Whenever I try to subtract datasets, I get (+) signs instead of the chromosome number in the BED format. Sometimes when I subtract datasets, the "chr" column is correctly displayed, however in some instances I just get a series of + signs. I am trying to subtract the simple repeats dataset from the mouse (mm10) genome from the dataset that shows the difference between the intergenic regions and the repeatmasker.
The tool Join, Subtract and Group > Subtract Whole Dataset from another dataset does not manipulate the data. Lines between datasets are directly compared. If a line contains unexpected data, it was present in the original inputs.
This assumes there is not a problem with the tool. At this time, none are known. Please check the outputs against the original input first. If you discover a problem (lines in the output that are NOT present in the inputs), please share the history, leave all datasets in the analysis undeleted - including any test datasets generated, and send that shared history link along with a link to this Biostars post to firstname.lastname@example.org.
How to share a history: https://wiki.galaxyproject.org/Learn/Share
Common troubleshooting & data investigation methods/tools: https://wiki.galaxyproject.org/Support#Troubleshooting_tool_errors_-_Review_THIS_before_submitting_bug_reports
Best, Jen, Galaxy team
Ah, just saw that you are using the tool Operate on Genomic Intervals > Subtract instead.
I am going to leave the above posted, but moved to a comment.
Please instead just share the history as described and we can review what the issue is. Please note that http://usegalaxy.org is undergoing some quick admin changes right now - so some functions like sharing may be difficult (result in a Galaxy could not be reached message). Wait a bit and it is expected to clear.
Here is the link to the history I am working on: https://usegalaxy.org/u/carlostorres/h/galaxy-101-2016 Best,
There are some items from the history that are correctly displayed but on some datasets like the Neutral Regions one, just has a series of + signs on the first column.
Thanks! Reviewing now. Jen
I see the problem - metadata assignments for BED datatypes is problematic. I am writing up a ticket to address this (and will share your history as a test example - if not OK - let me know) & trying to come up with a workaround for you to use now. Give me a bit to do all of this - is somewhat complicated.
Thank you. You can use the shared history as a test example.