[phenixbb] Problem in reducing R-factor

[Ryan Spencer]

Hi Nat and Phenix community,

	For high resolution structures, below 1.5 Å, is there a set of rules
for best practices for refinement? 

	I've been toying with different parameters and found that the
weighting didn't make a huge difference, which is probably expected, with a
1.3 Å dataset. Do stereochemical/ADP weights have any effect?

With weighting – 
Rwork 0.1573 
Rfree 0.1806
RMS (bonds) 0.0073
RMS(angles) 1.468
Clashscore: 6.1
Ramachandran: 100
Ramachandran outliers: 0
Rotamer outliers: 0
CC ~0.94

Without weighting-
Rwork: 0.1603
Rfree:  0.1840
RMS (bonds) 0.0049
RMS(angles) 1.272
Clashscore: 7.6
Ramachandran favored: 100
Ramachandran outliers: 0
Rotamer outliers: 0
CC ~0.94

	For the particular data set I'm working with, I'm refining directly
against anomalous data, more specifically against 12 Iodines. Refining the
f' and f" has worked well, generating a beautiful density map.

 	Refining against the model with hydrogens increased the Rfree/Rwork
values by about 2 points. From what I've read previously I thought refining
hydrogens in a non-"riding" model was only valid below ~1.0 Å.

	The previous phenix notes
http://www.phenix-online.org/presentations/latest/pavel_phenix_refine.pdf ,
slide 23 has the combination of anisotropic B-factors with TLS refinement
for atoms. This combination worked well for reducing the Rfree/Rwork values
by a few points. For water, I assume that at this resolution there is a
combination of anisotropic (big banana looking density) and isotropic waters
and that one cannot simply rely entirely on the water_update option.

	Partial occupancy has been hit or miss, during one refinement it
will appear that there is density for an alternative conformation. After
adding the alternative conformation the density will disappear or show that
no density should be there. In other instances refinement of the partial
occupancies has worked well.

	The twin law has worked great, usually dropping Rfree/Rwork values
by 5 points or so.

	I apologize if any of this is common knowledge and shows my relative
inexperience when it comes to best practices for crystal refinement. 

Ryan Spencer
University Of California, Irvine
Nowick Research Group




-----Original Message-----
From: phenixbb-bounces at phenix-online.org
[mailto:phenixbb-bounces at phenix-online.org] On Behalf Of Pavel Afonine
Sent: Friday, October 11, 2013 12:17 PM
To: phenixbb at phenix-online.org
Subject: Re: [phenixbb] Problem in reducing R-factor

I'm *thinking* (not doing it right now) to make phenix.refine do something
of this sort  automatically if the resolution is really 1.3A (and not in
fact say 1.5 or so but data file has a couple of reflections nearing 1.3A,
which would tend people claim it is 1.3A resolution!) and r-factors are not
too far to expected ones given *actual* data resolution. Of course "1.3A" in
this case is going to be a range between
1 and 1.5A in terms of data quality.

Pavel

On 10/11/13 8:39 AM, Nathaniel Echols wrote:
> What happens if you try refining protein atoms (but not waters) with
> individual anisotropic B-factors instead of TLS?  It isn't guaranteed to
> work at 1.3Å, but it usually does (in my hands, anyway).
>
> Also make sure you have explicit hydrogens added; this will have a
> significant effect at this resolution.  And turn on the water update if
> you haven't already.
>
>
> On Fri, Oct 11, 2013 at 6:19 AM, Singh, Bishal
> <b.singh at dkfz-heidelberg.de <mailto:b.singh at dkfz-heidelberg.de>> wrote:
>
>     Hi all,
>       I am solving a crystal structure at 1.3Å. Completeness of the
>     data-set and Figure of merit at this resolution are 99.5% and 0.62
>     respectively. Structure has two molecules per asymmetric unit. I
>     used TLS parameters and NCS restraints. WXC_scale and WXU_scale are
>     0.50 and 1.0 respectively. Rwork and Rfree are 18.09 and 19.62
>     respectively. I tried with reducing WXC_scale and WXU scale to 0.25
>     and 0.30 respectively but it did not help. Please suggest me if I
>     need to change some parameters. I shall be grateful for your help.
>
>     Thanking you,
>     Bishal Singh
>     Graduate student
>     University of Heidelberg, Germany
>     _______________________________________________
>     phenixbb mailing list
>     phenixbb at phenix-online.org <mailto:phenixbb at phenix-online.org>
>     http://phenix-online.org/mailman/listinfo/phenixbb
>
>
>
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
>
_______________________________________________
phenixbb mailing list
phenixbb at phenix-online.org
http://phenix-online.org/mailman/listinfo/phenixbb

-------------- next part --------------
A non-text attachment was scrubbed...
Name: Density_map.png
Type: image/png
Size: 249898 bytes
Desc: not available
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20131011/f81629d4/attachment-0001.png>