Friday Hackathon
By George Kalogiannis & Balig Panossian, Designed from the official QIIME2 tutorials
Scenario
Teams receive paired-end 16S rRNA FASTQ files from faecal samples of three human communities that differ in agricultural practice life styles (hunter-gatherer, agriculture/hunter-gatherer, agriculture). Companion metadata and manifest sheets are provided.
Your mission, by 12:30 on Friday, is to transform these raw reads into an ecological narrative and defend it in a four-slide deck (cover + three content slides, 5 minutes total). You must work around the scheduled times in the agenda.
Optional personal dataset: If your group already has a 16S amplicon dataset that you’d like to showcase, you may substitute it for the workshop dataset—provided it fits the same format we have illustrated (paired-end FASTQ files plus a valid sample-metadata TSV). The grading rubric and four-slide limit remain unchanged, so choose wisely: your own data must still let you demonstrate the full pipeline and draw clear ecological conclusions within the allotted time.
What We’ll Be Grading (implicitly)
-
Judgement, not button-mashing
Why those trim points? Why that diversity metric? -
Story compression
Can you merge QC, processing, and results into three coherent, uncluttered slides? -
Biological connection
Do your stats and graphics convincingly relate microbiome shifts to agricultural practice?
The Files
All files can be found at /mnt/lustre/groups/WCHPC/friday_data
on the CHPC.
Hints: Even though the QC plots may look good, denoising with too high of a length can remove samples that don’t meet certain length requirements. Think about setting the truncation length to near 150-175. When denoising, use
100000
reads to learn the errors from.