[gmx-users] Distance for window spacing in US?

[Erik Marklund]

On Feb 22, 2013, at 1:13 AM, Justin Lemkul wrote:

>
> On 2/21/13 6:24 PM, Yun Shi wrote:
>> So in umbrella sampling, does it really matter if the vector
>> connecting the COMs of protein and ligand is NOT parallel with the
>> vector of the pulling force (although the pull rate is 0)?
>>
>> I guess as long as the force in each sampling window is along the  
>> same
>> direction, we can in theory get potential of mean force along any
>> direction.
>>
>
> I honestly don't know how that would affect things.  Perhaps someone  
> else who has dealt with such cases can comment, but typically it is  
> easiest to define the reaction coordinate to be coincident with the  
> direction(s) along which you wish to apply the umbrella restraint.
>
> -Justin

The important thing is that your PMF is calculated along a sensible  
reaction coordinate. In the worst case there may be free energy  
barriers perpendicular to your coordinate which the system would have  
to cross. Those will be important for the kinetics and thermodynamics  
that you aim to study, but will not appear in your PMF.

Erik