Previous Thread
Next Thread
Print Thread
Page 1 of 2 1 2
Androstenedione parameters
#25620 10/15/10 04:29 AM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Dear charmmers,

I am going to parametrize a compound from androgen family namely androstenedione. I am going to adopt the cholesterol parameters for this compound. My question is how to get the parameters for the ketone oxygen attached to the benzene carbon in C3 and C17 positions. Is there any compound in the data bank that I could use for this purpose? It is grateful if anyone could give me a clue how to go about it.

Thank you very much

Manori

Re: Androstenedione parameters
Manori #25635 10/17/10 08:25 PM
Joined: Sep 2003
Posts: 8,507
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,507
Steroids are (mostly) not aromatic, so I don't think you should be looking for oxygen double bonded to a "benzene" carbon. Any ketone might be a good starting point (subject to testing) for C17, while for C3, a ketone conjugated to a C=C double bond would probably be better as an initial model. Always keep in mind that your final results may well depend on how careful you are when defining a new molecule; be critical, and test your assumptions.


Rick Venable
computational chemist

Re: Androstenedione parameters
rmv #25641 10/18/10 06:00 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
What Rick said. The most important model compound to be parameterized in this case would be butenone. We recently did phenylmethylketone (not released yet), which could be used as a source of initial guess parameters for butenone. But a potential energy scan around its rotatable bond would still be mandatory.

Brushing up on organic chemistry and basic math may also be helpful in preparation of the parametrization.

Re: Androstenedione parameters
rmv #25642 10/18/10 06:33 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Dear rmv,

Thank you very much for the information. I have question. Why do you suggest C=C bond for C3 and not a any ketone for that too?

Manori

Re: Androstenedione parameters
Manori #25643 10/18/10 06:37 PM
Joined: Sep 2003
Posts: 8,507
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,507
The partial charge distribution and the behavior of the single bond between the two double bonds are expected to differ from a ketone which is not conjugated to a double bond. Because of the ring constraint, cyclohexenone might be a better QM model for the A ring of a steroid.


Rick Venable
computational chemist

Re: Androstenedione parameters
rmv #25646 10/19/10 05:29 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
Originally Posted By: rmv
cyclohexenone might be a better QM model for the A ring of a steroid.
True, but if we want to treat the "-enone" as a new functional group (which I think is justified), the best starting point would be butenone, where we can study this functional group in the absence of any external influences. That gives us generic (and hopefully transferable) parameters for enones; this is important for a general force field as we have learned by getting burned because of not doing the more generic cases first. After doing that, we can move on to cyclohexenone, then build the whole compound. If the speed of the parameterization process is more important than the precision, I think it would be safer to skip the cyclohexenone step than to skip the butenone step.

Re: Androstenedione parameters
Kenno #25660 10/21/10 05:57 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Dear all,

Greatly thankful for both of you to clarify things. I have several more things bothering my mind,

1) I think the reason you are suggesting butenone because if I use Acetone (already in there) I couldn't get the dihedral angle. (Please explain whether I am correct) Can I start with Acetone to parametrize butenone?

2) Once you said that you already parametrized butenone. Do you think you can share these parameters with me?, then I can acknowledge you in our publications if any (anyway we will refer your papers)

3) Suppose I parametrize the butenone and obtain all the bond, angle, dihedral and charges, can I still use the rest of the cholesterol structure for my androstenedione.

4) Since cholesterol has different atom types how should I put it in the cgenff files.

Greatly appreciate your help.

Thank you very much

Manori

Last edited by Manori; 10/21/10 06:05 PM.
Re: Androstenedione parameters
Manori #25661 10/21/10 06:27 PM
Joined: Sep 2003
Posts: 8,507
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,507
There's also the conjugation of the double bonds in the butenone system, which makes it quite distinct from acetone (or butanone).

Upon reflection, I now think that I would probably do QM calculations for cis- and trans-pentenone, with cis as a model for the conjugated C=C-C=O in a ring.

Also, some steroids do have an aromatic A ring, but those have an -OH at C3 (phenol analog).


Rick Venable
computational chemist

Re: Androstenedione parameters
rmv #25664 10/21/10 08:45 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Thanks rmv,

I understand that we have to parametrize double bond conjugated to ketone. Can I use the parameters from C7H7NO 3-acetylpyridine for this purpose because it has this property.

I am kind of getting confusing right now.

So, what you have suggested is,

starting from phenylmethylketone (actually Kenno has suggested this) as initial paras for butenone. And then parametrize cyclohexenone using butenone and the parametrize the final compound.

This is a very long process (may be I understood wrong). Do I have to do all the parameters or only the dihedral?

I would be grateful if Kenno or yourself could explain this to me in details step by step the process of parametrization.

This is a great help

Thanks both of you very much

Re: Androstenedione parameters
Manori #25665 10/21/10 10:10 PM
Joined: Sep 2003
Posts: 8,507
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,507
I believe Kenno has a tutorial available for using CGenFF to add new molecules; see the "sticky" post at the top of this forum.

As I understand it, at the minimum you need to do a torsion scan of the central C-C single bond in butenone, to get dihedral parameters. Unless you have target data (expt), you could skip cyclohexenone.

I'm less certain about the charges, without some QM results as a guide.

I don't think the 3-acetylpyridine is a very good model for this.


Rick Venable
computational chemist

Re: Androstenedione parameters
rmv #25666 10/22/10 03:17 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Dear Kenno,

Can I use your parameters for phenylmethylketone to start parametrize the butenone? It would be great if you could let me know whether you could share it with me. Or else direct me to a good compound which is already in the database to begin with.

Thanks

Manori

Re: Androstenedione parameters
Manori #25667 10/22/10 04:36 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
Originally Posted By: Manori
4) Since cholesterol has different atom types how should I put it in the cgenff files.
Actually, there's a cholesterol topology in CGenFF as well so you can represent your entire molecule with CGenFF.

I would recommend doing a full parameterization - charges, equilibrium geometry, vibrations and potential energy scan - on butenone, (or, for even better accuracy and transferability, cis and trans 3-penten-2-one, as suggested by rmv). Must-reads before doing this are:
  • The CGenFF paper
  • The CGenFF tutorial
  • The CGenFF FAQ
  • Poker's molvib tutorial
  • Threads by various people on this forum who have successfully parameterized molecules


An experienced CGenFF worker can get through the whole parameterization procedure in one or two days. However, the first time usually takes much longer, about two weeks on average. And that's for people working in our group; if your only source of support is this forum, it may easily take you 3 or 4 weeks. I'm telling you this because the enone moiety is actually on the to-do list of a grad student in our lab. It's very difficult for me to estimate when she will be done with it because my boss can give her a different and urgent assignment any time, but if nothing goes wrong, she might be finished with it by the beginning of December. So you have the choice between committing a few weeks to this parameterization and be reasonably sure to have your parameters somewhere mid or end November, or gambling that everything will work out well at our end and that you will get the parameters from us by mid-December (but I have to stress again that I can make no guarantees whatsoever).

Re: Androstenedione parameters
Kenno #25688 10/26/10 06:45 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
Let me know what your decision is. If you are going to do the parametrization yourself, I can pull together the phenylmethylketone parameters for you to use as initial guess.

Re: Androstenedione parameters
Kenno #25690 10/27/10 09:20 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Dear Kenno,

I am going to start parametrize butenone. Once your graduate student finish the job please send me the data. Then I can compare to see my capabilities. I would be grateful if you could give me the initial para for this. Also, I have a difficulty of determining what atom types to use for the double bonded carbon in butenone. Can I use CG2RC7 for the double bonded carbon? It would be great if you could give a hint on this.

Thanks for all the valuable advice and help


Manori

Last edited by Manori; 10/28/10 03:54 AM.
Re: Androstenedione parameters
Manori #25691 10/27/10 09:45 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Dear Kenno,

Why we need to parametrize both cis and trans of 3-penten-2-one? I thought the cis conformation gives a half of the ring and it is what we need to do. I do not understand why need trans.

Thanks

Manori

Re: Androstenedione parameters
Manori #25701 10/28/10 05:31 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
Start with butenone. I ran it through the automatic assignment program, then replaced select charges and parameters with values from phenyl methyl ketone. The atom types in this file are correct; the charges and parameters should be considered an "initial guess". The most important thing to re-optimize is the rotatable C2-C3 dihedral.

Attached Files
beon_adjusted.str.txt (2.93 KB, 305 downloads)
Re: Androstenedione parameters
Kenno #25703 10/28/10 06:37 PM
Joined: Sep 2004
Posts: 251
M
Manori Offline OP
Forum Member
OP Offline
Forum Member
M
Joined: Sep 2004
Posts: 251
Thank you very much for your help. I am going to go ahead and use this as a initial para for butenone. I am thinking of doing a full optimization of charge, bonds and angles, and the c2-C3 dihedral. Please tell me is there anything specifically that I need to keep in mind when I am doing this.

Thanks again.

Manori

Re: Androstenedione parameters
Manori #25706 10/29/10 04:24 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
Actually, it would be an advantage if you could keep all charges except C2, O2 and C3 at their standard values during charge optimization, if the water interactions are not too bad (if they are, changing C4 can be considered). This is not set in stone, but it would make life somewhat easier when moving to the larger compounds.

Otherwise, the compound should be pretty regular and well-behaved.

Re: Androstenedione parameters
Kenno #26002 12/14/10 05:51 PM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
We optimized butenone (BEON) and added it to the new 2b6 release of CGenFF. I will make an formal announcement on these forums tomorrow, but you can already download it here.

Re: Androstenedione parameters
Kenno #26028 12/21/10 12:41 AM
Joined: Dec 2005
Posts: 1,535
Forum Member
Offline
Forum Member
Joined: Dec 2005
Posts: 1,535
As for the larger model compounds, you can submit them to the CGenFF program an see how high the penalties are. This is a very convenient way to assess how well a given molecule is supported by CGenFF, and which parts of the molecule are most in need of optimization.

Page 1 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: 0.022s Queries: 55 (0.012s) Memory: 1.0878 MB (Peak: 1.3108 MB) Data Comp: Off Server Time: 2020-10-31 18:40:42 UTC
Valid HTML 5 and Valid CSS