ZebraQ

Introduction to R / Python and bulk RNA-seq data analysis with DESeq2

An introductory course for beginners in programming and RNA-seq data analysis, built for biotechnology students with little or no bioinformatics background.

The course is available in two parallel tracks: R and Python. They cover the same material, use the same datasets, and reach the same biological conclusions. Which one is taught depends on the cohort; pick whichever suits you.

For more advanced users we recommend the DESeq2 vignette.

---

Course structure

Part 1 — programming and data handling

	Topic
S1	Data structures and basic operations
S2	Importing and exporting data
S3	Summary statistics and data visualisation

Part 2 — RNA-seq analysis

	Topic
S4	Differential expression analysis
S5	Exploring and visualising RNA-seq results
S6	Bonus: hands-on with a public dataset (GSE106118)

---

🐍 Python track

Runs in Google Colab with nothing to install - click a badge and start.

Lesson	Notebook	Open
S1	Python data structures
S2	Importing and exporting data
S3	Exploratory data analysis
S4	Differential expression with PyDESeq2
S5	Visualising RNA-seq results
S6	Bonus — GSE106118

Every notebook is self-contained: it fetches its own data and does not require you to have run any earlier lesson.

Setup instructions for running locally: setup/python/README.md.

---

📊 R track

R markdown sources and rendered HTML are in lessons-r/.

Lesson	Source	Rendered
S1	S1_R_Data_Structures.Rmd	html
S2	S2_Import_Export_Data_in_R.Rmd	html
S3	S3_Exploratory_Data_Analysis.Rmd	html
S4	S4_DESeq_analysis.Rmd	html
S5	S5_DESeq_visualization.Rmd	html
S6	S6_Bonus.Rmd	html

Required R packages: setup/Installation_and_setup.Rmd.

---

Do the two tracks give the same answer?

Close enough to teach either, and the difference is worth understanding.

DESeq2 (R) and PyDESeq2 (Python) are independent implementations of the same method, so they are not bit-identical. Running both on this course's zebrafish dataset (13,755 genes after filtering):

Comparison	Result
Genes tested	13,755 in both
log2 fold change, Spearman correlation	1.000000
Agreement on the significant set	Jaccard 0.996

S4 of the Python track reproduces this comparison live, so students can see it rather than take it on trust.

---

Dataset

Disclaimer: the data used in this course is for educational purposes only.

data/ is shared by both tracks:

File	Contents
chd_genes.annotations.tsv	276 genes associated with congenital heart disease, with gnomAD constraint metrics (also provided as .csv and .xlsx)
salmon.merged.gene_counts.filtered.tsv	Gene counts from RNA-seq of wild-type and mutant zebrafish (Danio rerio) hearts at 48 hpf. Dataset is incomplete and results are therefore uninterpretable
samplesheet.tsv	Sample metadata for the zebrafish experiment (3 WT, 3 MT)
GSE106118_HE10W.tsv.gz	Public human fetal heart expression data used in the S6 bonus lesson

---

Materials

Supporting slides: slides/Practical_Section_II.pdf

---

References

1. R for Data Science
2. Python Data Science Handbook
3. DESeq2
4. PyDESeq2

---

Maintainer

Enrique Audain (enrique.audain@uni-oldenburg.de)

Have fun!