ADP refinement in phenix.real _space_refine
While refining a protein-DNA complex in a cryoEM map (~2.5-7A), I have noticed the following behaviour of ADP refinement: - If NCS is on for the protein chains, no ADP refinement of the DNA takes place. - If NCS is off, ADP refinement behaves oddly: one strand of dsDNA is refined with atomic ADPs, the other strand with group ADPs (one per nucleotide). The group ADPs take up unphysical values (lower ADPs in poor map regions), whereas the atomic ADPs look sensible. - To get the ADP refinement of DNA to behave normally, I need to both leave NCS out and set nproc=1 explicitly (haven’t tried other values). I wonder if any of this has been seen before or am I not doing this correctly? Phenix-dev-5330 iMac, Catalina 10.15.7, 8-core Intel i9 Best wishes, Luca Luca Pellegrini, PhD Department of Biochemistry University of Cambridge Cambridge CB2 1GA UK
Hi Luca,
While refining a protein-DNA complex in a cryoEM map (~2.5-7A), I have noticed the following behaviour of ADP refinement:
- If NCS is on for the protein chains, no ADP refinement of the DNA takes place.
- If NCS is off, ADP refinement behaves oddly: one strand of dsDNA is refined with atomic ADPs, the other strand with group ADPs (one per nucleotide). The group ADPs take up unphysical values (lower ADPs in poor map regions), whereas the atomic ADPs look sensible.
- To get the ADP refinement of DNA to behave normally, I need to both leave NCS out and set nproc=1 explicitly (haven’t tried other values).
I wonder if any of this has been seen before or am I not doing this correctly?
this needs a closer look to diagnose the problem. Would you mind sending input files as well as .eff and .log files off list and I will investigate. Also, please indicate if the cryo-EM map you used was symmetrized (a symmetry was implied during the reconstruction) Pavel
participants (2)
-
Luca Pellegrini -
Pavel Afonine