Dear all, I am using phenix.refine for the first time to refine a 4.5A structure solved by MR with individual sites (stretching things a bit, I know) and a grouped ADP per domain. Below are the R-factor per cycle stats: REMARK ------------------------------------------------------------------------ REMARK R-factors, x-ray target values and norm of gradient of x-ray target REMARK stage r-work r-free xray_target_w xray_target_t REMARK 0 : 0.6464 0.6186 7.101344e+00 7.091899e+00 REMARK 1_bss: 0.3955 0.4006 6.825652e+00 6.806242e+00 REMARK 1_xyz: 0.3200 0.3622 6.675328e+00 6.720706e+00 REMARK 1_adp: 0.3201 0.3622 6.674498e+00 6.719426e+00 REMARK 2_bss: 0.3539 0.3804 6.736883e+00 6.752128e+00 REMARK 2_xyz: 0.3067 0.3427 6.639056e+00 6.687071e+00 REMARK 2_adp: 0.3062 0.3423 6.638917e+00 6.687168e+00 REMARK 3_bss: 0.3357 0.3685 6.706475e+00 6.735398e+00 REMARK 3_xyz: 0.3082 0.3469 6.652536e+00 6.702364e+00 REMARK 3_adp: 0.3076 0.3468 6.652984e+00 6.703972e+00 REMARK 3_bss: 0.3286 0.3659 6.701854e+00 6.743421e+00 REMARK ------------------------------------------------------------------------ You can see that after the intial scaling, each cycle of bulk solvent scaling and anisotropy correction leads to an increase in R-factors of ~2%. Has anyone seen this behaviour before and maybe suggest a solution? Alternatively, is there a way I can turn off the final cycle of bulk solvent scaling before outputting the pdb and maps? Thanks in advance, Steve Glynn