Episode 03: Variant Calling Workflow Changes

[fpsom]

Edits for Episode 03: Variant Calling Workflow

• Replace copying subsampled and trimmed data from hidden directory with downloaded a subsampled and trimmed data set from figshare with curl.
• add information on all of the flags used in the different commands
• replace samtools mpileup with bcftools mpileup
• Explain commandline structure for samtools mpileup and bcftools. See Issue #56 of original repo
• Using bwa mem instead of bwa aln. See also Issue #111 of original repo
• Add locations of true and false positive variant calls. See also Issue #60 of original repo
• Additional exercise about SAM format. See also Issue #50 of original repo
• Need exercises in variant calling. See also Issue #91 of original repo

[taylorreiter]

@fpsom @crazyhottommy I reimplemented this variant calling workflow for one of my own samples yesterday and noticed that the samtools and bcftools versions are out of date. Right now the lesson uses version 0.1.19 for both tools. samtools is now on v 1.3.1 and bcftools is on 1.8. Can we move to the newer versions of the tools? I have code to do so.

Note that bcftools v 0.1.19 does not have mpileup.

[fpsom]

I definitely agree with that. And I would propose that all tools involved are upgraded to a later (if not the latest) version - i.e. trimmomatic, samtools, bcftools and bwa afak.

The downside is that we'll probably need to upgrade the entire AWS image...

[taylorreiter]

Referencing this original issue.

@fpsom @ErinBecker I have added exercises to 02-variant-calling so there are now three. Would you still like there to be exercises about the SAM format?

[fpsom]

Thanks @taylorreiter ! It looks good as is, but I would propose the following; we keep the two existing exercises as part of the episode, and we also include an exercise on the SAM format as an "advanced exercise". Knowing SAM format is in itself a challenge, so only people who are genuinely interested should be encourage to do it. And I wouldn't remove it altogether, as the format is the base input of a significant number of tools.

[taylorreiter]

Ok! I think that is a great idea. It will be a while before I have a chance to work on this, but I'll keep the issue open so we can address it. Thank you!