> I have a few other questions. I understand that the PK values correspond to
> the coefficients that were obtained during the fitting of the Fourier
> series. The coefficients are divided by 2 and the resulting value is the PK
> value. The form of the series that I used did not contain a PHASE parameter
> (it was set to 0).I am assuming that this infers that the PHASE value must
> be set to 0 for the torsion parameter but wanted to clarify the matter.
>
> Form used : *PK**1/2 (1+COS(*PN**phi - 0))
I would suggest you to use a zero PHASE. At the end, if your PK is
negative, you can make the PHASE 180 and PK positive. Even if it stays in
its original negative value, it will be ok. But be careful on the PK
value! The equation that you wrote is not the right equation. You can look
at the 'Parameter Development' section in AMBER8 manual to see how the
torsional parameters are developed. The equation that is used is
Etors = ( PK / IDIVF ) * ( 1 + cos( PN * phi - PHASE) )
[Copy/pasted from the manual]
In a specific case representation, IDIVF=1. And what you fit will give
you PK/1 = PK. After the parameterization, do an MM PES calculation to see
if the torsional profile is reasonable. Do not confuse yourself with the
barrier hight, which is 2*PK! In a complex torsional energy surface, I do
not think that the barrier can be measured directly, because it will all
depend on how many cosine terms you will use. Fit to the above equation
and calculate PK with IDIVF=1. That value is what needs to be put to the
frcmod file as new torsional parameter. Also, set PHASE=0, which will make
things easier. I am not sure which fitting protocol you are planning on
using; I would suggest you to use the linear-least-square-fitting
protocol.
Good luck,
>
> Stephane Acoca
> McGill University
>
> On Mon, Feb 15, 2010 at 4:49 PM, stephane acoca <stephane.acoca.gmail.com>
> wrote:
>> Thank you both for the clarification. Especially with regards to the
>> negative terms implying a further expansion of the fourier series.
>>
>> On Mon, Feb 15, 2010 at 12:42 PM, Jason Swails <jason.swails.gmail.com>
> wrote:
>>> This is true.
>>>
>>> On Mon, Feb 15, 2010 at 11:07 AM, Ilyas Yildirim
>>> <yildirim.pas.rochester.edu> wrote:
>>>> What exactly do you mean with 'the function is only fit to a single
> fourier
>>>> term' ? The E_tors can have as many cosine terms in it as possible. For
>>>> instance, in the parm99.dat file, you have the following torsion
> defined:
>>>>
>>>> CT-CT-CT-CT 1 0.18 0.0 -3.
>>>
>>> The negative in the last column means include another term. The
>>> positive value is actually used for the parameter. Once a
>>> non-negative value is set for the last column, no further fourier
>>> terms are used for that torsion (and I'm guessing that if the values
>>> were positive, each successive term would overwrite the last if it
>>> didn't throw an error message).
>>>
>>>> CT-CT-CT-CT 1 0.25 180.0 -2.
>>>> CT-CT-CT-CT 1 0.20 180.0 1.
>>>>
>>>> This parameter set is to define the torsion energy of CT-CT-CT-CT with
>>>> different PHASE, PK, and PN values (IDIVF is equal to 1) which will have
> 3
>>>> cosine term to define the CT-CT-CT-CT torsional energy. You can increase
> the
>>>> level of accuracy by increasing the number of cosine terms used to
> define a
>>>> particular torsion.
>>>>
>>>> Ilyas Yildirim, Ph.D.
>>>> ---------------------------------------------------------------
>>>> = Hutchison Hall B#10 - Department of Chemistry =
>>>> = - University of Rochester =
>>>> = 585-275-6766 (office) - =
>>>> = http://www.pas.rochester.edu/~yildirim/ =
>>>> ---------------------------------------------------------------
>>>>
>>>> On Mon, 15 Feb 2010, stephane acoca wrote:
>>>>
>>>>> Hi, I would like to clarify the protocol for adjusting torsional
>>>>> parameters. The problem I am having is that after fitting to a fourier
>>>>> series the energy difference between the QM and MM energies (torsion
>>>>> not included in MM Energy), the frcmod specifications I find point to
>>>>> :
>>>>>
>>>>> IDIVF
>>>>> PK
>>>>> PHASE
>>>>> PN
>>>>>
>>>>> Meaning that the function is only fit to a single fourier term of the
>>>>> format :
>>>>>
>>>>> Etors = (PK/IDIVF)*(1+ COS(PN*PHI - PHASE)
>>>>>
>>>>> Aren't higher level fourier terms available to be specified torsion
>>>>> parameters. I couldn't find the information in either Amber manual
>>>>> 8,9, or 10. Any guidance would be appreciated.
>>>>>
>>>>> Stephane Acoca
>>>>> McGill University
>>>
>>>
>>> Good luck!
>>> Jason
>>>
>>> --
>>> ---------------------------------------
>>> Jason M. Swails
>>> Quantum Theory Project,
>>> University of Florida
>>> Ph.D. Graduate Student
>>> 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
>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Feb 17 2010 - 18:00:02 PST
