to put or not to put a tool suite into the toolshed

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

to put or not to put a tool suite into the toolshed

Wolfgang Maier
Dear all,
I am maintaining a (still young) suite of command line tools (written in
Python) for identification of mutations in model organism genomes
through whole-genome sequencing (https://sourceforge.net/projects/mimodd/).
MiModD aims at geneticists that do not have much background in
bioinformatics, and it's supposed to make WGS analysis for small model
organism genomes (anything from yeast to fish) possible on regular PCs,
so it's not a cloud/cluster solution.

We do support Galaxy though through a complete set of tool wrappers for
three reasons:
- to provide a graphical user interface
- to offer labs the possibility to install the software on one dedicated
machine, but run analyses from any machine (typically with Windows
installed)
- to keep analysis workflows documented and reproducible.

Currently, we advise users to take advantage of these features and
install a local instance of Galaxy even though it will be running then
only on a single desktop PC (or even just a notebook). After
installation of Galaxy our software simply copies its wrapper xmls over
to the tools folder and modifies the tool_conf file to integrate itself.
In addition, users have to install a bit of other third-party software
(samtools, snap aligner, optionally snpeff) that our code relies on.

[End of lengthy introduction]

So my question is: in what way could the package profit from being
uploaded to a Galaxy toolshed ? I guess it would mean quite some extra
work from my side since I'm not familiar with the whole procedure, so
are there benefits (visibility, ease of installation, etc.) that are
worth the effort ?

Thanks a real lot for any feedback,
Wolfgang

___________________________________________________________
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/
Reply | Threaded
Open this post in threaded view
|

Re: to put or not to put a tool suite into the toolshed

Hans-Rudolf Hotz
Hi Wolfgang


short answer:

Please, put your package into the Galaxy Tool Shed. This makes it
visible to the whole Galaxy community and everybody can profit from it.


long answer:

I have been in the same boat. We never put the Galaxy wrappers and the
perl scripts of our old NGS pipeline into the Tool Shed. I do feel bad
about this. My excuse has always been that too much local customization
would be necessary, eg making possible to store data outside the Galaxy
directory tree. Hence, we only distributed the package as bunch of files
on request. Obviously, there were hardly any requests.

Now we are in a similar situation. We are in the process of making our
new R/Bioconductor based NGS pipeline available on our internal Galaxy
server. And the code is far from being ready to be put into the Tool
Shed. But this time we will make the effort. So the Galaxy community can
benefit from it (@Dave Clements: Unfortunately, it won't be ready for
Baltimore).



Regards, Hans-Rudolf




On 06/05/2014 11:52 PM, Wolfgang Maier wrote:

> Dear all,
> I am maintaining a (still young) suite of command line tools (written in
> Python) for identification of mutations in model organism genomes
> through whole-genome sequencing (https://sourceforge.net/projects/mimodd/).
> MiModD aims at geneticists that do not have much background in
> bioinformatics, and it's supposed to make WGS analysis for small model
> organism genomes (anything from yeast to fish) possible on regular PCs,
> so it's not a cloud/cluster solution.
>
> We do support Galaxy though through a complete set of tool wrappers for
> three reasons:
> - to provide a graphical user interface
> - to offer labs the possibility to install the software on one dedicated
> machine, but run analyses from any machine (typically with Windows
> installed)
> - to keep analysis workflows documented and reproducible.
>
> Currently, we advise users to take advantage of these features and
> install a local instance of Galaxy even though it will be running then
> only on a single desktop PC (or even just a notebook). After
> installation of Galaxy our software simply copies its wrapper xmls over
> to the tools folder and modifies the tool_conf file to integrate itself.
> In addition, users have to install a bit of other third-party software
> (samtools, snap aligner, optionally snpeff) that our code relies on.
>
> [End of lengthy introduction]
>
> So my question is: in what way could the package profit from being
> uploaded to a Galaxy toolshed ? I guess it would mean quite some extra
> work from my side since I'm not familiar with the whole procedure, so
> are there benefits (visibility, ease of installation, etc.) that are
> worth the effort ?
>
> Thanks a real lot for any feedback,
> Wolfgang
>
> ___________________________________________________________
> 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/
___________________________________________________________
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/
Reply | Threaded
Open this post in threaded view
|

Re: to put or not to put a tool suite into the toolshed

Björn Grüning-3
In reply to this post by Wolfgang Maier
Hi Wolfgang,

> I am maintaining a (still young) suite of command line tools (written in
> Python) for identification of mutations in model organism genomes
> through whole-genome sequencing (https://sourceforge.net/projects/mimodd/).
> MiModD aims at geneticists that do not have much background in
> bioinformatics, and it's supposed to make WGS analysis for small model
> organism genomes (anything from yeast to fish) possible on regular PCs,
> so it's not a cloud/cluster solution.
>
> We do support Galaxy though through a complete set of tool wrappers for
> three reasons:
> - to provide a graphical user interface
> - to offer labs the possibility to install the software on one dedicated
> machine, but run analyses from any machine (typically with Windows
> installed)
> - to keep analysis workflows documented and reproducible.
>
> Currently, we advise users to take advantage of these features and
> install a local instance of Galaxy even though it will be running then
> only on a single desktop PC (or even just a notebook). After
> installation of Galaxy our software simply copies its wrapper xmls over
> to the tools folder and modifies the tool_conf file to integrate itself.
> In addition, users have to install a bit of other third-party software
> (samtools, snap aligner, optionally snpeff) that our code relies on.
>
> [End of lengthy introduction]
>
> So my question is: in what way could the package profit from being
> uploaded to a Galaxy toolshed ? I guess it would mean quite some extra
> work from my side since I'm not familiar with the whole procedure, so
> are there benefits (visibility, ease of installation, etc.) that are
> worth the effort ?

Yes, it would mean extra work for you, but it is worth the effort.

One main reason, you already mentioned, visibility. If done right and
you are supporting your wrapper these will probably be used by many
different Galaxy Servers out there. Just have a look at a few tools that
are from the same field as your tools and look at the download counter.

Easy installation is also one reason, as soon as you have a Galaxy
server it can be trivial to install your tools, suppose you define all
dependencies. -> But defining dependencies is easy nowadays, and even
better, many dependencies are already available in the Tool Shed. If you
depend on numpy, insert a few lines to point to the correct numpy
dependency and you are done.
We can help you if you encounter any problems in that step, or just drop
by during the GCC developer workshop. We will give a hands-on how to
wrap your tools.

Reproducibility: With the Tool Shed you get reproducibility for free.
You can install many different version of your tools and Galaxy will
take care to select the correct one to reproduce your results.

You can also share your workflows in the Tool Shed. Convince others from
your tools, export your workflows and and anyone can test it with a few
clicks.

Be part of a very nice community, which tries hard to make bioinformatic
(and more) tools accessible for everyone.

Convinced? :)
Cheers,
Bjoern


> Thanks a real lot for any feedback,
> Wolfgang
>
> ___________________________________________________________
> 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/
___________________________________________________________
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/
Reply | Threaded
Open this post in threaded view
|

Re: to put or not to put a tool suite into the toolshed

Wolfgang Maier
In reply to this post by Wolfgang Maier
Thanks a lot for all the encouraging feedback.

More than anything else the friendly atmosphere that seems to prevail on
this mailing list makes me think that it may indeed be worthwhile to
make the package available to this community !
I will have a closer look at the toolshed instructions and see if I get
along with them.

Have a nice weekend,
Wolfgang

___________________________________________________________
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/