- Start with the module cards. Each module has two or three key readings. If you read only those, you have the core of the subject.
- Most of it is free to read. Papers with a free full text link to PubMed Central or the publisher; the few that may need a subscription are marked.
- The questions have no answer key. Use them to check your understanding after a module, or to discuss with other learners.
Module by module
The key readings for each module, and questions to think about once you have finished it. Full references are under Papers and guides.
Module 1 · FoundationsR basics
R works on whole vectors and data frames at once. Tidy data, one row per observation and one column per variable, makes every later step simpler.
Think about- Why does R quietly reuse a vector of 3 values when you add it to one of 6, but warn you when the shorter vector has 4 values?
- What would make a sample sheet "untidy", and how would you fix it?
Module 2 · IntermediateStatistics
A test answers a narrow question. Effect sizes with confidence intervals say how big a difference is; p-values alone do not.
Think about- A study reports p = 0.04 with n = 8 per group. What else would you want to know before believing it?
- Why does testing 20,000 genes at p < 0.05 almost guarantee false positives?
Module 3 · IntermediateFigures that tell the truth
A figure answers one question. Show the data, say what error bars mean, use colour that everyone can read, and start bars at zero.
Think about- When is a bar chart the right choice?
- Which figure in a recent paper you read would you redraw, and how?
Module 4 · AdvancedRNA-seq analysis
DESeq2 models counts per gene, shares information across genes to estimate variability, and shrinks noisy fold changes towards zero.
Think about- Why does DESeq2 need raw counts, not normalised values?
- A gene has log2 fold change 5 but padj 0.4. How is that possible?
Papers and guides
The papers, books and guidance behind the course, grouped by module.
Module 1 · R basics
Use R for Data Science as a reference while you work through the lessons, and read Wilson et al. once you start your own projects.
- Wickham H, Çetinkaya-Rundel M, Grolemund G (2023). R for Data Science, 2nd edition. free online book. Free online book covering R basics, data import, tidying and visualisation.
- Wilson G et al. (2017). Good enough practices in scientific computing. PLOS Computational Biology 13:e1005510. Simple habits for reproducible projects.
Module 2 · Statistics
Greenland et al. is the longest of these; the other two are short and can each be read in one sitting.
- Altman DG, Bland JM (2005). Standard deviations and standard errors. BMJ 331:903. Which one to report, in one page.
- Greenland S et al. (2016). Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. European Journal of Epidemiology 31:337–350. A guide to what p-values and confidence intervals do and do not mean.
- Noble WS (2009). How does multiple testing correction work?. Nature Biotechnology 27:1135–1137. Multiple testing and false discovery rates, with a worked example.
Weissgerber et al. and Rougier et al. are quick reads to return to before you make any figure; the ggplot2 book is for looking things up.
- Weissgerber TL et al. (2015). Beyond bar and line graphs: time for a new data presentation paradigm. PLOS Biology 13:e1002128. The case for showing every data point.
- Crameri F, Shephard GE, Heron PJ (2020). The misuse of colour in science communication. Nature Communications 11:5444. Colour maps that do not mislead.
- Rougier NP, Droettboom M, Bourne PE (2014). Ten simple rules for better figures. PLOS Computational Biology 10:e1003833. Ten rules that apply to every figure.
- Wickham H, Navarro D, Pedersen TL. ggplot2: Elegant Graphics for Data Analysis, 3rd edition. free online book. The full ggplot2 reference, free online.
Module 4 · RNA-seq analysis
Start with the workflow by Love et al. 2015, which follows the same airway analysis step by step, then turn to the DESeq2 paper for the reasoning behind each step.
- Conesa A et al. (2016). A survey of best practices for RNA-seq data analysis. Genome Biology 17:13. Best practice for every step of RNA-seq analysis.
- Love MI, Huber W, Anders S (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology 15:550. The DESeq2 method paper.
- Love MI et al. (2015). RNA-Seq workflow: gene-level exploratory analysis and differential expression. F1000Research 4:1070. A complete Bioconductor workflow on the airway data.
- Himes BE et al. (2014). RNA-Seq transcriptome profiling identifies CRISPLD2 as a glucocorticoid responsive gene that modulates cytokine function in airway smooth muscle cells. PLOS ONE 9:e99625. The published airway experiment.
- Soneson C, Love MI, Robinson MD (2015). Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences. F1000Research 4:1521. The tximport method and why transcript-level estimates help.
- Leek JT et al. (2010). Tackling the widespread and critical impact of batch effects in high-throughput data. Nature Reviews Genetics 11:733–739. What batch effects are and how to design around them.
- Subramanian A et al. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. PNAS 102:15545–15550. The original GSEA paper.
The packages used in the course, with the versions the code was tested on. Newer versions usually work; if output differs, check the package's news page.
| Tool | What it does | Module | Version in the course |
|---|
| R | The language | All | 4.4.3 |
| ggplot2 | Figures | 1, 3 | 4.0.3 |
| dplyr, tidyr | Cleaning and reshaping data | 1–3 | Current CRAN |
| MASS | The birth weight data in Module 2 | 2 | Ships with R |
| patchwork | Multi-panel figures | 3 | 1.3.2 |
| pheatmap | Heatmaps | 3 | Current CRAN |
| DESeq2 (Bioconductor) | Differential expression | 4 | 1.46 |
| tximport (Bioconductor) | Importing Salmon and kallisto estimates | 4 | 1.36 |
| clusterProfiler (Bioconductor) | GO and GSEA | 4 | 4.16 |
Glossary
- Confidence interval
- A range of values for the true effect that is consistent with the data.
- Cross-validation
- Estimating how well a model predicts by testing it on data held out from training.
- Data frame
- R's table: columns of equal length, each with one type.
- Effect size
- How big a difference or association is, in meaningful units.
- False discovery rate
- The expected share of false positives among the results called significant.
- Fold change
- The ratio of a gene's expression between two conditions; usually reported on a log2 scale.
- Odds ratio
- How much the odds of an outcome change between groups.
- p-value
- How surprising the data would be if there were truly no effect.
- Power
- The chance a study detects an effect of a given size, if it exists.
- Regression
- A model of how an outcome changes with one or more predictors.
- Size factor
- DESeq2's per-sample scaling for sequencing depth.
- Standard error
- How precisely a mean is known: SD ÷ √n.
- Tidy data
- One row per observation, one column per variable.
- Vector
- R's basic object: values of one type in order.
- Volcano plot
- Fold change against −log10 p-value, one point per gene.
Going further