GATK update to 3.4

[kaktus42]

What do you think about updating?
There are a lot of improvements.
Someone already worked on this? Because I just started and didn't want to reinvent the wheel.
Also I'm not sure how thorough I can do it in the time window I have.

[bgruening]

Afaik, nobody of us is working on it. GATK licensing is not ideal and we have worked recently on other tools. If you want to help here this would be awesome! GAKT3 is now also matured enough I guess and the Best practise guide is not changing so much anymore.

I think we need a new datatype to support .g.vcf files.
Moreover, we implemented together with the GATK team a JSON descriptions of all parameters from all tools - available with every release. With this in place our idea was to generate these wrappers more or less automatically.

[kaktus42]

I found that in the gatk, a new bai is always generated when a gatk tool is started. This creates quite an overhead. Isn't there a better solution?
I guess this goes back to the datatype 'bam'. Where is the bai stored? How can it be accessed?
More on this also on the mailinglist "BAM datatype - bam & bai"

[hexylena]

https://github.com/galaxyproject/tools-iuc/blob/0887009a23d176b21536c9fd8a18c4fecc417d4f/tools/bedtools/multiCov.xml see L12 of that file for where the bai file is stored.

[kaktus42]

Ay. Perfect. Like a charm. That helps to tackle the issue I mentioned. 👍

[matthdsm]

Hi all,

I too am interested in an updated version of the GATK wrappers. Any idea if there is time/motivation to write an update from your side?
Where can I find the before mentioned JSON description for the programs and is there any code already in place for the automated wrapper generation?

[kaktus42]

@matthdsm I'm currently working on the new GATK.
You can find it here: https://github.com/kaktus42/galaxytools/tree/master/tools/gatk

I have implemented so far:

• RealignerTargetCreator
• IndelRealigner

[bgruening]

@kaktus42 are you still interested on this? What we could support is a subset that we are allowed to redistribute. Please have a look at: https://github.com/bioconda/bioconda-recipes/tree/master/recipes/gatk/3.5

If we are not allowed to redistribute I would say we can not guarantee reproducible and hence we can not put this into IUC.

[kaktus42]

Hi @bgruening - thanks for the update.
Unfortunately I'm not able to put more time into this currently