[phenixbb] Questions about phenix.fmodel

David A Case david.case at rutgers.edu
Sun Sep 10 08:00:32 PDT 2017


There are two things I'd like to do, for which phenix.fmodels look appropriate,
but which I've been able to do:

1. Run phenix.fmodel on a phenix-refined pdb file, and get the same FMODEL
parameters as phenix.refine itself put into its output mtz file.  Can I
extract from the log files the scaling factors that phenix.refine used, and if
so, how?

2. Run phenix.fmodel on a pdb file, and get (unscaled) FMODEL values that
include bulk solvent contributions as well as the ususal atomic-based
structure factors.

I think my problem arises from trying to understand the documentation at
phenix-online.org/documentation/reference/fmodel.html.  Specifically: the
equation at the very top of the documentation states:

   Fmodel = scale * exp(AnisoScale) * (Fcalc + k_sol * exp(-b_sol*s^2/4) *
Fmask)

But I had thought(?) that k_sol and b_sol were no longer being used.  I don't
see any reference to these in my phenix.refine log files.

Example 2 in the documentation page suggests using k_sol=0.35 and b_sol=50,
but I don't know how to find the "best" values for these, i.e. the ones that
would be closest to what phenix.refine is using.  If I try the values in
example 2, I don't get the same FMODEL values as phenix.refine reports.

Thanks in advance for pointers...dave case



More information about the phenixbb mailing list