Relatedly, but for a narrower audience, I also adapted the Snakemake SLURM executor plugin to perform automatic partition selection on the Harvard Cannon cluster. I was surprised by the amount of flexibility the plugins allowed, which may have finally won me over to them.

github.com/harvardinfor...

Each task has an associated tutorial on our group's website:

informatics.fas.harvard.edu/resources/#t...

Feel free to reach out if you want to use any of these workflows and need help getting started or run into any issues!

Hmm, yea that could work. I wonder if there is any mouse data that could work for this? @jeffreygood.bsky.social
Although, I'm not sure how to map shortbreads. Maybe some batches with BWA-kery? :D

Yea that would definitely help pinpoint which SNPs are being miscalled. I was also hoping to dig into why they are being miscalled - mis-mapped reads? unmapped reads? something else? Which I can't think of how to do without a truth set for the mappings themselves.

We've (@jeffreygood.bsky.social) implemented this in pseudo-it (github.com/goodest-good...), though hard to quantify the effects without good simulations, which I've yet to find the right read simulation program for. Ideally want simulated bam and vcf to compare mapping and SNP calls. Ideas welcome!