Previous Thread
Next Thread
Print Thread
Page 2 of 2 1 2
Re: Parameterizing adenosine-like drug molecule
rmv #25312 09/03/10 12:28 AM
Joined: Feb 2010
Posts: 29
Ascaris Offline OP
Forum Member
OP Offline
Forum Member
Joined: Feb 2010
Posts: 29
By hand it is, then. But I'm still not clear on where I can find out how to create an IC table by hand, even if it only lists dihedral terms... e.g.,
Quote:
How does one determine what dihedrals (proper and improper) need to be listed in the IC table of the rtf?
This is what I used the IC gene before, not the values (which indeed had to be edited as predicted, which I did). I had read what you quoted from intcor.doc and the subsequent numbered list.

Now, looking at the rtf for the adenine in AMP from the CGENFF rtf I see that improper dihedrals are defined in the IC table with central atoms C2, C4, C6, N6, C8, and N9. This is all of the impropers that I can see in this section of the molecule, except for the one with C5 as the central atom. Why is there no dihedral term with C5 as the central atom? If there were such a term listed, I feel like I'd have a pretty good sense of which ones I'd need to define for my fragment (those with central atoms C2, C4, C5, C6, N6, N7, C8, C9). Also, the lines with C6 and N6 are exactly duplicated; I assume these duplicates are extras that have not been deleted.

Then, for the (proper) dihedrals I see that one dihedral term for 6 pairs of central atoms are defined once (one is duplicated exactly). I can see not needing to define multiple terms for each pair of central atoms, but why are there no dihedrals at all for X-C6-N1-X or X-N1-C2-X or X-C2-N3-X or X-C4-C5-X or X-N7-C8-X ? It seems that understanding this is crucial for creating an IC table by hand...

(Looking back at my stream file, I see at least one typo or nonsensical dihedral created by ic gene, not sure which but I'll assume typo by me...)

Thank you for all your help.


Intestinal nematode
Bronx, NY
Re: Parameterizing adenosine-like drug molecule
Ascaris #25315 09/03/10 06:48 AM
Joined: Sep 2003
Posts: 4,797
Likes: 2
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 4,797
Likes: 2
Just to be sure, since you never actually mentioned anything about this: You do not happen to have x,y,z coordinates for your molecule? Would 3DP9 be of any use? You see, if you have x,y,z coordinates this whole frustrating IC-table business in the RTF can be skipped.


Lennart Nilsson
Karolinska Institutet
Stockholm, Sweden
Re: Parameterizing adenosine-like drug molecule
lennart #25320 09/03/10 05:24 PM
Joined: Sep 2003
Posts: 8,507
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,507
Assuming a planar molecule, consider the following--
Code:
IC  C6  N1  C2  N3   0.   0.   0.   0.   0.
IC  N1  C2  N3  C4   0.   0.   0.   0.   0.
IC  C2  N3  C4  C5   0.   0.   0.   0.   0.
IC  N3  C4  C5  N7   0.   0. 180.   0.   0.
IC  C4  C5  N7  C8   0.   0.   0.   0.   0.
IC  C5  N7  C8  C9   0.   0.   0.   0.   0.
IC  N7  C8  C9  C4   0.   0.   0.   0.   0.
IC  C8  C9  C4  C5   0.   0.   0.   0.   0.
IC  C9  C4  C5  C6   0.   0. 180.   0.   0.
IC  N1  C5 *C6  N6   0.   0. 180.   0.   0.


The above IC table should be sufficient to build the molecule, and the dihedrals are meaningful for analysis of ring deformation, NH2 wobble, and bending across the bridge. To build Cartesian coodinates for all atoms:

read sequ FG1 1
gener F setup
ic param
ic seed F 1 C6 F 1 N1 F 1 C2
ic build
hbuild sele hydrogen end




Rick Venable
computational chemist

Re: Parameterizing adenosine-like drug molecule
lennart #25322 09/04/10 07:45 PM
Joined: Feb 2010
Posts: 29
Ascaris Offline OP
Forum Member
OP Offline
Forum Member
Joined: Feb 2010
Posts: 29
Lennart- I do have x,y,z coordinates, and 3DP9 is indeed the source... I had tried skipping the IC table business earlier, had run into some problems, and—concerned that the lack of an IC table might be the source—then proceeded to attempt to make such a table...

Rick, I will try what you've suggested right away, including this table. Thank you very much for making this table and spelling things out. Much obliged!

At the same time, I think I'd better figure out how to do this without an IC table, too, for future endeavors. smile

I appreciate all of your help and patience.

Many thanks!


Intestinal nematode
Bronx, NY
Re: Parameterizing adenosine-like drug molecule
Ascaris #25323 09/04/10 08:35 PM
Joined: Sep 2003
Posts: 4,797
Likes: 2
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 4,797
Likes: 2
Without the IC table:

read rtf card name ...
read para card name ...
read sequence fg1 1
generate fg1
read coor pdb name fg1.pdb
hbuild sele hydrogen end

You may of course still have problems with missing parameters.


Lennart Nilsson
Karolinska Institutet
Stockholm, Sweden
Re: Parameterizing adenosine-like drug molecule
Ascaris #25327 09/04/10 11:56 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
Originally Posted By: Ascaris
Ah, yes, I believe I did use ic generate to create this IC table (because of what Kenno said in his tutorial about being able to use that command to generate an IC table).
To tell you the truth, my arm was twisted to put that "IC GENERATE" suggestion in the tutorial, and I hope the present thread will help me obtain permission to drop it. People relying blindly on IC GENErate already caused me a lot of headaches and I personally encourage all new CGenFF developers to build their IC table by hand. If you imagine the seed atoms already being in place, and you see the IC table as a recipe for placing all the other atoms, one by one, one per line, it becomes absolutely trivial. One only needs to understand the unconventional (but convenient) way impropers are defined in the IC table.

Originally Posted By: lennart
read coor pdb name fg1.pdb
In my experience, this will not work unless all fields in the pdb file are exactly what CHARMM expects them to be. And if your residue name is longer than 4 characters, that's not even possible. And CHARMM offers no way I know of to override these checks. We commonly use the following workaround:
  • Write out a crd file immediately after the "generate" command
  • Use the "system" command to call a script that replaces the 999 coordinates in the crd file with real coordinates from an external source (watch out for the order of the atoms!)
  • Read the resulting crd file back in.

Re: Parameterizing adenosine-like drug molecule
Kenno #25348 09/09/10 05:06 PM
Joined: Feb 2010
Posts: 29
Ascaris Offline OP
Forum Member
OP Offline
Forum Member
Joined: Feb 2010
Posts: 29
Thank you very much, Lennart and Kenno. I did end up needing to use Kenno's workaround, but that *did* work. smile

Kenno wrote:
Quote:
If you imagine the seed atoms already being in place, and you see the IC table as a recipe for placing all the other atoms, one by one, one per line, it becomes absolutely trivial.

Aha. Yes, now that I see this, it does seem trivial. And I can see how Rick came up with the above table too. Ok, great. Thanks, everyone, for all the help.


Intestinal nematode
Bronx, NY
Re: Parameterizing adenosine-like drug molecule
Ascaris #25349 09/09/10 10:52 PM
Joined: Sep 2003
Posts: 4,797
Likes: 2
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 4,797
Likes: 2
And was the structure in the end similar to the experimentally obtained structure?


Lennart Nilsson
Karolinska Institutet
Stockholm, Sweden
Page 2 of 2 1 2

Moderated by  alex, lennart, rmv 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.4
(Release build 20200307)
Responsive Width:

PHP: 5.6.33-0+deb8u1 Page Time: 0.015s Queries: 31 (0.008s) Memory: 0.9706 MB (Peak: 1.0882 MB) Data Comp: Off Server Time: 2020-10-31 20:15:28 UTC
Valid HTML 5 and Valid CSS