Dear colleagues,

 

I am trying to refine a protein-DNA complex. In this structure, 2 DNA double strands are present with 50% occupancy (or similar occupancy’s) and overlap at certain positions. Phenix.refine crashed  due to a nonbonded interaction distance issue. To fix that, I tried to exclude the DNA strands using:

 

pdb_interpretation {
 custom_nonbonded_symmetry_exclusions="chain E"
 custom_nonbonded_symmetry_exclusions="chain F"
 }

 

Nevertheless, phenix.refine crashed with the same error message. Changing the “nonbonded_distance_threshold” works, but results in terrible statistics (since this applies for the complete structure, I guess).

 

Is there a way to exclude only the DNA strands (in regard to the clash guard) or a similar approach to refine the structure?

 

Thank you in advance for your help,

 

Tobias