November 2012 - galaxy-dev - lists.galaxyproject.org

Folder permissions after cluster run
by Sajoscha Sauer 26 Nov '12

26 Nov '12

Hi, We are running galaxy on the cluster with the 'real user' submission. It all works fine, but at the end of the run we found something unexpected. All the .dat files in the file_path folder are owned by the user running Galaxy, but all the folders (e.g. dataset_25808_files) are still owned by the user that was running the tool on the cluster. Is this the expected outcome or is there something wrong? Thanks for your help, Sajoscha

2 1

Can't edit galaxy workflow
by Sanjarbek Hudaiberdiev 26 Nov '12

26 Nov '12

I tried to reply to similar posts, but couldn't figure out how to do it. So posting again: Galaxy giving this error when editing workflow, just after creating workflow from existing history: URL:http://localhost/galaxy/workflow/load_workflow?id=df7a1f0c02a5b08e&_=1352817462148 Moduleweberror.evalexception.middleware:*364*in|respond| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>app_iter*=*self*.*application*(*environ*,*detect_start_response*)*| Modulepaste.debug.prints:*98*in|__call__| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>environ*,*self*.*app*)*| Modulepaste.wsgilib:*539*in|intercept_output| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>app_iter*=*application*(*environ*,*replacement_start_response*)*| Modulepaste.recursive:*80*in|__call__| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>*return*self*.*application*(*environ*,*start_response*)*| Modulepaste.httpexceptions:*632*in|__call__| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>*return*self*.*application*(*environ*,*start_response*)*| Modulegalaxy.web.framework.base:*160*in|__call__| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>body*=*method*(*trans*,*****kwargs*)*| Modulegalaxy.web.framework:*73*in|decorator| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>*return*simplejson*.*dumps*(*func*(*self*,*trans*,****args*,*****kwargs*)**)*| Modulegalaxy.webapps.galaxy.controllers.workflow:*733*in|load_workflow| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>'tooltip'*:*module*.*get_tooltip*(*static_path*=*url_for*(*'/static'*)**)**,*| Modulegalaxy.workflow.modules:*258*in|get_tooltip| |>> <http://localhost/galaxy/workflow/editor?id=df7a1f0c02a5b08e#>*return*self*.*tool*.*help*.*render*(*static_path*=*static_path*)*| *AttributeError: 'NoneType' object has no attribute 'render'* Could anyone help me to solve this problem? Thanks, Sanjar.

5 7

WebDAV
by Adhemar 26 Nov '12

26 Nov '12

Hi folks, I've just set a WebDAV directory in order to let users upload files as an alternative to the good old FTP. Since it might be useful, here are the steps. I hope you enjoy! Adhemar *Configuring WebDAV* In order to provide a drag and drop interface to upload files to galaxy, one can choose WebDAV. To make it available, you'll need to install the following: # These modules are responsible for enabling the DAV protocol through the apache webserver mod_dav mod_dav_fs # This package is necessary in order to set galaxy uid/gid (AssignUserId) rather than the default apache user httpd-itk.x86_64 : MPM Itk for Apache HTTP Server # First, install httpd-itk (http://www.webtatic.com/packages/httpd-itk/) yum install httpd-itk # Then edit /etc/sysconfig/httpd and add the following line: HTTPD=/usr/sbin/httpd.itk # The galaxy user must have write permission in the directory containing the DavLockDB files otherwise: [error] [client 127.0.0.1] The locks could not be queried for verification against a possible "If:" header. [500, #0] [error] [client 127.0.0.1] Could not open the lock database. [500, #400] [error] [client 127.0.0.1] (13)Permission denied: Could not open property database. [500, #1] # Here's an example of apache's virtual host configuration, assuming ftp_upload_dir = /home/galaxy/database/ftp in the Galaxy config file: # include 'REMOTE_USER' in the directory path in order to limit users to their own directories RewriteRule ^/repository(.*) /home/galaxy/database/ftp/%{LA-U:REMOTE_USER}$1 <Directory /var/webdav> Options Indexes FollowSymlinks MultiViews AllowOverride None Order Allow,Deny Allow from all # set the galaxy user and group for your vhosts to use AssignUserId galaxy galaxy # here follows the PAM authentication in order to allow users registered in the server itself # can be changed to ldap, htpasswd file, ... AuthType Basic AuthName Repository AuthPAM_Enabled on Require valid-user # activate mod_dav Dav On </Directory> # It's necessary to create each user directory with the correct galaxy user permissions!!! # After that you can use a webdav client (cadaver) or map the network directory in a filemanager (nautilus, konqueror, orca) to view/edit/transfer files. cadaver http://hostname/repository

1 0

XGRID and Galaxy
by James Boocock 26 Nov '12

26 Nov '12

Hi Galaxy Dev Team. We are looking to put xgrid into the galaxy api as part of a summer project. We are hoping to add this functionality directly into galaxy so that all users can benefit. We are curious where to start investigating how galaxy interfaces with grids. We hope to integrate the functionality into galaxy so that a user can specify xgrid in the config file and galaxy will deal with the rest of the interfacing with xgrid. This would of course only be avaliable to MACOSX galaxy installs. Cheers James. Research Assistant Department of Biochemistry University Of Otago

3 2

FW: Re. RNA-Seq
by Songyan Liu 23 Nov '12

23 Nov '12

My account emailis soliu(a)cc.umanitoba.ca From: Songyan Liu [mailto:mfgc@cc.umanitoba.ca] Sent: Friday, November 23, 2012 1:41 PM To: 'galaxy-dev(a)lists.bx.psu.edu' Subject: Re. RNA-Seq I was using Cufflinks to analyze Clip-Seq data. The results as 'gene expression' and 'transcript expression' gave me almost same results (No. of genes ? No. of transcripts). I guess the number of genes should be relatively smaller than that in transcripts. I wonder if I need to get a special format for the annotation? I used '62: UCSC Main on Human: refGene (genome)'. Thanks, Songyan Liu

1 0

Re. RNA-Seq
by Songyan Liu 23 Nov '12

23 Nov '12

I was using Cufflinks to analyze Clip-Seq data. The results as 'gene expression' and 'transcript expression' gave me almost same results (No. of genes ? No. of transcripts). I guess the number of genes should be relatively smaller than that in transcripts. I wonder if I need to get a special format for the annotation? I used '62: UCSC Main on Human: refGene (genome)'. Thanks, Songyan Liu

1 0

help
by Magriet Van Der Nest 23 Nov '12

23 Nov '12

Dear Galaxy help-centre I want to use Tophat and Clifflinks to do an expression analysis. However, the reference genome for Heterobasidion annosum (it is available at JGI) is not listed. You have posted on your website that we can contact you if 'your genome of interest' is not listed. Is it possible that you can included the H. annosum genome on the list? Thank you very much Magriet

1 0

piping two scripts in the same xml tool wrapper
by Christophe Antoniewski 22 Nov '12

22 Nov '12

Hi, Is there a way to pipe in the same xml wrapper two scripts in different languages, the output of the first script becoming the input of the second ? I know that the spirit of galaxy would be to make a workflow, but the intermediate output in the present case is not interesting at all and may confuse the user. I am testing something like : <command interpreter="python">plotter.py $input $minsize $maxsize $factor $output</command> <command interpreter="bash">r_wrapper.sh $Rplotter</command> <inputs> <param name="input" type="data" format="tabular" label="Compute Plot table from this bowtie standard output"/> <param name="minsize" type="integer" size="3" value="20" label="Min size of small RNA to plot" help="'20' = 20 nucleotides"/> <param name="maxsize" type="integer" size="3" value="22" label="Max size of small RNA to plot" help="'22' = 22 nucleotides"/> <param name="factor" type="float" size="6" value="1.00" label="Normalization factort" help="leave at 1.00 for no normalization"/> <param name="title" type="text" size="15" value="Main Title" label="Main Titles"/> <param name="xlabel" type="text" size="15" value="Coordinates (nt)" label="x axis label"/> <param name="ylabel" type="text" size="15" value="Normalized number of reads" label="y axis label"/> <param name="yrange" type="integer" size="6" value="400" label="y axis range"/> </inputs> <configfiles> <configfile name="Rplotter"> ## Setup R error handling to go to stderr options( show.error.messages=F, error = function () { cat( geterrmessage(), file=stderr() ); q( "no", 1, F ) } ) tab = read.delim("${output}", header=TRUE) ## Open output PDF file pdf( "${outputFinal}" ) #determining y and x axis ranges MAXcoord = max(tab[,1]) + 100 MINcoord = min(tab[,1]) MMAX = ${yrange} plot(tab[,1], tab[,2], type="h", xaxt="n", xlim=c(MINcoord , MAXcoord), ylim=c(-MMAX , MMAX), frame.plot=FALSE, lwd=2, las=2, xlab= "${xlabel}", ylab="${ylabel}", main = "${title}") lines(tab[,1], tab[,3], type="h", lwd=2) axis(1) ## Close the PDF file devname = dev.off() </configfile> </configfiles> <outputs> <data name="output" format="tabular" label= "Data Frame"/> <data name="outputFinal" format="pdf" label= "PDF plot"/> </outputs> but the second script output is empty. I suspect that the second script is launched when the output of the first script is not available yet. Thanks for the help Chris

4 3

Bug when using data_column
by aristos aristodemou 22 Nov '12

22 Nov '12

Dear galaxy developers and users, When using the data_column type for a tool in galaxy, there is a bug that gives the message "An invalid option was selected, please verify". The problem occurs when for example you had selected a dataset with 6 columns in it and selected as a choice the 6th column and then changed your mind and selected another dataset with 5 columns. This happens because when it checks in basic.py if the value selected is in the accepted values of the selection list, the value it uses is the one that was selected on the previous dataset. Does anyone know how this issue could be solved so that the check is not performed when the dataset is selected or if there is a way to reset the value when a new dataset is selected? Regards, Aris

1 0

History will not refresh after workflow submission via the API
by MSc. Sascha Kastens 22 Nov '12

22 Nov '12

Hi devs! After updating my local instance to changeset 19cbbaf566216cb46ecc6a6d17e0f1e0ab52978e my tool which submits a workflow via the API and set force_history_refresh to TRUE does not refresh the history anymore. The job is submit via DRMAA to SGE. Unfortunately the job seems to run forever because it is still in yellow state (even after SGE tells Galaxy that the job has finished. As soon as I refresh the history manually the according history entries get green and all entries for the submitted workflow are shown. I have no idea where I can have a look at. If you need any further information let me know! Thanks for your help in advance! Cheers, Sascha

2 2