Hi Rui, This error probably means that the binaries from UCSC are not installed or the path to the tools is incorrect. "ucsc_tools" is a package requirement for the tools that use these binaries in the .xml portion of the wrapper (for those in the dist/devteam). If you are running your own instance, without any dependencies on other institutional resources or repositories, then any binaries needed should in most cases come from UCSC in the Downloads area under source and be installed. Some wrappers are in the distribution (such as wigToBigWig in galaxy-dist / tools / filters ), but there are others are in the tool shed. http://hgdownload.cse.ucsc.edu/admin/exe http://usegalaxy.org/toolshed Where the other repo comes in, I am not sure. Maybe from a wrapper installed from the Tool Shed (or another source) that needs a UCSC binary. If it is dependent on a particular version, that could be a reason for a distinct repo. Contacting the wrapper author may help to clarify. Tools in the distribution and from "devteam" in the Tool Shed all use the standard UCSC source to my knowledge. Maybe others will add in comments to this thread, if they recognize this. Or, your sysadmin can help if you contact them, if there shared resources to make use of. On 1/15/14 10:30 AM, ruiwang.sz wrote:
Hi Jennifer,
Thanks for the note.
A related question, do we have a 'ucsc_tool'? I sometimes saw the warning of failed dependency on ucsc_tool, but I'm not sure what it is. I found that there are many binary/script utilities:
https://github.com/adamlabadorf/ucsc_tools
should I install all these by running
python setup.py install
or I could just copy all those executables to my own utility dir?
Thanks, Rui
On Wed, Jan 15, 2014 at 7:51 AM, Jennifer Jackson <jen@bx.psu.edu <mailto:jen@bx.psu.edu>> wrote:
Hello Rui,
The problems you are describing have to do with the format of the input wig dataset.
It looks as if you have corrected the chromosome names to be identical to the reference genome build used (required). There are options to overcome the other issues:
1. verify that your data has no browser lines, and has track and definition lines in the correct format. UCSC is the definitive source for this info, as both the underlying tool and this format were developed by them. Links to their information and the general format rules in Galaxy can be found here https://wiki.galaxyproject.org/Learn/Datatypes#Wig_and_bigWig
2. send only a single wig file to the tool at a time, when using the Galaxy wrapper.
3. use the 'full parameter' option 'Clip chromosome positions:' to removing overhanging coordinates (known to be produced by several common tools). This is after confirming that the build is correct - overhanging coordinates can be a clue that there is a genome mismatch problem. https://wiki.galaxyproject.org/Support#Reference_genomes
4. note that variable step data comes in two formats fixed and variable - and that variable has two format versions, those with a a span definition and those without (see examples in wig examples in #1). I have only seen this tool run successfully, in Galaxy, on the type without "span" included. If you find that the span variable is problematic after correcting any other format issue issues, switch to the format without span.
Good luck,
Jen Galaxy team
On 1/14/14 9:52 PM, ruiwang.sz <http://ruiwang.sz> wrote:
Hi All,
I'm having an error at this:
*Dataset 18: Wig/BedGraph-to-bigWig on data 12*
Tool execution generated the following error message:
grep: writing output: Broken pipe
grep: writing output: Broken pipe
grep: writing output: Broken pipe
grep: writing output: Broken pipe
.. put: Broken pipe grep: writing output: Broken pipe grep: writing output: Broken pipe
..
grep: writing output: Broken pipe
grep: write error
Error running wigToBigWig.
The tool produced the following additional output:
hashMustFindVal: '1' not found
I searched and found this link:
http://redmine.soe.ucsc.edu/forum/index.php?t=msg&goto=10745&S=2a335135e76cf9b7160c0e9d41353767
which says that there is a naming convention difference.
I followed what he did and replaced chrom=1 to chrom=chr1 etc, now it goes further, but
still dies with error:
line 18020168 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276760
line 18020169 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276770
line 18020170 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276780
line 18020171 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276790
line 18020172 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276800
line 18020173 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276810
line 18020174 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276820
line 18020175 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276830
line 18020176 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276840
line 18020177 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276850
line 18020178 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276860
line 18020179 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276870
line 18020180 of /tmp/t3: chromosome chr1 has 195276750 bases, but item ends at 195276880
Unrecognized line 18020181 of /tmp/t3:
variableStep chr10 span=10
Error running wigToBigWig.
The command line is: wigToBigWig /tmp/t3 /home/bioinfoadmin/app/galaxy-dist/tool-data/shared/ucsc/chrom/ucsc_gg4.len /home/bioinfoadmin/app/galaxy-dist/database/files/000/dataset_769.dat -clip 2>&1 || echo "Error running wigToBigWig." >&2
Quite puzzled...wondering if anyone has seen it before and could give me a hand. :-) I'll really appreciate!
Thanks,
Rui
___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
-- Jennifer Hillman-Jackson http://galaxyproject.org
-- Jennifer Hillman-Jackson http://galaxyproject.org