This session will introduce participants to the statistical foundations required for analyzing RNA-seq count data. We will focus on generalized linear models (GLMs) and the principles of statistical testing that underpin differential gene expression analysis. Participants will learn how to pre-process and filter a count matrix, fit appropriate models, and interpret key statistical outputs that set the stage for downstream analyses.

1. Experience with R (RStudio Installed)
2. Understand how an RNA-seq count matrix is generated

• RStudio: https://posit.co/download/rstudio-desktop/

• edgeR
• ggplot2
• dplyr
• MASS

• Task of differential gene expression: pairwise vs multigroup comparisons
• Preprocess for fair comparison: Filtering / Normalization
• Statistical testing: Exact test for pairwise
• Linear regression
• Poisson and negative binomial distributions
• Generalized linear model
• Sample analysis using edgeR (R notebook)

• Materials created by Ryan Huang, with figures from the following sources:
• Past MiCM slides: Intro to RNA-seq and Statistics in R (Adrien Osakwe)
• QLSC600 slides: myself and Megan Ng
• RNA-seq lecture by Peter N. Robinson
• Tutorial from Berge and Clement: https://statomics.github.io/SGA/sequencing_countData.html

Workshop created as part of the McGill Initiative in Computational Medicine