Galaxy's dependency on old samtools vs tools wrapping later versions?
Hello all, Galaxy currently requires samtools on the $PATH in order to sort and index BAM files automatically, and samtools 0.1.19 works fine. Unfortunately later versions of samtools index have a regression: https://github.com/samtools/samtools/issues/199 This has caught several people out already, e.g. https://biostar.usegalaxy.org/p/7928/ and https://biostar.usegalaxy.org/p/9335/ While eventually samtools will be fixed, right now this means we can't have samtools 1.1 as the first samtools on the $PATH used by Galaxy. I am working on a wrapper for "samtools bam2fq": https://github.com/peterjc/pico_galaxy/tree/master/tools/samtools_bam2fq https://testtoolshed.g2.bx.psu.edu/view/peterjc/samtools_bam2fq The bam2qf command in samtools 0.1.19 has a number of bugs, so I want to target samtools 1.1. However this has complicated my testing since for my BAM input files Galaxy will call "samtools index", and if it calls samtools 1.1 this will fail. I'm not using the tool shed dependencies during development so instead came up with the following hack: https://github.com/peterjc/picobio/blob/master/sambam/samtools_auto.py My question is, what is expected to happen with a Tool Shed installed wrapper for samtools 1.1 and Galaxy's attempts to automatically call samtools to index any BAM output file? Would the tool environment put samtools 1.1 on the (local) $PATH which would then break setting the metadata as part of the same job? Regards, Peter
Peter, For the automated indexing of bam files, Galaxy uses the samtools version linked to as "default" under tool-dependencies/samtools/ This should normally be 0.1.19 or older, due to the not-yet-implemented handling of bam_index_build and other potential regressions that could be uncovered in the future. --Dave B. On 11/03/2014 01:52 PM, Peter Cock wrote:
Hello all,
Galaxy currently requires samtools on the $PATH in order to sort and index BAM files automatically, and samtools 0.1.19 works fine.
Unfortunately later versions of samtools index have a regression: https://github.com/samtools/samtools/issues/199
This has caught several people out already, e.g. https://biostar.usegalaxy.org/p/7928/ and https://biostar.usegalaxy.org/p/9335/
While eventually samtools will be fixed, right now this means we can't have samtools 1.1 as the first samtools on the $PATH used by Galaxy.
I am working on a wrapper for "samtools bam2fq": https://github.com/peterjc/pico_galaxy/tree/master/tools/samtools_bam2fq https://testtoolshed.g2.bx.psu.edu/view/peterjc/samtools_bam2fq
The bam2qf command in samtools 0.1.19 has a number of bugs, so I want to target samtools 1.1. However this has complicated my testing since for my BAM input files Galaxy will call "samtools index", and if it calls samtools 1.1 this will fail.
I'm not using the tool shed dependencies during development so instead came up with the following hack: https://github.com/peterjc/picobio/blob/master/sambam/samtools_auto.py
My question is, what is expected to happen with a Tool Shed installed wrapper for samtools 1.1 and Galaxy's attempts to automatically call samtools to index any BAM output file? Would the tool environment put samtools 1.1 on the (local) $PATH which would then break setting the metadata as part of the same job?
Regards,
Peter ___________________________________________________________ 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/
OK, so this should work then... :) Thanks Dave, Peter On Mon, Nov 3, 2014 at 7:06 PM, Dave Bouvier <dave@bx.psu.edu> wrote:
Peter,
For the automated indexing of bam files, Galaxy uses the samtools version linked to as "default" under tool-dependencies/samtools/
This should normally be 0.1.19 or older, due to the not-yet-implemented handling of bam_index_build and other potential regressions that could be uncovered in the future.
--Dave B.
On 11/03/2014 01:52 PM, Peter Cock wrote:
Hello all,
Galaxy currently requires samtools on the $PATH in order to sort and index BAM files automatically, and samtools 0.1.19 works fine.
Unfortunately later versions of samtools index have a regression: https://github.com/samtools/samtools/issues/199
This has caught several people out already, e.g. https://biostar.usegalaxy.org/p/7928/ and https://biostar.usegalaxy.org/p/9335/
While eventually samtools will be fixed, right now this means we can't have samtools 1.1 as the first samtools on the $PATH used by Galaxy.
I am working on a wrapper for "samtools bam2fq": https://github.com/peterjc/pico_galaxy/tree/master/tools/samtools_bam2fq https://testtoolshed.g2.bx.psu.edu/view/peterjc/samtools_bam2fq
The bam2qf command in samtools 0.1.19 has a number of bugs, so I want to target samtools 1.1. However this has complicated my testing since for my BAM input files Galaxy will call "samtools index", and if it calls samtools 1.1 this will fail.
I'm not using the tool shed dependencies during development so instead came up with the following hack: https://github.com/peterjc/picobio/blob/master/sambam/samtools_auto.py
My question is, what is expected to happen with a Tool Shed installed wrapper for samtools 1.1 and Galaxy's attempts to automatically call samtools to index any BAM output file? Would the tool environment put samtools 1.1 on the (local) $PATH which would then break setting the metadata as part of the same job?
Regards,
Peter ___________________________________________________________ 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/
Fingers crossed - perhaps I jumped the gun uploading this to the main tool shed without seeing the test results on the Test Tool Shed: https://toolshed.g2.bx.psu.edu/view/peterjc/samtools_bam2fq https://testtoolshed.g2.bx.psu.edu/view/peterjc/samtools_bam2fq I look forward to the automated test results from the Tool Sheds ... http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-November/020792.html Thanks, Peter On Tue, Nov 4, 2014 at 8:38 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
OK, so this should work then... :)
Thanks Dave,
Peter
On Mon, Nov 3, 2014 at 7:06 PM, Dave Bouvier <dave@bx.psu.edu> wrote:
Peter,
For the automated indexing of bam files, Galaxy uses the samtools version linked to as "default" under tool-dependencies/samtools/
This should normally be 0.1.19 or older, due to the not-yet-implemented handling of bam_index_build and other potential regressions that could be uncovered in the future.
--Dave B.
On 11/03/2014 01:52 PM, Peter Cock wrote:
Hello all,
Galaxy currently requires samtools on the $PATH in order to sort and index BAM files automatically, and samtools 0.1.19 works fine.
Unfortunately later versions of samtools index have a regression: https://github.com/samtools/samtools/issues/199
This has caught several people out already, e.g. https://biostar.usegalaxy.org/p/7928/ and https://biostar.usegalaxy.org/p/9335/
While eventually samtools will be fixed, right now this means we can't have samtools 1.1 as the first samtools on the $PATH used by Galaxy.
I am working on a wrapper for "samtools bam2fq": https://github.com/peterjc/pico_galaxy/tree/master/tools/samtools_bam2fq https://testtoolshed.g2.bx.psu.edu/view/peterjc/samtools_bam2fq
The bam2qf command in samtools 0.1.19 has a number of bugs, so I want to target samtools 1.1. However this has complicated my testing since for my BAM input files Galaxy will call "samtools index", and if it calls samtools 1.1 this will fail.
I'm not using the tool shed dependencies during development so instead came up with the following hack: https://github.com/peterjc/picobio/blob/master/sambam/samtools_auto.py
My question is, what is expected to happen with a Tool Shed installed wrapper for samtools 1.1 and Galaxy's attempts to automatically call samtools to index any BAM output file? Would the tool environment put samtools 1.1 on the (local) $PATH which would then break setting the metadata as part of the same job?
Regards,
Peter
participants (2)
-
Dave Bouvier -
Peter Cock