Update of Ascat and INSTALL_BIANCA.md

[malinlarsson]

Update of the Ascat process:
In bin/run_ascat.r, in the call to ascat.runAscat() the parameter gamma is set to 1. Gamma was previously not set, which means that the default value 0.55 was used. Gamma =1 is recommended when running Ascat on NGS data (please see https://www.crick.ac.uk/research/labs/peter-van-loo/software). The documentation of Ascat is updated accordingly.

The documentation of INSTALL_BIANCA.md was updated with some information that I learned during my own installation.

PR checklist

• This comment contains a description of changes (with reason)
• If you've fixed a bug or added code that should be tested, add tests!
• Ensure the test suite passes (./scripts/test.sh -p docker -t ALL).
• Documentation in docs is updated
• CHANGELOG.md is updated
• README.md is updated

Learn more about contributing: https://github.com/SciLifeLab/Sarek/blob/master/.github/CONTRIBUTING.md

[maxulysse]

Thanks a lot for the PR.

Have you consider updating to a more recent ASCAT version?
If which cases would we want to change this gamma parameter?

[malinlarsson]

Hi!

Have you consider updating to a more recent ASCAT version?

I noted that there is a new release out (v5.2.5) , so I can test it.

If which cases would we want to change this gamma parameter?

We should always use gamma=1. Other values of gamma should only be used when the input comes from SNP genotyping arrays. Here is an explanation of the gamma parameter from the Ascat documentation: "The parameter represents the drop in LogR for a change from two copies to one copy in 100% of cells. Gamma theoretically should equal 1, but due to array background signal and bespoke array normalisation procedures, in practice it is often significantly lower. …. For massively parallel sequencing data, gamma should always be set to 1. Best, Malin

[maxulysse]

Thanks a lot, that would be marvelous.

Did you manage to make some test data so I can set up some test for ASCAT in Sarek?

[malinlarsson]

Thanks a lot, that would be marvelous. Did you manage to make some test data so I can set up some test for ASCAT in Sarek?

No, I have not yet done that. When I tried to run Ascat for a small chunk of the genome before, but it didn’t work. Do you think we can use the HCC1143 WGS sample as test data? The problem is that it takes quite long to run. Best, Malin

[malinlarsson]

Another update included in this PR (forgot to mention it above):
the output file from Ascat containing all the CNV regions (sampled.cnvs.txt) has been updated. It now includes all segments across the genome, also those with normal copy number (i.e. where nMajor = 1 and nMinor =1). In the previous version the segments with normal copy number were excluded, but I think that is confusing since the native output from Ascat includes all segments.

[malinlarsson]

The do provide data here for ASCAT: https://github.com/Crick-CancerGenomics/ascat/tree/master/ExampleData <https://github.com/Crick-CancerGenomics/ascat/tree/master/ExampleData> And also here for Allelecount: https://github.com/cancerit/alleleCount/tree/dev/testData <https://github.com/cancerit/alleleCount/tree/dev/testData> Maybe we can work something out with that?

I can check what is there. How long time can the test-runs take?

That's good to know, did you add that in the docs?

I added a paragraph about the output files from Ascat in ASCAT.md. But I didn't write out specifically that all segments are included since this is the "default output format" from Ascat.

…

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#628 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADGC5aZ8NsKSfuDqYy8ngdBp98lmcO52ks5uV-NqgaJpZM4WTTMo>.

--------------------------------------------------------------------- Malin Larsson, PhD National Bioinformatics Infrastrucrure Sweden SciLifeLab Bioinformatics Long-term Support (WABI) Department of Physics, Chemistry and Biology House B, Room 2B:586, Campus Valla 581 83 Linköping Telephone: 070-3276037 E-mail: malin.larsson@scilifelab.se ---------------------------------------------------------------------

[maxulysse]

I can check what is there. How long time can the test-runs take?

Typically less than 40 min for a test.
So in our case, it would be from bams to ascat results, as we can't cut it in more pieces.

But I didn't write out specifically that all segments are included since this is the "default output format" from Ascat.

But since this is now the default output format, it could be good to update the docs accordingly and specify that as well, no?

[malinlarsson]

So in our case, it would be from bams to ascat results, as we can't cut it in more pieces.

OK. I have to check if that is possible. Perhaps if we generate a really “shallow” bam file for the full genome.

But I didn't write out specifically that all segments are included since this is the "default output format" from Ascat. But since this is now the default output format, it could be good to update the docs accordingly and specify that as well, no?

Yes, I can ad a note on this.

…

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#628 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADGC5ekKgEe3zgoeW5V1ej52QnEIIWvIks5uV_iugaJpZM4WTTMo>.

--------------------------------------------------------------------- Malin Larsson, PhD National Bioinformatics Infrastrucrure Sweden SciLifeLab Bioinformatics Long-term Support (WABI) Department of Physics, Chemistry and Biology House B, Room 2B:586, Campus Valla 581 83 Linköping Telephone: 070-3276037 E-mail: malin.larsson@scilifelab.se ---------------------------------------------------------------------

[maxulysse]

OK. I have to check if that is possible. Perhaps if we generate a really “shallow” bam file for the full genome

You don't have to do that with this PR, don't worry too much about it.

Yes, I can ad a note on this.

That would be perfect.

[malinlarsson]

OK. I have to check if that is possible. Perhaps if we generate a really “shallow” bam file for the full genome You don't have to do that with this PR, don't worry too much about it.

OK, perhaps when updating Ascat then.

Yes, I can ad a note on this. That would be perfect

I have updated my ASCAT.md file with this info. Do I need to create a new PR?

[maxulysse]

@malinlarsson
I changed the base for the PR from master to dev, as we're now doing that for Sarek.
I also fixed some merging issues.
If you're OK with all the changes, I'll merge it as soon as I have updated iGenomes as well.
Thanks a lot for the PR

[malinlarsson]

This PR now includes GC correction for Ascat. The process has been documented in docs/ASCAT.md, including how the GC correction files were generated.

Still to do: Update to Ascat 2.5.2, but can we do that in a separate PR?