[gmx-users] Re: request for dihedral PMF test system or complete alanine dipeptide topology file

[Vojtěch Spiwok]

Dear Chris,

I have done similar calculation. Look at the supporting
info for http://dx.doi.org/10.1021/jp068587c. If you find
anything useful for you feel free to contact me.

Best regards

Vojtech Spiwok

> 
> Message: 1
> Date: Wed, 25 Apr 2007 15:39:53 -0400
> From: Chris Neale <chris.neale at utoronto.ca>
> Subject: [gmx-users] request for dihedral PMF test system or complete
> 	alanine dipeptide topology file
> To: gmx-users at gromacs.org
> Message-ID: <462FAE89.90601 at utoronto.ca>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> Does anybody have a good test system for reproducing the PMF about a 
> dihedral? I believe that my procedure is correct, and I have 
> successfully reproduced a 1d dihedral PMF for a 4 atom chain system 
> simulated in the absence of nonbonded interactions.
> 
> However, my results using an oplsaa interpretation of the alanine 
> dipeptide ACE-ALA-NAC disagree with the literature both in unconstrained 
> runs and in a 2D phi-psi PMFs. For example my unconstrained runs of 25ns 
> only sample beta space and my PMF shows a 6kcal/mol barrier for 
> transition over psi from beta to alphaR. On the other hand 
> (Hu,Elstner,Hermans,Proteins 2003) show that 6ns is enough to 
> significantly sample both Beta and alphaR space and (Ponder,Case, 
> Adv.Prot.Chem 2003) indicates that the barrier for this beta-alphaR 
> transition should be between 1.5 and 2 kcal/mol.
> 
> If somebody has a good test system that would be greatly appreciated. I 
> am also including my alanine dipeptide topology file, but I am fairly 
> sure that it is correct. The only thing that I still question is the 
> difference between the C-N-CT-HC dihedral parameters (c-terminal in 
> alanine dipeptide) and C-N-CT_2-HC dihedral parameters (involved where 
> i+1 is an amino acid residue and not NAC) in ffoplsaabon.itp. However, I 
> have tested the system while modifying C-N-CT-HC to all zeroes and it 
> does not change the gross morphology of my results.
> 
> In addition I have defined the following types at the beginning of my 
> .top file to correspond to the CT=CT_2 case as I have previously posted 
> here: 
> http://www.gromacs.org/pipermail/gmx-users/2006-September/023875.html. 
> As a further test I have generated PMFs without the inclusion of these 
> additional parameters, allowing them to default to all zeroes and it 
> does not change the gross morphology of my results.
> [ dihedraltypes ]
>   CT     C      N      CT_2    3     30.28798  -4.81160 -25.47638   
> 0.00000   0.00000   0.00000 ; peptide - V1 changed to 2.3
>   CT_2   C      N      CT      3     30.28798  -4.81160 -25.47638   
> 0.00000   0.00000   0.00000 ; peptide - V1 changed to 2.3
> 
> Before moving on to my .itp file, here are a couple of other points of 
> interest for anybody embarking on a dihedral PMF determination:
>   (i) g_chi uses non-standard dihedrals, use g_rama instead.
>   (ii) the default xtc_precision=1000 may not be large enough to get the 
> most out of your data. The default value here is fine for distances, but 
> the precision in a dihedral will be less than the precision in the 
> coordinates when the dihedral is calculated from the saved coordinates.
> 
> Many thanks,
> Chris.
>