Hi Abhinav,

thanks again for sending me the files off list. This let me to identify the problem and fix it. The difference between R-factors in phenix.refine and phenix.model_vs_data was due to a slightly different accounting for scattering contribution from H atoms. In both cases it was correct though. If I check in my changes today then next available nightly build will have it fixed.

Pavel

On 6/10/13 11:42 AM, Abhinav Kumar wrote:

I am getting a significant difference between the R-factors written by phenix.refine and phenix.model_vs_data.

1. Phenix.refine
phenix.refine "in.pdb" "Fobs.mtz" "refinement.input.xray_data.labels=FP SIGFP" "refinement.input.xray_data.high_resolution=2.50" "fix_rotamers=True" "ordered_solvent=True" "strategy=tls+individual_sites+individual_adp"

   Final R-work = 0.1829, R-free = 0.2141

2. Phenix.model_vs_data
phenix.model_vs_data out.pdb Fobs.mtz f_obs_label="FP" r_free_flags_label="FreeR_flag" high_resolution=2.50 comprehensive=true

    r_work(re-computed)               : 0.1808
    r_free(re-computed)                : 0.2099

How can I reproduce the phenix.refine R-factors in phenix.model_vs_data?

--
Thanks,
Abhinav
JCSG@SSRL, SLAC
(650) 926-2992