Skip to content

Lecture notes. Written from the material of StatOmics Sga21, licensed CC BY-NC-SA 4.0. These are notes, not a transcript: the material has been reorganised and rewritten. This adaptation carries the same licence, and the original is linked above.

10. Course Description and Program

What this covers

This chapter is the course's own front matter: what the Statistical Genomics Analysis (SGA) course is about, who it is for, and how its sessions are organised into modules. It assumes nothing about proteomics or RNA-sequencing — that material starts in the modules described below — but does assume the general statistics and R background listed under Prerequisites.

Scope

High-throughput 'omics studies generate large, complex datasets, and this course is built around the statistical concepts needed to interpret them correctly: preprocessing, quantification, and differential analysis. Two technologies carry the material throughout — shotgun proteomics, and (bulk and single-cell) RNA-sequencing. Experimental design runs through both: the course covers how to design a statistically efficient experiment, and how the design shapes the model used on the resulting data, introducing concepts such as blocking as a consequence of design rather than an afterthought. All computation uses free, open-source tools in R/Bioconductor.

The source page opens with a workflow diagram (wpGraph.jpeg) situating the course among preprocessing, quantification and differential analysis; the image itself is not part of the converted material, so it is not reproduced here.

Audience and prerequisites

The course targets biologists and bioinformaticians with a particular interest in differential analysis of quantitative 'omics data.

It assumes a basic statistics course covering data exploration and descriptive statistics, and statistical modelling and inference: linear models, confidence intervals, t-tests, F-tests, ANOVA, and the chi-squared test. Two online refreshers are pointed to for this material: an English-language course at gtpb.github.io/PSLS20 and a Dutch-language one at statomics.github.io/statistiekCursusNotas.

It also expects some programming ability in R. A primer and a data-visualisation course, both hosted on Ghent University's Dodona platform, are given as a starting point for students who need to pick this up (R Basics and R Data Exploration).

Software

Everything runs in R, version 4.1.1 or later, with RStudio recommended. The course repository supplies a single installation script that pulls in every package used later in the course:

source("https://raw.githubusercontent.com/statOmics/SGA21/master/install.R")

Anyone who cannot get a local installation working can fall back to a cloud RStudio instance with the packages pre-installed — offered for the duration of the course, not for routine use.

Structure of the course

The programme runs in a recap plus three modules, moving from a shared statistical toolbox (linear models) into two 'omics technologies that both quantify abundance from high-throughput measurements, then into a technology — single-cell RNA-seq — where the unit of measurement changes from a bulk sample to a single cell.

Recap: linear models (week 1)

A lecture recapping the general linear model, paired with a tutorial that applies multiple regression to the KPNA2 dataset. This is the statistical base the rest of the course builds on.

Module I: proteomics data analysis (weeks 1-5)

  • Bioinformatics for proteomics — an introductory lecture (with an accompanying video) and tutorials on peptide/protein identification.
  • Preprocessing and analysis of simple designs (week 3) — a lecture on preprocessing quantification data, a matching tutorial, and a wrap-up on peptide-based models for robust summarisation.
  • Statistical inference and analysis of factorial designs (weeks 3-4) — a lecture on differential abundance analysis, a tutorial on experimental design, and a wrap-up on blocking.
  • Advanced and reading material — technical details on inference after summarisation, stage-wise testing, and three papers: Sticker et al. (2020) on robust summarisation and inference in label-free quantification, Ludger Goeminne's PhD-introduction background chapter on proteomics data analysis, and Van den Berge et al. (2017), the stageR stage-wise testing paper.
  • Worked solutions — the CPTAC study, a 6x6 cancer design, a mouse study, and a heart study, each solved in an accompanying wrap-up or dedicated page.

Module II: bulk RNA-sequencing

  • Introduction to sequencing technology, raw data and preprocessing, with a shell-script tutorial for preprocessing raw RNA-seq data and a recommended review ("RNA Sequencing Data: A Hitchhiker's Guide to Expression Analysis").
  • Working with count data and generalized linear models, including a tutorial that analyses bike-sharing count data as a warm-up before fitting a GLM to a single gene.
  • Analysis of RNA-seq data: the four major challenges in analysing RNA-seq datasets, plus a page arguing for using offsets rather than scaling the counts directly.
  • Technical topics in bulk RNA-seq differential expression analysis.

Module III: single-cell RNA-sequencing

General concepts and the analysis workflow for single-cell RNA-seq data, covering the variance-stabilising transformation for a Poisson random variable, an external explainer on UMAP, and a simulation study of post-selection inference — the problem of testing on the same data used to select which genes or clusters to test. These are tied together in a reproducible workflow on the Macosko Drop-seq dataset.

Instructors and license

The course is taught by Koen Van den Berge, Lieven Clement, and Lennart Martens.

The course material is licensed CC BY-NC-SA 4.0 (Creative Commons Attribution-NonCommercial-ShareAlike): reuse must give attribution, may not be commercial, and any derivative must carry the same licence.

Sources

  • Course description: docs/omics-statistics/statomics/sga21/home/01-course-description.md, converted from index.Rmd in the statOmics/SGA21 GitHub repository (commit 0ad787d).
  • Detailed program: docs/omics-statistics/statomics/sga21/home/02-detailed-program.md, converted from the same index.Rmd.
  • Named but not contained in these two pages, and not covered by this chapter: the individual lecture slide decks, tutorials, and shell scripts linked from the detailed program, the three papers cited (Sticker et al. 2020; Van den Berge et al. 2017; Goeminne's background chapter), and the wpGraph.jpeg workflow diagram referenced but not embedded in the converted page.

← 9. Illumina Next-Generation Sequencing Overview · Contents · 11. Course Description →