Re: [ccp4bb]: Refmac5 Make H All

[Garib Murshudov <garib@ysbl.york.ac.uk>]

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

David Buckler wrote:

> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
>
> Dear All,
>
> I am using Refmac5 through CCP4I 4.1.1 and am seeking advice on using the option
> Make H All.
>
> I note that Make H All is the default, but the documentation suggests using this
> toward the end of refinement and only with high resolution data.

>
> In my case, near the final stages of refinement with a 1.50 A dataset, I have
> applied Make H All for several cycles and reached R=18.4 and Rfree=21.6.  Using
> this same model and doing a single round of refinement from it with Make H No
> results in increases to R=19.6 and Rfree=22.3.
>
> Here are a few specific questions:
>
> 1. Is Make H All actually increasing the number of fitting parameters ("riding
> positions" suggests to me "no")?
>

No. Hydrogens are used for structure factor calculation and in geometry. They are
not refined in a usual manner.
They do not contribute to gradients and second derivatives

> 2. Can anyone provide specific suggestions for when to use this option in terms
> of data resolution -- or other general guidelines?

I have had good results (decrease of R and R-free and a bit better map without
refinement) at around 2.0 or slightly
less resolutions at the end stages of refinement. When atoms are positioned
correctly then hydrogens are predicted
correctly also.

>
> 3. If my above observation about effect on R is general, is it more proper to
> report the R's for a final refinement run with Make H No since H coordinates are
> not in the final model?
>
>

Header file says that hydrogens have been added in their riding positions. There is
ooption to output hydrogens
also (MAKE HOUT Yes). But it could be used just to add few more hydrogens if you
see them in the difference map.
I am not sure that they should be deoposited as hydrogen positions can be
calculated from other atoms.

Garib


--
------------------------------------------------------------------
Garib N. Murshudov, Chemistry Department, University of York, U.K.
Tel: Home +44 (1904) 42 62 87, work:  +44 (1904) 43 25 65
------------------------------------------------------------------