Hello again Peter, More great suggestions that we will consider for later enhancements. Thanks for the feedback! Best, Jen Galaxy team On 9/9/10 6:38 AM, Peter wrote:

Hello,

Are there any existing Galaxy wrappers for the command line tools (or the online services) TMHMM v2.0 (Trans-Membrane Hidden Markov Model) and SignalP v3.0 (Signal Peptide)?

http://www.cbs.dtu.dk/cgi-bin/nph-sw_request?tmhmm http://www.cbs.dtu.dk/cgi-bin/nph-sw_request?signalp

Both take protein FASTA files for input, but there are a few extra considerations here like limiting each call to TMHMM to 4000 proteins (IIRC), the possibility to split the job in order to take advantage of multiple cores, and parsing the output into a Galaxy friendly tabular form.

Thanks,

Peter _______________________________________________ galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev

-- Jennifer Jackson http://usegalaxy.org

On Fri, Oct 1, 2010 at 10:43 PM, Peter <peter@maubp.freeserve.co.uk> wrote:

Hi Jen,

I was hoping for someone to say "yes, we did that for our Galaxy, here is the wrapper code", but I'll settle for encouragement. ;)

As you will have seen, I'm focusing on the NCBI BLAST+ wrappers at the moment, but wrapping the command line versions of TMHMM v2.0 and SignalP v3.0 would probably be next on my list of things to do in Galaxy.

Peter

Hi all, The TMHMM v2.0 and SignalP v3.0 tools are free to download and use for academics, but commercial or other users would need to negiotiate a license: http://www.cbs.dtu.dk/cgi-bin/nph-sw_request?tmhmm http://www.cbs.dtu.dk/cgi-bin/nph-sw_request?signalp There is nothing to stop me writing and sharing wrappers for the tools, and installing and using TMHMM v2.0 and SignalP v3.0 on a local Galaxy server for academic use. However, I guess their licence would prevent TMHMM v2.0 and SignalP v3.0 being included in the "public" Galaxy server? One option might be to ask CBS if they would allow Penn State University to make these tools available via their public Galaxy server. Has this issue of "academic only" licences come up before for Galaxy wrapped tools? Alternatively I might be able to write wrappers for the CBS hosted TMHMM and SignalP webservice which appears to be free to use by anyone, but that seems less robust and they may be unhappy with the likely increased usage. Peter

On Fri, Oct 22, 2010 at 9:12 PM, Jennifer Jackson <jen@bx.psu.edu> wrote:

Hi Peter,

Good question about licenses! Sorry for the delayed answer, we had a bit of discussion to review the existing strategy.

The general guiding policy is that if a tool is not completely public, then it will not be released to the main code base/servers.

As I expected - thanks for the clarification.

That said, you are correct - if something can be worked out with tool author, coordination for a fully integrated wrapper is possible. Same would be true for remote use of a 3rd party web tool - coordination with the authors/owner would be needed first (although this is not the ideal solution for galaxy main).

OK

Meanwhile, a great place to add wrappers/remote access tools that have special licensing is the tool shed (http://community.g2.bx.psu.edu/). The user comments field (displayed as Description) is free-form text and any credits/license pointers can be added there for now. We may add in a special field for this info later on, but for immediate use this is a visible key that will likely satisfy many licensing rules.

Good idea. Can you (or someone at Galaxy) add another category? Something like "Sequence analysis", or "Protein sequence analysis" seems appropriate to me. Thanks, Peter

On Mon, Nov 8, 2010 at 9:49 AM, Peter <peter@maubp.freeserve.co.uk> wrote:

Thank you to Jennifer for adding that section.

I submitted a "tool suite" containing my TMHMM 2.0 and SignalP 3.0 wrappers last week - currently in status "waiting" so not publicly visible yet.

And it is now online at http://community.g2.bx.psu.edu/ Thanks to whomever from Galaxy did the review, I guess there were no comments which I hope is a good sign. Peter