Integrating Annocript into Galaxy

Question: Integrating Annocript into Galaxy

19 months ago by

Dear Galaxy community,

I would like to integrate my tool Annocript into Galaxy. Annocript was wrote in PERL language and needs a fasta file and a text configuration file in input. It runs BLAST software, DNA2PEP and Portrait. It creates a MySQL database that is used to integrate the annotations with UniProt.

Do you think it is possible to integrate it in Galaxy? Can you redirect me to a tutorial where a similar software has been integrated?

Thanks in advance

Francesco

perl mysql • 492 views

ADD COMMENT • link •

modified 19 months ago • written 19 months ago by francescomusacchia • 20

Dear Martin,

thanks for the quick reply.

The script takes in input a configuration text file where you have to write the paths of the programs it needs. You can both write a full path or a single program name. In the latter case of course it should be in the $PATH. So, to respond your first question: it isn't mandatory that DNA2PEP, BLAST and Portrait are in the $PATH. But they must be installed on the user machine.

About the 2nd question, I try to explain the functioning because it is not clear to me. When you run Annocript the first time it downloads UniProt databases, creates a MySQL database (command CREATE) and inserts proteins annotations into tables. This database is created only the first time. Further times you run my software it will fetch information from the database to construct the transcriptome annotation. So, I guess also using Galaxy the user will have to install MySQL (client and server) on her/his machine.

The database is both kind of data structure needed for Annocript to run and a final useful artifact for those user, with an SQL expertise, who want to query directly the annotation database.

Hope to have been clear.

ADD REPLY • link written 19 months ago by francescomusacchia • 20

In order to answer you I have two main questions:

It runs BLAST software, DNA2PEP and Portrait.

Does your perl script invoke these as binaries (expecting them on the $PATH)?

It creates a MySQL database that is used to integrate the annotations with UniProt.

Is there (going to be) a Galaxy Tool that takes this DB as an input (and in what format if so) or is this going to be the final artifact of the analysis in Galaxy? How can Galaxy user benefit from this?

ADD REPLY • link written 19 months ago by Martin Čech ♦♦ 4.9k

19 months ago by

Martin Čech ♦♦ 4.9k

United States

Martin Čech ♦♦ 4.9k wrote:

For some background on how Galaxy treats dependencies please see my answer here

Galaxy user has little control over the environment Galaxy runs in, that is a job of an admin to maintain, thus user cannot give you the configuration file with paths or even install the needed dependencies. With MySQL as a dependency you are getting even deeper into sysadmin area as not every system user can create a database even when MySQL has been installed - permissions need to be set up front.

Based on your current workflow you described I do not see a good way to integrate your tool into Galaxy. If you want to try I would consider using sqlite instead of MySQL (cummerbund tool might be an example to explore how this can be done).

To be clear: It is possible to do this, but to deploy such tool you need a lot of Galaxy administrator overhead and you do not get much benefit of the Galaxy framework

ADD COMMENT • link written 19 months ago by Martin Čech ♦♦ 4.9k

We moved from SQLite to MySQL because it could not meet computational needings.

Thanks the same Martin!

ADD REPLY • link written 19 months ago by francescomusacchia • 20

Similar posts • Search »