AMBER Archive (2005)

Subject: Re: AMBER: reading frcmod torsional parameters

From: David A. Case (case_at_scripps.edu)
Date: Mon Dec 12 2005 - 10:56:02 CST


On Mon, Dec 12, 2005, ivan_at_mmb.pcb.ub.es wrote:
>
> When I create a frcmod file that is going to have a new general torsion
> parameters and a specific torsion parameter that is included in the
> general (let's say X-CT-CI-X and CT-CT-CI-HC)
>
> two things:
>
> - Do they have to be written in the frcmod file in a certain order, I mean
> firstly the general and then the specific ones ?

Basically, you should be very careful in using wildcards for torsions,
especially in frcmod files. If you can, avoid them. If this is not possible,
please check (using rdparm, for example) that you are getting exactly what you
want in the output prmtop file.

>
> - If idivf = 9 for the general torsion, do I have to change this to 8 if I
> am explicitly including one specific torsion or this is already taken
> into acount by the program itself?
>

The torsion force constant as listed is divided by idivf. So if you have nine
terms, you need to check that the overall barrier to rotation is what you want
it to be. So, if you originally had nine equivalent terms, but now one is
different, if you want all the remaining ones not to change, you would leave
idivf alone. Alternatively, you could logically think that all of the idivf
values are 1, and just treat the force constants as the adjustable parameters,
often chosen to get an rotational barrier to match some taget value.

The idivf value were originally there to make it easier for humans to read the
parm.dat files. There is no logic in the programs that checks this value
against a list of torsions, or modifies it based on the actual molecular
topology.

...hope this helps....dac

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu