Learning RNA-seq

Most RNA-seq tutorials teach the commands. A student who follows one can produce a volcano plot and cannot tell you what the y-axis means, why the p-values were adjusted, or whether the result is real. That student gets stuck the first time the data misbehaves.

I wrote this to train an undergraduate summer student with a strong quantitative background and no molecular biology. It is concept-first: the statistics come before the code, and every plot is introduced by asking what it should look like if the experiment worked, before producing it.

What is in it

• A concept-only slide deck covering what RNA-seq measures and why, from gene to transcriptome to counts to differential expression. No pipelines, no code to run yet.
• A prequel on what comes off a sequencer, how to read a FASTQ, what read QC catches, and how nf-core/rnaseq produces the count matrix everything else starts from.
• A standalone sandbox on simulated data for the hardest idea in the subject: what a p-value is, why twenty thousand tests break it, and what Benjamini-Hochberg does about that.

It worked with the student it was written for, so I rebuilt it on the public Bioconductor airway dataset and released it. The count matrix is regenerated in the exact shape nf-core emits, so a student who learns on it recognizes real pipeline output.

Code MIT, teaching material CC BY 4.0.