Hi Ruben,

The cryst1 card of the pdb is the next: CRYST1 98.500 104.050 415.200 90.00 90.00 90.00 P 21 21 21

I tried phenix.pdbtools --geometry-regularization again after adding this CRYST1 card. The result looks good, i.e. symmetry interactions don't seem to be the problem.

For non-bound interactions the distance is sometimes less than the VDW-radius. Could this be the problem?

No, I don't think so. Your histogram of nonbonded interaction distances looks reasonable.

The log file can be downloaded here: files.me.com/ruben.vandermeeren/9tuifv

Clearly, the .cif file is read, otherwise phenix.refine would stop very early on. Sorry, I still don't know what's going wrong. Could you send me the phenix.refine inputs for debugging (to me directly, not to phenixbb)? Ralf