[Bio-linux-dev] Upgrade to Bio-Linux 7 (at last)

[Tim Booth]

Hi Mike,

Thanks for reporting the problems with QIIME.  I've patched the code to
make it work with the newer RDP classifier and I've also fixed the issue
with KING not working.

I've also applied the patch that I saw on the QIIME blog to fix the
problem with compare_alpha_diversity reporting the wrong t-statistic.

The new packages are now on the package server for BL7.

Cheers,

TIM

On Wed, 2012-11-07 at 16:12 +0000, Michael Cox wrote:
> Hi Tim,
> 
> 
> it looks like something else is going on too - King is now not working
> within the pipeline to generate the kinemage files for PCoA analysis.
>  Here's the error:
> 
> 
> Qiime> beta_diversity_through_plots.py -i otus/otu_table.biom -m
> aran.r1.map.txt -o wf_bdiv_even692/ -t otus/rep_set.tre -e 692 
> 
> 
> Traceback (most recent call last):
>  File "/usr/lib/qiime/bin/beta_diversity_through_plots.py", line 166,
> in <module>
>    main()
>  File "/usr/lib/qiime/bin/beta_diversity_through_plots.py", line 163,
> in main
>    status_update_callback=status_update_callback)
>  File "/usr/lib/python2.7/dist-packages/qiime/workflow.py", line 867,
> in run_beta_diversity_through_plots
>    close_logger_on_success=close_logger_on_success)
>  File "/usr/lib/python2.7/dist-packages/qiime/workflow.py", line 135,
> in call_commands_serially
>    raise WorkflowError, msg
> qiime.workflow.WorkflowError: 
> 
> *** ERROR RAISED DURING STEP: Make 3D plots (continuous coloring,
> weighted_unifrac)
> Command run was:
> python /usr/lib/qiime/bin//make_3d_plots.py -p
> wf_bdiv_even692//prefs.txt -i wf_bdiv_even692//weighted_unifrac_pc.txt
> -o wf_bdiv_even692//weighted_unifrac_3d_continuous/ -m
> aran.r1.map.txt 
> Command returned exit status: 1
> Stdout:
> 
> Stderr
> Traceback (most recent call last):
>  File "/usr/lib/qiime/bin//make_3d_plots.py", line 443, in <module>
>    main()
>  File "/usr/lib/qiime/bin//make_3d_plots.py", line 419, in main
>    shutil.copyfile(os.path.join(jar_path,'king.jar'),
> os.path.join(jar_dir_path,'king.jar'))
>  File "/usr/lib/python2.7/shutil.py", line 82, in copyfile
>    with open(src, 'rb') as fsrc:
> IOError: [Errno 2] No such file or directory:
> '/usr/lib/python2.7/dist-packages/qiime/support_files/jar/king.jar'
> 
> 
> There is no "support_files" directory in that directory.  I have
> installed King myself independently in /usr/local/bioinf/king/king and
> this available as king from the command line.  I guess I could create
> a symlink to this, but it would break on update?  For the moment I'm
> going to get this user to load his distance matrices into R as they
> will have been generated and I'm evil ;)
> 
> 
> Cheers
> 
> 
> Mike
> 
> On 7 Nov 2012, at 10:27, Tim Booth <tbooth at ceh.ac.uk> wrote:
> 
> > Hi Mike,
> > 
> > Well, if I'd actually tested the thing I guess I'd have found this
> > out
> > for myself.  I just couldn't quite believe that Qiime would find the
> > JAR
> > that you've told it to run but then refuse to run it because it
> > doesn't
> > like the name.
> > 
> > The discussion on the forum makes it quite clear that they added
> > this
> > check because they wanted to stop people using the older version of
> > the
> > classifier.  It is highly likely that the new version works fine.
> >  Maybe
> > the 2.2 RDP version added some option that Qiime relies on, in which
> > case it's unlikely this would be removed in a later version.
> > 
> > My "normalised" comment means that the rdp-classifier package now
> > meets
> > proper packaging standards.  It was a bit of a fudge before, and I
> > couldn't put all the configuration bits in one place.  It also means
> > that the version of the package given by dpkg gives you the version
> > of
> > the classifier - ie. 2.5.
> > 
> > My inclination is to do a bit more testing and then if 2.5 seems to
> > work
> > I'll patch Qiime to accept it.  I'll also add a warning when Qiime
> > is
> > started that the classifier being used is not the default.  Let me
> > know
> > if you have any problems with the renamed 2.5 classifier or if you
> > think
> > you need to switch back to 2.2 in order to ensure a "standard"
> > analysis.
> > 
> > Cheers,
> > 
> > TIM
> > 
> > > I did say below!  :P
> > > 
> > > assign_taxonomy.py does a crappy name check
> > > https://groups.google.com/d/msg/qiime-forum/o7CaXnfS1J0/DgsFm-Q-s6YJ
> > > hence the problem
> > > 
> > > Apparently they only support 2.2 as well, do you know whether
> > > Biolinux
> > > has the most recent version of RDP classifier installed?  I tried
> > > to
> > > determine the version number, but helpfully it's not included in
> > > the
> > > help output :) 
> > > 
> > > I am fairly sure other versions aren't supported, and after
> > > changing
> > > the name to rdp_classifier-2.2.jar
> > > 
> > > and the path to
> > > 
> > > # This was previously set in bio-linux-rdp-classifier but can now
> > > live
> > > here
> > > # as the rdp-classifier package has been normalised.
> > > export RDP_JAR_PATH=
> > > ${RDP_JAR_PATH:-/usr/share/rdp-classifier/rdp_classifier-2.2.jar}
> > > 
> > > It does seem to run and produce a sensible result, but I've not
> > > checked within the context of the whole pipeline yet.  Will let
> > > you
> > > know if I hit a problem with it and have to install 2.2.  What
> > > does
> > > the normalised comment refer to?
> > > 
> > > Thanks!  QIIME is such fun isn't it?  Is Galaxy more or less fun?
> > > 
> > 
> > -- 
> > This message (and any attachments) is for the recipient only. NERC
> > is subject to the Freedom of Information Act 2000 and the contents
> > of this email and any reply you make may be disclosed by NERC unless
> > it is exempt from release under the Act. Any material supplied to
> > NERC may be stored in an electronic records management system.
> 

-- 
Tim Booth <tbooth at ceh.ac.uk>
NERC Environmental Bioinformatics Centre 

Centre for Ecology and Hydrology
Maclean Bldg, Benson Lane
Crowmarsh Gifford
Wallingford, England
OX10 8BB 

http://nebc.nerc.ac.uk
+44 1491 69 2705