I understand what you mean now. It still seems strange that one dihedral
type would be changed to another without much explanation why this was
done and why the other form was chosen.
In any case thank you for your explanation.
Regards,
Alex
On 04/05/2011 03:42 PM, Thomas Cheatham III wrote:
>> I am not sure I understand your arguments for the second dihedral.
>> This is how I understood the procedure. New atom type CI was introduced
>> at the C5' atom so that alpha and gamma angles would be unique, and then
>> they were fitted. Then all other dihedrals which contain atom C5' were
>> rewritten in the frcmod.parambsc0 file with the same parameters as in
>> param99 but with CT replaced with CI wherever C5' atom was. That way
>> only alpha and gamma would be changed but all other dihedrals would be
>> the same as in param99. However, dihedrals OS-CT-CI-OH and OS-CT-CI-OS
>> (OS-CT-CT-OH and OS-CT-CT-OS in param99) were not included, so these are
>> now (unintentionally) read as the general type X -CT-CI-X.
> What you say is logical and intuitive, however this is not what I think
> was done. The key text is:
>
>>> dihedrals. This differs slightly from the original parm99 force field,
>>> where the gamma is defined by the O5'-C5'-C4'-O4' atoms using the smae
>>> set of atom types (OS-CT-CT-OS) as the sugar ring torsion
>>> O4'-C4'-C3'-O3' and all other anomeric torsions. To avoid altering
>>> other conformational profiles (such as that of sugar repuckering) a new
>>> atom type (CI) was introduced and assigned to C5'. Defining a new atom
>>> type for the C5' make intuitive sense because it is expected that the
>>> O5'-C5'-C4'-O4' anomeric torsion, adjacent to the phosphorous, should be
>>> distinct from the standard (OS-CT-CT-OS) anomeric torsion."
> The writing says that OS-CT-CI-OS does not equal OS-CT-CT-OS i.e. the
> O5'-C5'-C4'-O4' anomeric torsion is being treated differently than the
> others. OS-CT-CI-OS was defined as a 3-fold (via the generic torsion
> match) and then alpha/gamma were fit. The 3-fold is how it is different.
>
> I agree it seems odd; if I were to reparameterize gamma/alpha today I
> would start by assuming the O5'-C5'-C4'-O4' had the gauche anomeric
> character (i.e. the additional 2-fold term) but then allow its parameters
> to vary during the fit / optimization too... This way we could more
> intuitively include the anomeric effect other than indirectly including
> it via the coupled gamma torsion.
>
> To date with the AMBER ff nucleic acid fixes, researchers have tended to
> only optimize one (like chi) or two angles (alpha/gamma or epsilon/zeta)
> at a time. In the future it is very likely you will see simultaneous
> optimization of more than one or two dihedrals at a time since this is
> probably necessary to better represent the energetic space of nucleic
> acids...
>
> In short, my understanding is that since the frcmod.parmbsc0 matches the
> paper, this *is* parmbsc0. Of course, one is free to argue whether this
> is correct or not for DNA or RNA and/or argue if the data in the paper is
> correct or not and/or propose new force fields that hopefully are better!
>
> --tec3
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 05 2011 - 14:00:03 PDT
