chip-seq pipeline fails ataqv because of the old 1.0.0 version

[naumenko-sa]

Hello everyone!

Happy New 2020!

When running a CHIP-project with many samples, I found that it fails ataqv step for one sample.
A similar memory issue was described and solved before:
ParkerLab/ataqv#2

We are still using ataqv 1.0.0 in bcbio
https://github.com/bioconda/bioconda-recipes/blob/master/recipes/ataqv/meta.yaml

I have built ataqv 1.1.1 from source and it passes the QC step for that sample.

In our installation I manually fixed ataqv to run v.1.1.1 by default to finish the project,
but we need the updated one installed via bioconda.

Easy to suggest, but not so easy to achieve.
First, I asked ParkerLab to make a release:
ParkerLab/ataqv#9

The next step would be to fix bioconda recipe and bcbio requirement with >=1.1.1

Sergey

[naumenko-sa]

Ataqv developer suggested to call ataqv with human and without --autosomal-reference-file and it was indeed better at least for one sample which failed otherwise.

Should we get rid of --autosomal-reference-file or there are other reasons to call ataqv this way?

[naumenko-sa]

We have fixed it by updating ataqv to 1.1.1