J. Chem. Phys., 106(12) (1997) which is "The 
>> torsional conformations of butane: definitive energetics from ab 
>> initio methods".  The difference in energy between the anti and 
>> gauche conformations is 0.62 kcal/mol.  I performed an energy 
>> minimisation of butane in the anti and gauche confirmations and found 
>> that the energy difference using the GROMACS forcefield is 1.58 
>> kcal/mol!  The energy for the dihedral is over twice of what it 
>> really should be, no wonder the chains are straight!
>>
>> How was the value decided on?  It just seems to me that is a huge 
>> difference, even if the force field isn't meant for hydrocarbons, 
>> proteins/amino acids still contain chains of CH2-CH2 and those 
>> sections are going to be too rigid.
>>
>> Catch ya,
>>
>> Dr. Dallas Warren
>> Research Fellow
>> Department of Pharmaceutical Biology and Pharmacology
>> Victorian College of Pharmacy
>> Monash University
>> 381 Royal Parade
>> Parkville VIC 3010
>> dallas.warren at vcp.monash.edu.au
>> +61 3 9903 9076
>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20021114/69f5c24f/attachment.html>