[gmx-users] error in hydrocarbon dihedrals of opls-aa?

[Markus Tusch]

Am 07.05.2010 16:03, schrieb David van der Spoel:
> On 5/7/10 8:45 AM, Markus Tusch wrote:
>> Hi everybody,
>>
>> I've just compared the following opls-aa torsional parameters for
>> hydrocarbons given in J. Comp. Chem., Vol. 18, No. 16, 1955 - 1970 
>> (1997)
>> as well as in J. Comp. Chem., Vol. 23, No. 15, 1416 - 1429 (2002)
>> with those in ffoplsaabon.itp (Gromacs 4.0.7):
>>
>> ##literature (these are also the original values for alkanes published
>> by Jorgensen in 1996 I believe):
>> V1 V2 V3
>> H-C-C-H 0.0 0.0 0.318
>> H-C-C-C 0.0 0.0 0.366
>> C-C-C-C 1.74 -0.157 0.279
>>
>> ##ffoplsaabon.itp, keyword "hydrocarbon", (translated according to eq.
>> 4.64 of the manual):
>>
>> V1 V2 V3
>> H-C-C-H 0.0 0.0 0.300
>> H-C-C-C 0.0 0.0 0.300
>> C-C-C-C 1.299 -0.05 0.200
>>
>> There's obviously a difference.
>> Is there another source for the latter parameters which I just wasn't
>> able to find? I know that the numbers in Gromacs were updated in 1999.
>>
>> Cheers,
>> Markus
>>
> Hi Markus,
>
> Erik Lindahl made those conversions based on a file provided by 
> Jorgensen. This did not make it into gromacs before 3.1 which was well 
> after 1999, but there was a downloadable force field earlier.
>
> Now the question is whether
> - the OPLS force field changed after 1996 when this was published
> - Erik made an error when converting to Ryckaert Bellemans
> - You made an error when converting back :).
Hi David,
thanks for your reply.

Well, for all the other 10 dihedrals (alcohols, ethers, acetals, 
carbohydrates) in the reference from 1997 (see above) my results match 
:). It's just the three hydrocarbon related ones that were significantly 
off.
So I guess it's likely that exactly these parameters were changed at 
some point in between, although van Gunsteren still used them in 2002 
(see above).

>
> Looking at the ffoplsaabon.itp file I can not but notice that there 
> are more than 1 H C C C dihedrals which are all different:
> C   CT CT   HC 3 -0.20934 -0.62802 0.0 0.83736 0.0 0.0 ;
> C   CT CT_2 HC 3 -0.15910 -0.47730 0.0 0.63639 0.0 0.0
> C_3 CT CT_2 HC 3 -0.15910 -0.47730 0.0 0.63639 0.0 0.0
>
> Neither of which seems to be consistent with the Eqn 4.64
> Based on H-C-C-C 0.0 0.0 0.300 I would expect
> H C C C 3 0.6276 1.8828 0 -1.2552 0

The one I am talking about is
CT     CT     CT     HC      3      0.62760   1.88280   0.00000  
-2.51040   0.00000   0.00000 ; hydrocarbon all-atom
which was also used for
CO     CT     CT     HC      3      0.62760   1.88280   0.00000  
-2.51040   0.00000   0.00000 ; acetal
and
CT     CT     CO     HC      3      0.62760   1.88280   0.00000  
-2.51040   0.00000   0.00000 ; acetal
amongst others ...

Except for the factor 2 in C3 these are very much consistent with your 
result, so maybe you just missed the 2 in C3 = -2 F3 in Eqn 4.64?
>
> So I am lost here...
>


-- 
Markus Tusch M.Sc.
University of Paderborn
Faculty of Science
Department of Chemistry
Warburger Str. 100
D-33098 Paderborn
Germany

Tel. +49-5251-60-5754
e-mail mtusch at mail.upb.de