[gmx-users] Fwd: sds topology // prodrg vs. am1-bcc // charges

[Mark Abraham]

Louic Vermeer wrote:
> Hi gromacs users,
> 
> I want to use a topology file for SDS and obviously I am not the first person 
> doing simulations in SDS. Previous comments on this list suggested using the 
> prodrg server, but I also read that the partial charges calculated by prodrg 
> are not always reliable.
> 
> What would be the most intelligent thing to do?
> 1. use the prodrg topology file
> 2. use the prodrg file, but change the partial charges, using for example the 
> am1-bcc method. This gives charge-groups that do not sum to a charge of zero 
> (one group has +0.2 and another one -0.2, which seems quite a lot to me). On 
> the other hand, the charges in the conjugated SO4 system are more distributed 
> according to bcc, which seems better to me than the original prodrg topology.
> 3. hope for someone using this list to send me a "correct" topology file for 
> sds

http://wiki.gromacs.org/index.php/Parameterization

> I would also like to know whether charge-groups are used at all when using 
> PME. Are they only for cut-off? I did not get that from the manual.

See section 3.4.2 - they are used for neighbour searching, which is 
still necessary for L-J interactions. My guess is that they are also 
used keeping the direct-space Ewald sum over an integral charge, but I'd 
have to read the code to be sure.

Mark