Raw data is not a table
but a heap of letters
What goes into a paper is a tidy count table. What the sequencer actually hands you is twenty-five million fragments of 76 letters.
This course joins the two up. You open real FASTQ (SRA SRR31831920–26), measure its quality, catch the ribosomal RNA that got mixed in, build a k-mer index by hand to attach reads to genes, and check the table you made against the one the paper deposited at GEO.
How you'll learn
Not a made-up example. 44,000 actual reads from SRR31831926, downloaded from SRA, come into your browser — instrument name, quality scores and contamination all exactly as they are.
What the commonest sequence is, why there are only six distinct quality scores, why the SRA record disagrees with the instrument name — none of it gets waved past.
A k-mer index, the reverse complement, first match. You write the skeleton of what salmon does yourself and attach 44,000 reads to 290 genes.
“Is r = 0.76 a good score?” cannot be answered as it stands. Only after a simulation shows the ceiling is 0.91 can you call it 84%.