[gmx-users] prodrg 4.5beta generated topologies and exclusions

[Mark Abraham]

Dean Cue bas wrote:
> Hello all. Just a quick clarification, please.
> 
> Reading the original GROMOS53A6 paper, it appears that 2nd neighbor 
> (1-3) interactions are always excluded, and that third neighbor (1-4) 
> non-bonding interactions are used, yet modified in some circumstances. 
> The paper also states that all (1-4) interactions should be explicitly 
> excluded for atoms either within or directly bonded to aromatic rings to 
> help keep planarity. I could confirm this in the adenine topology in the 
> ffG53a6.rtp file where there were the above described exclusions in the 
> [exclusions] section for that residue.
> 
> Now my questions.
> 
> Prodrg4.5beta produces .itp files for all my ligands where under the 
> [moleculetype] section it states that nrexcl is 3.
> 
> Doesn’t this automatically exclude all (1-4) interactions for that ligand?  

Yes, such interactions are excluded.

> If so, then doesn’t that automatically negate the need to explicitly 
> exclude the (1-4) interactions for planar aromatic systems as described 
> above in the ff paper?

The authors of that paper were probably not pre-supposing the use of any 
particular topology-generation tool.

> If so, then doesn’t excluding all these third neighbor (1-4) 
> interactions for ligand topologies produced by Prodrg ignore 
> intra-molecular interactions that are important in the simulated 
> behavior and properties of these ligands?

It would seem so. If this bothered me, I would start by reading the 
PRODRG documentation - but I would have done that before using anything 
it generated.

> Doesn’t this imply that the default for ffG53a6 intra-molecular 
> protein-atom/protein-atom non-bonded interactions is nrexcl =2 ?

That might depend on the mechanism that is being used for the different 
and/or missing 1-4 interactions. Pre-excluding and then adding might be 
easier than pre-including and then excluding.

> Thanks in advance for any clarification in this area.  I just want to be 
> sure I’m accounting for my exclusions properly.

Mark