Hi Jan, have you tried to reset the B-factors? I sometimes encounter that the minimizer is trapped in a local minimum and |phenix.refine|struggles to bring them back down. You might also try Optimize ADP weights. Hope this helps! Best regards, Georg. Am 03.10.2026 um 04:32 schrieb John Bacik via phenixbb:

Hi Jan, it might be helpful to provide some screen shots of the positive density you are observing. I also wonder if you have somehow introduced some model bias into your data. All the best,John On Friday, October 2, 2026 at 11:14:34 AM CDT, jan.abendroth--- via phenixbb wrote:

Hi all, I have seen this odd behavior with a couple of ligands in phenix 2.2.1-6174:

B-factors of some well-defined ligands increase significantly during a refinement, in the range of 2-fold compared with surrounding protein residues. The density hints for full occupancy - just about as strong as for the neighboring residues. As a test, I changed the occupancy of the ligand to 0.8.  Phenix refines back to 1.0. If I turn off occupancy refinement, B-factors for the ligand increase even more.

While the geometry looks fine, I am left with significant positive difference density. I have refined several structures for both targets with different ligands and never ran into this.

Has this been seen before and is there anything I could try? I have already tested versions 6190 and 6048 and see the same behavior.

Thanks, Jan _______________________________________________ phenixbb mailing list [email protected] To unsubscribe send an email [email protected] Unsubscribe: phenixbb-leave@%(host_name)s

_______________________________________________ phenixbb mailing list [email protected] To unsubscribe send an email [email protected] Unsubscribe: phenixbb-leave@%(host_name)s