[gmx-developers] free energy with urey-bradley

[Mark Abraham]

On 29/10/2011 7:17 AM, Michel Cuendet wrote:
>
> Hi Gromacs People,
>
> I'm trying to set up a thermodynamic integration simulation with the 
> charmm force field to calculate the free energy of mutating a side 
> chain to another (similar) one. But apparently, the urey_bradley 
> parameters for topology B are not taken into account... I checked in 
> the code and we have :
>
> in ifunc.c  (l.105) :
> def_angle   ("UREY_BRADLEY","U-B",          3, 4, 0,  0,          
> urey_bradley ),
> meaning zero parameters are read for the B topology.
>
> and in bondfree.c (l.835) :
>     *dvdlambda += harmonic(kth,kth,th0,th0,theta,lambda,&va,&dVdt);  
> /*  21  */
>     *dvdlambda += harmonic(kUB,kUB,r13,r13,dr,lambda,&vbond,&fbond); 
> /*  19  */
> meaning the same parameters are used for both A and B.
>
> Is there a particular reason for which the lambda-interpolation of the 
> urey-bradley term was not implemented?
>
> Has anyone implemented it in a development version?
>
> There should at least be a warning in grompp about this !


I'd guess it is an historical accident - U-B was implemented many years 
before CHARMM forcefield was added. Table 5.5 correctly notes that U-B 
cannot be interpolated in free energy calculations. It looks fairly 
straightforward to add such interpolation.

Mark