[phenixbb] Phenix.ensemble_refinment questions

Tom Burnley tom.burnley at gmail.com
Thu Nov 13 07:18:45 PST 2014

Hi Joseph,

1) IIRC the fully phased SFs weren't deposited and we asked the
authors for them.  Whilst I was developing ER we hadn't allowed for
phase information to be used in the target function.

2) It is possible that a optimal parameters are found by grid
searching all three parameters at once.  If no value for tx is
supplied it will estimate Tx based on resolution and this is typically
very reliable.

3) With a cluster I would suggest using the auto Tx setting as above
and running a grid of temperature offsets and ptls values: e.g 12
parallel of 1, 2.5, 5 (temp) and 0.95, 0.9, 0.8, 0.6 (ptls).

More details can be found here:

I haven't been actively developing ensemble refinement for a while now
so other may have more up-to-date answers....

Hope that helps,


On 13 November 2014 14:46, Joseph Brock <joseph.brock at ki.se> wrote:
> Dear Phenix community,
> I was hoping that some of you may be able to clarify some questions about
> Phenix.ensemble_refinement:
> 1. In the associated publication (Burnley et al. eLife 2012;), the ensemble
> refinement is validated by comparing the correlation of the ensemble
> generated map, with the map generated from the experimental phases for PDB
> entry 1YTT ("The overall correlation coefficient between the
> electron-density map from the ensemble model (obtained without experimental
> phases) and the experimentally phased electron-density map was 0.903,
> compared with 0.873 and 0.895 for the published and re-refined single
> structures. These seemingly small improvements in over- all quality
> indicators allow for significant local improvements."). I am confused how
> one computes an experimentally phased from structure factors deposited in
> the PDB that contain only anomalous intensities/amplitudes and not
> Hendrickson-Lattman coefficients. Is there a program within the phenix
> package that can do this?
> In addition, the phenix.ensemble_refinement keywords include:
> experimental_phases
> file_name = None
> labels = None
> Is it possible to include experimental phases during the rolling average
> refinement process and could this be beneficial (if the phases were of a
> sufficient quality)?
> 2. Is the correct procedure in parameter optimisation to optimise pTLS, then
> temperature offest and then tx? Or is it possible that the best ensemble may
> be generated from any (non-directional) combination of the above parameters?
> 3. What is the function of the "nproc" keyword? If this is the number of CPU
> cores that can be used in parallel, what is the most efficient way of using
> phenix.ensemble_refinement on a cluster?
> 4. Can anybody suggest a good way to compare the ensembles generated by
> phenix.ensemble_refinement with the trajectories of traditional MD
> simulations (for e.g. produced by NAMD? ;)
> Finally, I noticed that I cannot run phenix.ensemble_refinement using a
> "my_parameters.eff" file, it is necessary to type on the command line.
> Alternatively, the queueing system for the supercomputer I use is managed by
> SLURM, however, I cannot submit batch jobs via the appropriate command
> sbatch with anything other than default parameters (keywords in a
> parameters.eff or that are typed explicitly following the mycoords.pdb
> mydata.mtz myrestraints.cif files are not incorporated into the resulting
> parameters of the .log file).
> Could these issues be due to my local environment setup?
> I am running Phenix 1.9-1692 Platform: intel-linux-2.6-x86_64 redhat-e6.5.
> Sorry for the long email and many thanks in advance for you help!
> Best regards,
> Joseph.
> Joseph Brock | PhD
> Division of Physiological Chemistry II
> Department of Medical Biochemistry and Biophysics
> Karolinska Institutet
> Scheeles väg 2
> SE-171 77 Stockholm, Sweden
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb

More information about the phenixbb mailing list