[phenixbb] setting resolution for 2 datasets in AutoSol

Thomas C. Terwilliger terwilliger at lanl.gov
Mon Jul 9 20:36:49 PDT 2007

Hi Luca,

It's easiest to do this in 2 steps:

1a. You can set the resolution separately for each dataset in the AutoSol

In the GUI, just go through each dataset and set the overall resolution to
your desired values as you go.  Ignore the density modification and
building that occurs in this run.

1b. Then you'll get an output exptl_fobs_phases_freeR_flags_2.mtz or
equivalent that is phased to the resolution you specified, but that has
all the data (with no phases) to higher resolution.

2. Now run the AutoBuild wizard, and set the resolution to your
high-resolution limit.  That should use the phases you generated, and
include the high-res data for density modification and refinement.

Equivalently, you can run AutoSol at low resolution then run AutoBuild at
higher resolution by supplying an input_hires_file with the hires data.

A key point here is that the building in AutoSol is just a preliminary
build...the building in AutoBuild is the real thing.

You can also run 2 datasets and set the resolution for each in a script. 
Here is a script to read in 2 SAD datasets and solve them together as one,
setting the resolution of phasing to 3 A for the first and 2.9 for the

cat two_datasets.inp

input_seq_file seq.dat
resolution 2.9                             #  Resolution
expt_type       sad
input_file_list  se.sca
############## NEW DATASET ################
run_list        start                      # run "start" method.
run_list        read_another_dataset       # starting a new dataset here
resolution  3.1                            #  Resolution
input_file_list  br.sca

Let me know if that does not do it!
-Tom T

>> From: lp212 at mole.bio.cam.ac.uk
>> Date: July 9, 2007 3:38:28 PM PDT
>> To: phenixbb at phenix-online.org
>> Subject: script for Zn and Se phasing
>> Hi,
>> I have collected two peak datasets on two different crystals, for a
protein that contains zinc:
>> 1. peak dataset at the Zn edge (lambda=1.285A, 2.5A reso)
>> 2. peak dataset at the Se edge (lambda=0.98A, 2.8A reso)
>> The protein has 2 Zn and 11 SeMet, and 4/au. When I input both
>> datasets in the AutoSol GUI, Hyss finds correctly 8/8 Zn sites, and
~34/44 Se sites, but only when I cut back the resolution (4.2A for Zn, and
4.7A for Se). What I can't figure out is how to tell the Autosol GUI to
use a low-res cutoff for finding the sites, and then use all the data for
model building. Could someone please explain how to do this in the GUI, or
send me a script?
>> Thanks a lot! Apologies if this is trivial...
>> Luca

More information about the phenixbb mailing list