New year plans

One of the things I plan to start doing is "xia2 howtos" here on the blog - interesting test cases and how to resolve them. On my hit list are:

• working with weak data
• handling multi-crystal data sets

Any other suggestions? Please post in the comments section.

Cryptic error message from xia2 which means "you need new cctbx"

Happy new year!

Had this error reported a couple of times, which results in xia2 needing new features in cctbx - particularly a problem if you use e.g. Phenix 1.8.2.

Error ends:

  File "/Applications/PHENIX-1.8.2-1309/Contents/phenix-1.8.2-1309/cctbx_project/libtbx/phil/__init__.py", line 1906, in process_includes
    phil_path=phil_path).objects)
  File "/Applications/PHENIX-1.8.2-1309/Contents/phenix-1.8.2-1309/cctbx_project/libtbx/phil/__init__.py", line 1946, in process_include_scope
    where_str=object.where_str)
  File "/Applications/PHENIX-1.8.2-1309/Contents/phenix-1.8.2-1309/cctbx_project/libtbx/utils.py", line 873, in __init__
    error_prefix, module_path, where_str))
ImportError: include scope: no module Wrappers.XDS.XDSIntegrate (input line 4)

(not very helpful!) Best way to resolve this: install Phenix 1.8.4 or new cctbx nightly build. (I need to improve xia2 error reporting)

Sorry for any annoyance, Graeme

Interactive control of indexing images with XDS processing

Now incoming in a nightly build...

[gw56@ws133 interactive]$ xia2 -3dii -interactive -image /data1/gw56/i04-bag-training/th_8_2_0001.cbf
Environment configuration...
XIA2_ROOT => /home/gw56/svn/xia2.trunk
XIA2CORE_ROOT => /home/gw56/svn/xia2.trunk/core
CCP4 => /dls_sw/apps/ccp4/x86_64/6.4.0/6nov2013/ccp4-6.4.0
CLIBD => /dls_sw/apps/ccp4/x86_64/6.4.0/6nov2013/ccp4-6.4.0/lib/data
CCP4_SCR => /tmp/tmpOOG0z9
Working directory: /tmp/interactive
Free space:        15.02 GB
Host: ws133
Build: 4468
Contact: xia2.support@gmail.com
XIA2 0.3.6.3
Command line: xia2 -3dii -interactive -image /data1/gw56/i04-bag-training/th_8_2_0001.cbf
Existing images for indexing th_8_2_0###.cbf: 1-540
>1-100, 441-540
New images for indexing: 1-100, 441-540

Essentially allows you to assign which images to use for XDS indexing for the case where you have a "good bit" you want to use.

Move to git?

In the spirit of not breaking xia2 for users while wanting to make some substantial changes, have been thinking of moving from subversion to git: this would affect you if you like the most bleeding edge xia2 versions and pull xia2 from svn.

So: has anyone an opinion on this?

Incoming: parameters to control xds integration / spot profile

Adding more Phil parameters this morning to control reflecting range, beam divergence etc. however these are not for common use: only when the automatic determination of parameters has failed are these useful.

Please see XDS documentation for details:

http://xds.mpimf-heidelberg.mpg.de/html_doc/xds_parameters.html#REFLECTING_RANGE=

In build 4480 & later having phil file containing

xds.integrate.delphi = 10.0
xds.integrate.reflecting_range = 0.31
xds.integrate.reflecting_range_esd = 0.045
xds.integrate.beam_divergence = 0.25
xds.integrate.beam_divergence_esd = 0.025

will allow you to control these via -phil file.phil

Controlling XDS Refinement in Integration with Phil

By default in xia2 the REFINE= for integration refines ORIENTATION and CELL - however if you want to override this you can do so by using Phil parameters. Create a file

whatever.phil

containing

xds.integrate.refine=CELL
xds.integrate.refine_final=CELL

for example - and run with -phil whatever.phil. The former is the first pass, the second the pass when a refined orientation is used in case you want these to be different. You can select from ORIENTATION CELL BEAM DISTANCE AXIS - best to look at the XDS documentation if you want to change these.

In general you don't need to do this but some people would like the choice...

Very big but not functional commit coming

Warning to anyone who uses xia2 from subversion - forthcoming commit will be very large as I change the indentation from 4 spaces to 2 spaces, to make the code more consistent with the cctbx way of things. I have run a bunch of tests and manually inspected at least 1000 lines of the changes to make sure that they are sensible (and they appear to be) but if you have any local modifications in your code you may / will find updating via svn causes conflicts.

Any problems please contact me at the usual address xia2.support@gmail.com