[phenixbb] Can Phenix fix rotamers for several residues only?

[Nathaniel Echols]

On Sat, Aug 28, 2010 at 9:54 AM, Edward A. Berry <BerryE at upstate.edu> wrote:
> Is "fix_rotamers" a strategy? Can I limit it to a selection of residues?

I think it's only global right now.

> I get the traceback quoted below.
> If this is a memory problem, would fixing fewer rotamers help?
>  File "/raid/bin/phenix-1.6-289/cctbx_project/cctbx/miller.py", line 3016,
> in __init__
>    conjugate_flag=True)
> MemoryError

Sorry, it is crashing during an FFT, which is required to generate the
maps used for fitting, and I believe this has to be done over the
entire unit cell.  I suspect there are some inefficiencies in the
code, because I've seen this same error before in GUI bug reports.
(For instance, the grid spacing is 1/4 of resolution, and changing to
1/3 might fix the problem - although it will probably also have a
negative effect on the real-space refinement.)  Regardless, the only
solution is to upgrade the RAM, and since you already have 2GB
installed, you will also need to upgrade to a 64-bit OS and Phenix
installer if you aren't already using that.  (This assumes that your
processor is compatible, of course.)

PS. Please update to version 1.6.4, I believe Pavel made some major
improvements to the rotamer fitting since 1.6 was released (and of
course there are many other new features/fixes as well).

-Nat