[gmx-users] constrained mdrun about linear rigid molecule ??

[Mark Abraham]

On 12/01/2012 8:40 PM, Kiwoong Kim wrote:
> Hi,
>
> I'm not good at handling Gromacs yet.
>
> My question is about virtual interaction.
>
> I'm doing simulation dealing with CO2 which is linear rigid molecule 
> as you already know.
> As pervasively known, I introduced virtual sites.
>
> However, I don't know how to give initial coordinate correctly of 
> dummy particles.
>
> Let us consider following system.
>
> x1-----d1-----x2
>
> There are two real atom and one dummy particle between them with 
> appropriate length ratio, a, from x1, and (1-a) from x2.
> If I give the coordinates of three particles by the rules of above 
> statement, i.e., dummy particle between them with appropriate length 
> ratio, a, from x1, and (1-a) from x2..
>
> Is it correct?

Looks fine.

> Isn't there any unstable energy?

Why should there be?

>
> Do I have to run any preceding run before equilibration run??
> Otherwise, Is there any other way giving the correct coordinate of 
> dummy particle ???

I expect you can give it any coordinates you like, and the interaction 
any value of a you like. Thereafter, it's in the hands of physics. If 
you want something reasonable, then probably choose the coordinates and 
the value of a in a consistent way.

Mark