Lesson 6: Statistics for Differential Expression in Microarray Studies

Printer-friendly versionPrinter-friendly version
Key Learning Goals for this Lesson:
  • Developing a design matrix
  • Understanding the moderated t-test
  • Understanding why increasing power improves both the FDR and FNR

After preprocessing and normalization, the data are in an expression matrix usually with genes in the rows and samples in the columns.  The basic analysis is a test for each gene. We need to account for the experimental design in the analysis - for example we may have paired samples or technical replicates.

We are going to demonstrate with the ColonCA data.