[gmx-users] Lennard Jones spheres

[Mark Abraham]

Chiara Parravicini wrote:
> Dear Mark,
> thanks for your reply!
> 
> I'll try to better focus my issue: my intent is to perform simulations 
> similar to which described in some recently published papers in which 
> cavities are expanded by means of  LJ spheres, using methods called 
> "pressure guided MD" or "active-site pressurization" (Kimura et al., 
> Proteins 2008 Jan 3; Mazanetz et al., Biochemical Society Transactions 
> 2008). I'm working with a receptor model but experimental informations 
> about the ligand binding mode are not available to date. I don't want to 
> introduce bias in my MD simulations using a structure with the ligand 
> already inserted so, inspired by that papers, I was thinking about LJ 
> spheres.

Whatever you do to achieve this biases the receptor somehow, and I'd 
guess that doing it right is an even darker art than much of MD already 
is :-) There's various strategies possible to generate your starting 
structure. If you already have some gap, then using genbox to place 
waters in the gap and then using the coordinates of the water atoms to 
place your LJ-sphere, is possible. Or you could place them with your 
favourite molecule builder. Or do it by hand in a text editor.

You don't need anything fancy in the topology - just a thorough 
knowledge of Chapter 5 of the manual, and a .itp file for your "LJ 
sphere" molecule that defines LJ and nothing else. You're on your own 
for guessing how big the parameters should be, though! You can even use 
ions.itp as a template for that .itp file. Then generate your receptor 
.top with pdb2gmx from a structure file that has only the receptor, and 
edit that .top file to #include the .itp, change the [ molecules ] 
section, and present the new .top file with your 
spheres-added-to-receptor structure file to grompp. Easy! :-P

Mark