[phenixbb] phaser_x.mtz, resolve_x.mtz and exptl_fobs_phases_freeR_flags.mtz give different Rfree's using same model?

Tom Terwilliger terwilliger at lanl.gov
Tue May 20 08:47:04 PDT 2008

Hi Francis,

I don't know how to explain the refinement results, but here are some comments:

1. if you use defaults in autosol, then resolve_xx.mtz will have all 
zeros for the HL coeffs. If you want the HL coeffs you have to say


2. I'd recommend refining using the HL coeffs from phaser_xx.mtz or 
solve_xx.mtz, but yes, you can refine against the HL coeffs from 
resolve_xx.mtz and it seems to work ok. Paul Adams did a test a while 
back and the two worked about the same if I recall correctly. My 
logic for using experimental phases in MLHL refinement is that this 
is the independent information that is available; anything from 
density modification as information about where the model is located 
and flatness of solvent mixed in with the experimental information, 
and this duplicates information from the refinement process.

3. So in general I'd use  exptl_fobs_phases_freeR_flags_xx.mtz from 
autosol, or if you don't build a model in autosol, 
exptl_fobs_phases_freeR_flags.mtz from autobuild.

All the best,
-Tom T

At 09:17 AM 5/20/2008, you wrote:
>Ok. I think I have a case where I can't refine unless I use density
>modified experimental phases.
>I did an autosol run where I specified build=False so I don't have an
>exptl_fobs_phases_freeR_flags.mtz, but I do have phaser_1.mtz and
>However, when I refine my model using mlhl my Rfree drops below 30
>when using resolve_x.mtz and not phaser_x.mtz nor using the original
>d*trek ref using the same model? Using any of the other mtz's results
>in Rfree's in the mid 30s. Interestingly refining against the original
>d*trek results in an increase of Rfree with macro cycle?
>I checked the log file for resolve_x.mtz and although it says it's
>using the mlhl refinement target, it seems to be using FP, SIGFP (not
>the HL coefficients that I thought phased refinement would use) as
>evidenced by the .log file under 'Non-default parameters'. In this
>section it doesn't read in the HL cofficients? Is there any indication
>that it's actually refining against the phases?
>Any insight as to which mtz to use for final refinement would be
>What is the refinement target ml_sad?
>Francis Reyes M.Sc.
>215 UCB
>University of Colorado at Boulder
>gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D
>8AE2 F2F4 90F7 9640 28BC  686F 78FD 6669 67BA 8D5D
>phenixbb mailing list
>phenixbb at phenix-online.org

Thomas C. Terwilliger
Mail Stop M888
Los Alamos National Laboratory
Los Alamos, NM 87545

Tel:  505-667-0072                 email: terwilliger at LANL.gov
Fax: 505-665-3024                 SOLVE web site: http://solve.lanl.gov
PHENIX web site: http:www.phenix-online.org
ISFI Integrated Center for Structure and Function Innovation web 
site: http://techcenter.mbi.ucla.edu
TB Structural Genomics Consortium web site: http://www.doe-mbi.ucla.edu/TB

More information about the phenixbb mailing list