AMBER Archive (2005)

Subject: Re: AMBER: Antechamber/Leap, problem with Improper angles

From: David A. Case (case_at_scripps.edu)
Date: Tue Nov 15 2005 - 15:03:11 CST


On Mon, Nov 14, 2005, Christophe Guilbert wrote:
>
> I have a problem using antechamber/Leap on some ligands, especially when
> it comes to assign improper angles.

There are some subtle (and non-obvious) things going on here. I'll try to
be as clear as possible.

>
> - When I run antechamber (distribute with amber 8):
> "antechamber -nc 0 -rn lig -i tobeconvert.mol2 -fi mol2 -o lig_bcc.prepi
> -fo prepi -c bcc -j 4 -at gaff -pf y"
> "lig_bcc.prepi" indicated 15 improper angles.

The "non-obvious thing" is that the impropers specified in the prepin file
are ignored later on -- they are printed out by LEaP but never used.

Here is how LEaP determines what impropers to use: it looks for all
three-coordinate atoms. If there is an improper in the force field for that
atom and its three neighbors, it adds the improper. If there is no force
field term that matches, it skips it. It will print a warning (but still do
nothing) if an sp2 atom is missing a force field term.

> the one I particularly miss is "C2 C4 C3 O1" which should be planar.
>
> ** Warning: No sp2 improper torsion term for cd-c3-cc-os
> atoms are: C4 C2 C3 O1

Here is the warning LEaP is giving you that it thinks there should be an
improper torsion here, but that it could not find any matching parameter.

But beyond this, I don't have all the answers. Maybe Junmei can weigh in
here:

(1) Why didn't parmchk find the missing parameters?

(2) What is the following line in gaff.dat really doing:

X -X -X -X 1.1 180. 2. dac, 10/94

(This is at the end of the improper torsion list. In spite of having my
initials, I don't think this line is really mine....)

(3) Also, the following line in gaff.dat looks funny:

X -X -ca-hc 1.1 180. 2. bsd.on C6H6 nmodes

(Shouldn't "ha" always be bonded to "ca"?)

As a work-around, you can manually add the desired impropers to your frcmod
file. I'm cc-ing this to Junmei to get his comments -- it is very possible
that I am misinterpreting something here....

...thanks for the report....dac

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