[Avogadro-devel] [ avogadro-Bugs-3372106 ] Error during optimization with gaff

[SourceForge.net <no...@so...>]

Bugs item #3372106, was opened at 2011-07-20 10:35
Message generated for change (Tracker Item Submitted) made by gustafdo
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=835077&aid=3372106&group_id=165310

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Open Babel
Group: v 1.0.3
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Gustaf Olsson (gustafdo)
Assigned to: Geoff Hutchison (ghutchis)
Summary: Error during optimization with gaff

Initial Comment:
Sorry if this turns up two times, I could not find my submission after the last time and I also forgot my name and email address.

My name is Gustaf Olsson (gus...@ln...) 

I am running Avogadro v.1.0.3 on Mac OS X 10.6.8

First of all, thank you for providing this excellent pice of freeware! One of the best molecular visualizing, building and analyzing packages I have tried so for.

Though I have encountered a problem/bug that I cannot seem to sort out. It seems that GAFF has become a \"reactive\" forcefield in the latest version and I\'m pretty sure that it is not supposed to be!

This is what I\'m doing:

	File -> Import -> Fetch chemical structure…
		-> methacrylic acid (figure1, attached file)

(Though it does not matter if I draw the molecule myself, the same error occurs)

	-> Auto Optimization Tool (Energy minimization)

I then continue to run an energy minimization using the GAFF forcefield (since I\'m running AMBER for MD simulations using GAFF) with 4 steps per update and running either \'Steepest Descent\' or \'Conjugate Gradients\' (does not matter which method). 

This is where the problem occurs, the oxygen atoms of the carboxyl group suddenly attract each other and the carbonyl oxygen atom of the group swallows the acidic proton, forms a new bond to the \"hydroxyl\" oxygen and the minimization stops (figure2, attached file). This should not happen, from my experience with GAFF it should even be impossible for the FF to actually represent bond formation.

If I delete the \"hydroxyl\" oxygen atom (indicated in figure2, attached) the acidic hydrogen reemerges and the energy minimization continues with the incorrect structure (figure3, attached). It does not matter if I elongate the carbon chains, the carboxylic acid group continues to display this behavior. 

Interestingly, when importing benzoic acid instead of a hydrocarbon chain acid, the problem disappears. When performing the minimization/optimization for this structure the acid functionality behaves the way one would expect (figure4, attached).

I found the gaff.dat file and tried replacing it with the newer GAFF (2010) file from my amber distribution and removing the headings (MASS, ANGLE and so on…) but this gave the same result as with the previous GAFF version so I guess this is somehow related to the Avogardo distribution itself. I cannot remember that I had this problem with previous version of the software, though I cannot find the DMG for v. 1.0.2 so I cannot confirm this. I should add that the problem does not occur with any of the other available forcefields either.

Any help/solutions/input on this matter would be greatly appreciated!

 

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=835077&aid=3372106&group_id=165310