[Openmd-users] OpenMD - can we do this?

[Dan Gezelter]

Brett,


  That's the form I implemented, although in looking it over I notice a sign error in the first derivative.   Hold on...  OK, just fixed it in svn.  You'll want to do an "svn update" to get the fix.

If you want to check it, Cosine bend should be in  src/types/CosineBendType.hpp, and the code that parses the force file should be in src/io/BendTypesSectionParser.cpp

There are some oddities in bends when the angles are close to zero or 180.  The forces in bends have a 1/sin(theta) term, and we handle the divergence by implementing a "floor" when the sin approaches zero: 

    if (fabs(sinTheta) < 1.0E-6) {
      sinTheta = 1.0E-6;
    }

This code is in src/primitives/Bend.cpp.

 --Dan




On Aug 5, 2010, at 10:26 AM, Brett Donovan wrote:

> Hi Dan,
> 
> Looks like it is indeed equation 4 that we would like to replicate. Is
> it definitely this that is implemented:
> 
> Vbend = Kbend * 0.5(cos(\theta) - cos(\theta_0) )^2 ?
> 
> If so this is the ticket!! I did not know this was already in the
> repertoire! So that's fantastic. I'm more used to implementing cosine
> forms that are quadratic at 180 degrees, as I said I can understand
> the need for this function at 120 degrees between, but not at 180
> degrees as the small angle expansion is not quadratic.
> 
> Thanks again...best wishes....
> 
> 
> Brett
> 
> On Thu, Aug 5, 2010 at 2:47 PM, Dan Gezelter <gezelter at nd.edu> wrote:
>> Brett,
>> 
>>  I'm looking at that paper now [J. Phys. Chem. B 111, 7812-7824 (2007)].  I don't see anything like the formula you are describing.  Equation 4 is the cosine bend that we've already implemented.  To use it you'd have a line like this in the BendTypes section of your force field file:
>> 
>> A B C  Cosine theta0  ktheta
>> 
>> where theta0 is in units of degrees, and ktheta would be in units of kcal/mol (Marrink's paper has k_angle in units of KJ mol^-1 rad^-2, but these units can't be correct if they are using a harmonic function of the cosine.)  To mimic the chain behavior in Martini, I think you'd want:
>> 
>> C C C Cosine 180.0  5.975
>> 
>> If there's another reference I should be looking at, please let me know.
>> 
>>  --Dan
>> 
>> On Aug 5, 2010, at 4:45 AM, Brett Donovan wrote:
>> 
>>> Hi Dan
>>> 
>>> The reference for the paper is:
>>> 
>>> "The MARTINI Force Field: Coarse Grained Model for Biomolecular
>>> Simulations", and it is the Vbend potential which they use for two
>>> adjacent bond vectors. I believe they use the standard convention that
>>> two adjacent bond vectors point towards/way from each other.
>>> 
>>> Its a odd function is some ways as 180 degree's the small angle is
>>> quartic-like and at 120 degrees for a unsaturated bond its
>>> quadratic-like. It would seems it become popularized due to Marrink.
>>> 
>>> Let me know if you need anything from me,
>>> 
>>> Best wishes
>>> 
>>> 
>>> Brett
>>> 
>>> 
>>> 
>>> On Wed, Aug 4, 2010 at 8:41 PM, Dan Gezelter <gezelter at nd.edu> wrote:
>>>> Brett,
>>>> 
>>>>  Actually, we already implemented basic cosine bends:
>>>> 
>>>>  V = 0.5 * kbend (cos(theta) - cos(theta0))^2
>>>> 
>>>> (which is I think the one Marrink has been using).  What form do you guys want?
>>>> 
>>>> 
>>>>  --Dan
>>>> 
>>>> ***********************************************
>>>>  J. Daniel Gezelter
>>>>  Associate Professor of Chemistry
>>>>  Director of Graduate Admissions
>>>>  Department of Chemistry and Biochemistry
>>>>  251 Nieuwland Science Hall
>>>>  University of Notre Dame
>>>>  Notre Dame, IN 46556-5670
>>>> 
>>>>  phone:  +1 (574) 631-7595
>>>>  fax:    +1 (574) 631-6652
>>>>  e-mail: gezelter at nd.edu
>>>>  web:    http://www.nd.edu/~gezelter
>>>> ************************************************
>>>> 
>>>> 
>>>> 
>> 
>> ***********************************************
>>  J. Daniel Gezelter
>>  Associate Professor of Chemistry
>>  Director of Graduate Admissions
>>  Department of Chemistry and Biochemistry
>>  251 Nieuwland Science Hall
>>  University of Notre Dame
>>  Notre Dame, IN 46556-5670
>> 
>>  phone:  +1 (574) 631-7595
>>  fax:    +1 (574) 631-6652
>>  e-mail: gezelter at nd.edu
>>  web:    http://www.nd.edu/~gezelter
>> ************************************************
>> 
>> 
>> 

***********************************************
  J. Daniel Gezelter
  Associate Professor of Chemistry
  Director of Graduate Admissions
  Department of Chemistry and Biochemistry
  251 Nieuwland Science Hall
  University of Notre Dame
  Notre Dame, IN 46556-5670

  phone:  +1 (574) 631-7595
  fax:    +1 (574) 631-6652
  e-mail: gezelter at nd.edu
  web:    http://www.nd.edu/~gezelter
************************************************