How can I get phenix to spit out not just PhiModel but an estimated FOM or fake HL coefficients (without experimental phases - just based on how badly my model agrees with FOBS)? And I'd like the weights in their own personal mtz column(s), not just precombined in a "weighted phi" column. In case anybody cares why, I'd like to use solvent massaging and domain-wise averaging to remove bias from a low-resolution model-phased map, preferably with DM, unless anybody has a better suggestion. I did rigid body, TLS, etc on the model in phenix, so it seems like phenix has already calculated all the info I want if I could only squeeze it out. thanks, Phoebe ===================================== Phoebe A. Rice Dept. of Biochemistry & Molecular Biology The University of Chicago phone 773 834 1723 http://bmb.bsd.uchicago.edu/Faculty_and_Research/01_Faculty/01_Faculty_Alpha... http://www.rsc.org/shop/books/2008/9780854042722.asp
On Mon, Mar 21, 2011 at 2:23 PM, Phoebe Rice
How can I get phenix to spit out not just PhiModel but an estimated FOM or fake HL coefficients (without experimental phases - just based on how badly my model agrees with FOBS)? And I'd like the weights in their own personal mtz column(s), not just precombined in a "weighted phi" column.
It wouldn't surprise me if there is some command that already does this that I don't know about. However, it looks like we have almost everything we need to do this in CCTBX, and we're only missing a command to actually output the desired data. I threw together a script that appears to do the job, but I'm not 100% confident that the output is correct. I'll add it as "phenix.fmodel_phases" as soon as I can verify that there aren't any stupid coding errors. (You're not the first person to ask for this, but it wasn't obvious to me before how to write it.) If there is an equivalent command in CCP4 that does this, I'd recommend using that for now. (I'm pretty sure that Resolve can do this too, but I don't know the input keywords to use for running it without density modification.) -Nat
Hi Phoebe,
How can I get phenix to spit out not just PhiModel but an estimated FOM or fake HL coefficients (without experimental phases - just based on how badly my model agrees with FOBS)?
regardless why one may want to do this (output HL and FOM computed from a model), I will add this functionality to phenix.fmoodel once I'm back to office on Wednesday. I see how it can be useful... So it should be available be the end of this week/beginning of next week. Nat: yes, you are right, all the pieces are there: it is a method of fmodel (see mmtbx/f_model.py) - one just needs to put them together along with a test. Pavel.
participants (3)
-
Nathaniel Echols -
Pavel Afonine -
Phoebe Rice