Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
Martin Cech
@martenson
are you familiar with any of these two?
Dannon
@dannon
That'd get you make, and if this was working previously, it should continue to.
Dong-Hun Lee
@aivndv_twitter
Not really... hmm
Martin Cech
@martenson
This is your first time trying Galaxy, right?
Dong-Hun Lee
@aivndv_twitter
Yes
Dannon
@dannon
Yeah, I'd do what @martenson suggested and try docker.
Docker on windows should work well, and there's a ready-to-go image for you with Galaxy preinstalled.
Dong-Hun Lee
@aivndv_twitter
okay i will try
thanks
Dannon
@dannon
https://github.com/bgruening/docker-galaxy-stable This is a link to the repository with some more information in the README file, about how it runs, etc.
Definitely ping us here, though, if you run into more trouble.
Martin Cech
@martenson
You can also check out https://www.lifewire.com/run-ubuntu-within-windows-virtualbox-2202098 for how to run an ubuntu VM, Galaxy can be installed there easily.
if you want to just try Galaxy, try the docker solution. If you want to develop Galaxy tools and get more familiar with Linux - try the VM.
There is also a virtualbox image with ubuntu and Galaxy preinstalled at https://images.galaxyproject.org/planemo/latest.ova
Good luck.
Dong-Hun Lee
@aivndv_twitter
Thanks much!
galaxybot
@galaxybot
[gronke] I am getting this error in the galaxy.log when I try to send a file to Jbrowse. The error in Galaxy is "Exit code 1" https://pastecode.xyz/view/8a47f3f0
[gronke] Any idea what is causing it?
[gronke] Actually hold on, I have more information, here is the error from jbrowse right before, from the log
galaxybot
@galaxybot
[gronke] I'm running a test now but I am quite sure our jbrowse runs fine when I have connected other tools to it, I am surmising that a certain tool might be corrupting the files before it reaches jborws
MoHeydarian
@MoHeydarian

Hi Everyone!

I would like to remind everyone that the 2019 Galaxy Admin Training will be offered January 28 through February 1 at Penn State University (where the Galaxy Project got started over ten years ago). The workshop offers a 2 day introductory session followed by a 3 day advanced topics session. Participants can register for one or both sessions.

Registration is open and starts at $40 / day for participants from non-profits and academia. Early registration ends TODAY October 31, register now before rates increase.

The 2016 and 2018 admin trainings were both full, so you are strongly encouraged to register now.

galaxybot
@galaxybot
[gronke] yep, jbrowse runs fine if I do cutadapt ---> bwa --> jbrowse, but running varscan and lofreq apparently cause errors with it
Jennifer Hillman-Jackson
@jennaj
@martenson @mvdbeek Interesting question about BAM indexes. Any ideas for a solution? In main distro or tuning their local: https://biostar.usegalaxy.org/p/29884/
Ghost
@ghost~5772e7e2c2f0db084a206e1b
Yes, we can generate .csi indexes, but I don't think any galaxy tool would be able to do anything with those at the moment.
I assume the issue is mostly that metadata setting doesn't work for that genome ?
Jennifer Hillman-Jackson
@jennaj
Yes, that is how I understand it. They may have custom tools. I pointed the user to this thread, and will prompt them for more info/intended usage.
SergejN
@SergejN
hi all, thanks for quick response. What I do in most of the custom/adjusted tools is to load the SAM header and check if there is any coordinate beyond 2^30. If there is, then I create the index using "-c -m 14". Many users use Galaxy to map their *seq data to the genome and upload the data in the genome browser to viewing. However, since Galaxy tries to create the bai index automatically, the tools fail and I had to define a custom data format "unsorted.bam" to prevent the default behavior
then it gets really messy, since I can create tools that sort the unsorted bam files, but the type should again be "unsorted.bam", which is misleading and won't allow me to discriminate between the sorted and unsorted bams any more.
SergejN
@SergejN
I just checked- the UCSC genome browser accepts bai/csi
galaxybot
@galaxybot
[gronke] Still trying to pin down what's messing up my files before they reach jbrowse
[gronke] The three ones that fail have this in common: MarkDuplicatesWithMateCigar and AddOrReplaceReadGroups. Both of those run fine, but the jbrowse afterwards fail
Ghost
@ghost~5772e7e2c2f0db084a206e1b
Thanks for the report @SergejN, i've opened galaxyproject/galaxy#6956 for this. Do you know if .crai index for cram files have the same limitation ?
SergejN
@SergejN
@mvdbeek thanks! Unfortunately, I don't know. I've never used CRAM files, sorry
Ghost
@ghost~5772e7e2c2f0db084a206e1b
Depending on what type of sequencing data you have that might be another option. Cram files use a different compression scheme to bam files and are interconvertible for short read sequencing (you might loose some information on very long reads if converting from cram to bam). Galaxy supports these as well, and you can switch minimap2 for instance to produce cram files
But it isn't clear from the samtools manual whether they'd have the same limitation
SergejN
@SergejN
I am not sure if all tools for Hi-C, ATACseq, ChIP-seq peak calling etc would support CRAM. And as soon as I convert CRAM to BAM, I would have the same issue with the bai file again
Robert Leach
@hepcat72
Is there a way to (or a way planned to) add help text to workflow parameters that shows up with the parameter entry boxes at the top of a "run workflow" form? I try to make my variables verbose, but I can't exactly put an equation in there for the user to know how to calculate the value that should be entered...
Dannon
@dannon
@hepcat72 No, but that's a cool idea for an enhancement if you want to add a ticket
galaxybot
@galaxybot
[jebediah_] If i wanted a galaxy tool to output a yml file, would I have to add it as a file type in the config file?
jhl667
@jhl667
Hello! We are attempting to improve the performance of our compute cluster, and are exploring ways to offload stuff that usually occurs in the job_working_dir (on the shared portion of the filesystem) to compute node local scratch space. Does anyone have suggestions regarding Galaxy configuration that can help us achieve this?
Khadidja
@khadidja01_gitlab
Good afternoon,
I wanted to perform the cgSNVanalysis for 925 genomes, but my data can't be analysed in one history. So, I put them in 4 separated histories and I created a list of datasets pairs for each of them. In this case, SNVphyl works for each list separately. However, when I merged all these lists using Merge Collections into single list of datasets tool (Galaxy Version 1.0.0), SNVphyl doesn't work on this collection. Could you, please, tell me how I can analyse more than 250 isolates using SNVphyl in Galaxy? Because when we have more than 250 genomes they are usually distributed on the separate pages of the same history and we can't create a list of datasets pairs.
Thanks,
Best regards,
Slugger70
@Slugger70
@jhl667 I'm very interested in using worker node local scratch also! Not sure how to get it done though...
galaxybot
@galaxybot
[gronke] Alright, I think I've narrowed down the problem. JBrowse seems to be crashing when I try to send VCF SNPs to it. I have bcftools, varscan, and lofreq all sending them, and all three are crashing.
galaxybot
@galaxybot
[gronke] I have a workflow that sends bams to jbrowse as well as VCF snps, the bam one finishes but the jbrowse one crashes.
[gronke] or rather the bam jbrowse finishes, the vcf snps jbrowse crashes
jhl667
@jhl667
@Slugger70 Only thought I had so far was to switch job_working_directory parameter in the config file over to some location on the local scratch space. I will try this out as soon as a few jobs finish up.
galaxybot
@galaxybot
[gronke] Alright, I'm at my wits end here. I have tried every possible which way for the past week. I CANNOT get Jbrowse to take VCF SNPs as input without crashing. I get an Exit code 1 () fatal error.
[gronke] I am hoping someone out there can help me figure out what the heck is wrong here.