Re: [ccp4bb]: Geometry problems in Refmac

["Eleanor J. Dodson" <ccp4@ysbl.york.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***



Ben Sandler wrote:
> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
> 
> 
> It appears now that my ligand geometry problems arise from two sources:
> 
> * I'm refining a low (3.1A) resolution structure, and I didn't have the 
> matrix weight low enough.  Now that I've set it to 0.1, things are 
> looking much better.  Is this roughly the weight I should be expecting 
> to use for 3.1A?


  You really have to check the RMSD for bonds etc - this should stay 
about the SD - viz 0.02

> 
> * Even with the weight set to 0.1, I'm still having problems with a 
> hydroxyl group in the ligand.  The hydroxyl is sandwiched between two 
> iodines, and there's no electron density for it.  (I'm assuming that 
> this is the result of series termination errors stemming from the two 
> iodines having such dense electron density.)  As a result, the hydroxyl 
> is strongly pulled into the electron density of one of the iodines, 
> causing a consistent bend of about 10 degrees.  Is there any good way to 
> prevent this from happening?  The most obvious seems to be to set the 
> occupancy of the hydroxyl very low until the very last cycle of 
> refinement, but that doesn't feel entirely kosher to me.


Hmmm - you may need a LINK record to describe the hydroxal - IODINE bonding.
If you dont describe it - there  will be VDW repulsions invoked 
automatically. Check the output to see what the program thinks of your 
geometry - turn on MONITOR MANY and look for all mentions of the ligand..

Eleanor

> 
> -Ben Sandler
> 
>