Tool data tables and permissions

Question: Tool data tables and permissions

4.2 years ago by

France

abretaud • 20 wrote:

Hello,

I have just discovered the Data Managers to manage the Tool Data Tables, and I find it very useful!

However, I would like to use this system with some genomes that are not yet public. Is there some way to restrict the access to the genome and the generated indexes to a specific group of users of my galaxy instance?

I know about the Data Libraries that allow to tweak the permissions easily, but with this solution, the various tools must be able to use indexes that are in the history, which is not always the case.

Thanks

Anthony

permission data manager • 1.1k views

ADD COMMENT • link •

modified 4.2 years ago • written 4.2 years ago by abretaud • 20

4.2 years ago by

Jennifer Hillman Jackson ♦ 25k

United States

Jennifer Hillman Jackson ♦ 25k wrote:

Hello,

You are close to a potential solution. Would putting just the reference genomes in a shared data library, then allowing each user to make use of the Custom Reference Genome function work for you? This should function with nearly all tools, even those that require a "database" assignment (by going further and creating a "Custom Build" and assigning it to datasets).

Help for Custom Reference Genomes is here:
http://wiki.galaxyproject.org/Support#Custom_reference_genome
http://wiki.galaxyproject.org/Learn/CustomGenomes

Best, Jen, Galaxy team

ADD COMMENT • link written 4.2 years ago by Jennifer Hillman Jackson ♦ 25k

4.2 years ago by

abretaud • 20

France

abretaud • 20 wrote:

Hello,

That is an interesting solution, but in this case the indexes will be regenerated each time a user uses one of the reference genomes?

In fact we already generate all the indexes outside galaxy because some users use them with command line. I thought using data managers would allow me to load both fasta files and pregenerated indexes into galaxy without needing to recompute them (and without storing indexes multiple times too). In this case, the only thing that is missing is a permission system. Is it something that is planned some day?

Anthony

ADD COMMENT • link written 4.2 years ago by abretaud • 20

Permissions per reference genome are an interesting idea. Implementing this is probably non-trivial, as genomes are not a core "object" in the same way that datasets, histories, workflow, etc. are. I created a Trello card for it for the team to review: https://trello.com/c/E33Z1cvt

Meanwhile, if the genomes are small (bacterial, etc) the custom ref genome option shouldn't be too big of problem. Scaling up memory will help with larger genomes.

Good idea, hope that you are able to find a solution for now, then follow the ticket to see future development. Take care, Jen

ADD REPLY • link written 4.2 years ago by Jennifer Hillman Jackson ♦ 25k

Please log in to add an answer.

Similar posts • Search »