Great for clearing that up.
Thank you Jason and others who have contributed their time and knowledge to this thread.
Best regards,
Jesper
On May 2, 2012, at 5:00 PM, Jason Swails wrote:
> On Wed, May 2, 2012 at 7:21 PM, Jesper Soerensen <jesorensen.ucsd.edu>wrote:
>
>> Hi Jason,
>>
>> I know and that does make sense, since the changes only pertain to the
>> protein. But then mbondi2 were not made for ligands either - correct?
>> So should the ligand radii not come from the gaff parameters, or am I
>> misunderstanding something and gaff does not contain/specify its own set of
>> GB radii but instead use the mbondi2 radii?
>>
>
> As far as I know, they were not developed with small ligands in mind.
> However, my understanding is that GB does reasonably well for small
> ligands, since it's not a significant perturbation from the Born solution
> to the PB equation (from which GB is based). The complexity arises when
> you try to treat very complex systems like proteins.
>
> Globular proteins are obviously significantly different than a simple
> sphere, and so the correct estimation of the effective Born radius (which
> more or less measures proximity to the bulk solvent) becomes essential to
> properly characterize the solution-phase behavior. This is what the
> intrinsic radii from %FLAG RADII are used for (and *only* this in GB
> calculations). As a result, you should consider these radii a parameter of
> the GB model you're using, not of the force field you're using. (So no,
> gaff does not contain its own set of GB radii, nor does any other Amber
> force field).
>
> A gaff carbon that is sp2-bonded carbon with a double-bond to an oxygen is
> for all intents and purposes identical to the backbone C atom of an amino
> acid with respect to relative size and is therefore treated the same with
> the LEaP-assigned radii. Moreover, I argue this is the appropriate
> behavior given the purpose these radii serve.
>
>
>> Also, is mbondi2 the standard PB radii in leap now?
>>
>
> No, 'mbondi' is (at least using ff12SB).
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
> _______________________________________________
> 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 Thu May 03 2012 - 09:00:03 PDT
