Re: [Btk-developers] generic molecule self-energy

[<EJ...@lb...>]

----- Original Message -----
From: Tim Robertson <timr@u.washington.edu>
Date: Wednesday, July 24, 2002 2:01 pm
Subject: Re: [Btk-developers] generic molecule self-energy

> This might be tricky to implement.  Unless, by "undefined" you mean 
> "energy function will spit out a value, but value will be 
> meaningless." 
>  Then it's not so hard.

I think it's actually doable given a topology file, although either the Molecule 
itself would have to carry extra topo info, or the e functions would need a topo 
table passed in.  The latter case is nice because it leaves things up to the user - 
for example I might want to scale down interactions between backbone atoms of 
adjacent residues which should work without any info about topology.

> Regarding topology files -- the plan is not to include anything 
> charmm 
> or amber specific in the molecule class itself, right?   I'm 
> thinking 
> that you need functions like Molecule::define_connectivity() (or 
> some 
> such thing), that an outside method would call to set up the tetrads.
> 
> Is this the plan (or something like it) ?

This was exactly my plan.  It would be nice to include support (outside Molecule) 
for at least one of the standard formats, and if a user wanted to add support for 
OPLS or GROMOS or Gromacs, etc. later then they could easily do so.

So we'll need some functions like Molecule::define_connectivity() that would 
probably interface with Tetrads directly.  If we're talking about saving info about 
atoms connected by 2 or 3 bonds then we would obviously need to add some 
new data structure to Molecule.  The latter case is something that I hadn't really 
thought about before Chris mentioned it.

-E