[gmx-users] GAFF and alkynes (Justin Lemkul)

[Justin Lemkul]

On 2/13/15 2:52 PM, Rebeca García Fandiño wrote:
> Hi,
> sorry for insisting about this issue...but it is not clear for me yet. I have increased (in 10 times) the force constant of the two angles involved in my molecule (those involving the alkyne) and I obtain a correct geometry without no errors during the simulation, and perfectly linear.
>
> Is there anything dangerous using this approximation? I don´t understand very well Justin´s answer to my previous suggestion:
>

I would suggest some proper parametrization rather than an ad hoc change to 
force a certain geometry.

>> A restraint is just a biasing potential towards some desired outcome.  It's not
>> necessarily even different in form from a normal harmonic interaction, but I'd
>> be hesitant to try to use some angle definition and a restraint on top of that.
>>    The GROMACS code indicates there is a linear angle potential, but it doesn't
>> seem to be documented.  This was likely added to circumvent the previous
>> problems with these linear species.
>
> You suggest that maybe it is not a good idea but I cannot understand very well the reasons. Do you see anything very 'weird'  increasing the force constant of an angle during a simulation to obtain a linear functionality?
>

Same reason as above.  If the bonded parameters are misbehaving, layering more 
approximations on top is generally just covering up problems in the parametrization.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================