[phenixbb] DEN refinement

[Jeff Headd]

Hi Jan,

Did you run your DEN refinement with optimize=True? If you haven't run the
grid optimization, it's unlikely that you found the optimal gamma/weight
pair for your specific case.

Also, I would recommend not using reference model restraints while running
DEN refinement. This could potentially prevent your model from "deforming",
if you are restraining it to be the same as a reference model.

To use a reference file with DEN, you would use:

den.reference_file="path_to_reference_file"

That will handle the chainID/resID matching automatically, and correctly
use your reference in a DEN context.

Jeff


On Wed, Oct 2, 2013 at 4:43 PM, Jan van Agthoven <janccp4 at gmail.com> wrote:

> Hi everyone,
> I ran a DEN refinement of my low resolution structure (3.6) to improve
> my model. I typed in the following command:
>
> phenix.den_refine model.pdb data.mtz den.reference_file=reference.pdb
> reference.params
>
> with reference.params saying:
>
> refinement.reference_model.reference_group {
>      reference = chain A and resseq 1:953
>      selection = chain A and resseq 1:953
>    }
> refinement.reference_model.reference_group {
>      reference = chain B and resseq 1:687
>      selection = chain B and resseq 1:687
>    }
>
> because the reference file only contains chain A and chain B.
>
> Unfortunately I obtained much worse R factors then with a regular
> refinement:
>
> Start R-work = 0.2798, R-free = 0.3176
> Final R-work = 0.3733, R-free = 0.4089
>
> Does someone have an idea what could have go wrong?
>
> Thanks,
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20131002/d11e8ebe/attachment.htm>