@ahda We only recommend FilterAlignmentArtifacts for users who really want precision at the cost of sensitivity. For that reason it's turned on in the Mutect2 pipeline (WDL) by default. This will likely not be the case with an upcoming future version, but that's another story.
Anyway, this is not a bug, just a fact of trying to find mapping errors from short read data. The Sanger sequencing presumably uses reads of 800 bp or so and is therefore much more mappable.
Terra questions would be best asked on the Terra forum found here.
When a task uses a preemptible machine and it gets preempted that means the VM your command was running on was stopped by another job needing its resources. Terra will restart your task on another preemptible VM or regular VM until the job finishes, hence the multiple messages about preemption.
The first error message you posted was probably the last message before the workflow stopped, which is the real reason why the workflow failed. It says for some reason the task failed before it completed, which indicates the problem doesn't lie with Terra's platform but with the action being run on the VM. For more details on why the job failed you would have to click on the paper and folder icons to look at the logs for that particular scattered task job.
@zjwang The -R argument is expecting a .fasta file rather than a .con file.
@mshand thank you for catching the error and sorry I missed it!
@zjwang correcting the input to -R argument should fix the issue.
Try to add a file path that puts the outputs somewhere easy to find on your system.
I think the syntax error of two ".." in your the output designation "output.snp..plots.R " is not putting the outputs in a place you can find them.
In the instructions on how to use the tool, it is " -rscriptFile output.plots.R" if the file is not Allele Specific and " -rscriptFile output.plots.AS.R" if the file is Allele Specific.
@crackiee I agree with you that well-known oncogenic variants should not be filtered as germline variants. One way to achieve this would be by running Funcotator on the output of FilterMutectCalls and then removing the germline filter from any variants annotated as oncogenic.
The command lines above look plausible.
Probably the best place to start with the new CNV pipeline is the online documentation here:
HaplotypeCaller should automatically set fixed seed. There is a bug in reproducibility in some cases when using very large number of bams, and this is being worked on by the dev team. I am not sure about the timeline on when this would be worked on, but for all intents and purposes HaplotypeCaller should automatically set fixed seed.
@ahda @bhanuGandham I traced this to a mistake in our code where a < should be a <=, causing one extra variant to be filtered. Normally this extra variant is, by definition, on the border between filtering and not filtering and so this isn't noticeable, but in your case there are three obvious somatic calls and thirteen obvious calls that should be filtered, in which case dropping one of the somatic calls is clearly the wrong thing to do. Thank you for spotting this and I will try to fix it this week.
Hi @SkyWarrior, yes, your assessment is correct. We changed the output format of some of the files to add output of concatenated denoised copy ratios in https://github.com/broadinstitute/gatk/pull/5823, but neglected to mention that this breaks backwards compatibility. Thanks for bringing it to our attention, we'll amend the release notes.