Hi Oliver,
Please keep this on the gromacs mailing list. Please feel free to send
me a separate email to notify me that you have made a post and to ask
for my reply on that list. My answers here are submitted alongside your
text below.
Oliver Mirus wrote:
Hi Chris,
I've followed your short tutorial on the Gromacs Users mailing list.
As this is my first "force field hacking" I have a question regarding
your advice in point 3 in
http://www.gromacs.org/pipermail/gmx-users/2006-May/021416.html
"3. Added [dihedraltypes] from lipid.itp to ffoplsaabon.itp.
- Prior to running ensure that the non-RB dihedral does not exist
for these groups."
I have downloaded lipid.itp, popc.itp and the corresponding coordinate
file from Tieleman's website to give it a try. However, I am having a
hard time identifying the non-RB dihedrals you were mentioning.
In lipid.itp I found these dihedraltypes:
[ dihedraltypes ]
LP2 LP2 3 9.2789 12.156 -13.120 -3.0597 26.240 -31.495
LH2 LH2 3 9.2789 12.156 -13.120 -3.0597 26.240 -31.495
If I understand you correctly, I have to delete all non-RB dihedrals
involving LP2 and/or LH2? This means I have to check in the popc.itp
file in the dihedrals section, that no non-RB dihedral exists, which
LP2 and/or LH2 are a part of? Several dihedrals are already of
RB-type, but e.g. a dihedral involving LH2 is still of type 1.
First the simple answer: Ignore that statement.
Then a bit more info: I meant to go through the *protein* ff and ensure
that there are no non-RB dihedrals. Of course anything that is in
lipid.itp is there for good reason.
Now the complete reply: I have re-read my old post and I understand your
confusion.
Let's call the forcefields i) the protein forcefield (e.g. ffoplsaa.itp)
and ii) the lipid forcefield (e.g. lipid.itp)
What I meant by the note in section3 can be reduced to this: The atom
types that you are adding from the lipid forcefield must satisfy one of
the following two conditions: a) they must have novel atomtypes from all
atoms in the protein forcefield or at least not overlap or conflict in
the parameter definitions that utilize them; or b) they must not have
any special features defined in the lipid forcefield.
An example of (a) above being true is for LH2 and LP2. There are no
protein atoms that I know of with atomtypes of these names. This is
thanks to the good design of the initial creators.
[ dihedrals ]
; ai aj ak al funct phi0 cp mult
1 4 5 6 1 0.0 3.76 3
4 5 6 7 1 0.0 5.85 3
5 6 7 8 1 0.0 3.76 3
6 7 8 11 1 0.0 1.05 3
[...]
But what makes a dihedral a dihedral of LP2 or LH2? Only when the
number in the first column (ai) maps to LH2 or LP2?
I actually couldn't remember this for sure so I looked it up. A RB
dihedral is defined by aj and ak (with ai and al being wildcards). Note
that it can also be defined by all 4 (ai aj ak al) but it is not in this
case as we just get "LP LP" or "LH LH"
The relevant code is from src/kernel/toppush.c around line 611 in
gromacs3.3.1
if(alc[2][0]=='2') {
/* improper - the two atomtypes are 1,4. Use wildcards for 2,3 */
strcpy(alc[3],alc[1]);
sprintf(alc[2],"X");
sprintf(alc[1],"X");
/* alc[0] stays put */
} else {
/* proper - the two atomtypes are 2,3. Use wildcards for 1,4 */
sprintf(alc[3],"X");
strcpy(alc[2],alc[1]);
strcpy(alc[1],alc[0]);
sprintf(alc[0],"X");
}
I am sorry for my ignorance, but I have to start somewhere. ;-)
Another good place to start is a gromacs search for: lipid.itp oplsaa
That would not have given you the specific info that you requested in
this email, but it will give you some other good info that will no doubt
be useful.
Also check out the following link:
http://www.gromacs.org/pipermail/gmx-users/2006-September/023639.html
Thank you very much for your help!
best regards
Oliver Mirus
-------------------------
Institute of Botany (LMU)
Menzinger Str. 67
80638 Munich
Germany
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://www.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to [EMAIL PROTECTED]
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
