[gmx-users] ED sampling and overall rotation

[Thomas Stockner]

Hi Bert,

thanks for your answer. One more thing is unclear.

The eigenvectors used for enhanced sampling define directions in space. 
During a simulation, my protein will move and tumble and the initial 
orientation will not match with the orientation at a later time point of 
the simulations.
How is this overall rotation of my molecule taken into account when 
projecting the MD step onto the subspace spanned by the eigenvectors 
that are used for enhances sampling?

ciao

Thomas
----------------
Thomas Stockner
University of Calgary
Department of Biological Sciences, Biol 416
2500 University Dr
Calgary, AB, T2N 1N4
Canada

Tel:++1-403-2202439
Email:stockner at ucalgary.ca

>
>Thomas Stockner wrote:
>  
>
>>Dear Gromacs users,
>>
>>I am trying to use ED sampling, but some details remain unclear.
>>How is the parameter "slope" (nm/step), specified with make_edi, exactly 
>>calculated (defined)? Is it an average over the last X steps?
>>    
>>
>
>it's a running  average over the last 500 steps (usually ps).
>
>  
>
>>Are bond constraints (LINCS) applied after atoms (e.g. CA) are projected 
>>along the eigenvectors in case of rejection of the MD step?
>>
>>    
>>
>
>edsam is called after lincs (or shake), as an additional coordinate constraint.
>no MD step is "rejected" btw. edsam either accepts a step as is, or it corrects
>the position along selected eigenvectors, if required.
>
>  
>