wxu_scale is ignored with TLS
When using TLS, the wxu_scale is changed, even if it is given in the input file. IMHO, TLS should only modify the default value, and not a user-defined wxu_scale. Thanks, Joe Krahn
Hi Joe, thanks for the feedback! The way it works right now is that when TLS is combined with individual ADP refinement the wxu_scale factor is set up by the program and a user cannot change it. This is because at some point a few years ago I re-refined all structures in PDB where TLS selections are available and for each structure I obtained an optimal weight scale, wxu_scale. Then I derived a dependency of optimal wxu_scale as a function of resolution and this is what is now used. So it is supposed to be good in average, but of course there may be outliers. And once again, this is only specific to TLS+individual_adp refinement. It is in my to-do list to change this behavior so a user can still change the weight to whatever he/she likes. In fact this was suggested by many users in the past. Pavel. On 9/24/09 1:28 PM, Joe Krahn wrote:
When using TLS, the wxu_scale is changed, even if it is given in the input file. IMHO, TLS should only modify the default value, and not a user-defined wxu_scale.
Thanks, Joe Krahn _______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
Hi Pavel, I reprocessed data and got a new mtz file. I removed the r-free_flags remark from the pdb. Whe I try the refinement with the new .mtz I got this error. Sorry: R-free flags not compatible with F-obs array: missing flag for 1 F-obs selected for refinement: h k l data sigmas 6 1 0 1370 20.29 Maia
Hi Maia, if you got a new fobs data file, where does the new array of free-R flags come from? If you are using the old one, are you sure that the set of Miller indices (hkl) in the old free-R flags array is exactly the same as the set of Miller indices corresponding to the new fobs? I guess the root of the problems is here. That is there is no one-to-one correspondence between {h,k,l}_fobs<->{h,k,l}_flags... there may be a few extra or missing reflections, so you get the mismatch. Pavel. On 9/25/09 9:03 AM, Maia Cherney wrote:
Hi Pavel, I reprocessed data and got a new mtz file. I removed the r-free_flags remark from the pdb. Whe I try the refinement with the new .mtz I got this error.
Sorry: R-free flags not compatible with F-obs array: missing flag for 1 F-obs selected for refinement: h k l data sigmas 6 1 0 1370 20.29
Maia _______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
participants (3)
-
Joe Krahn -
Maia Cherney -
Pavel Afonine