5.4.8-9. Differential gene expression and gene set analysis.

5.4.8. Finding differentially expressed genes (DEG)

• Common analysis when working with bulk RNA-seq
  • 2 conditions, e.g. mutant and wild-type
• Common packages:
  • DESeq2 and
  • edgeR,
• But single-cell data is somewhat different, so

​

Differential gene expression in single-cell data

Differences

• Usually do not have a defined set of experimental conditions
  • Rather you derive the clusters
    • Cluster vs background (all)
    • Cluster 1 vs Cluster 2
• The distribution of data is very different: think of the zeros.
  • → Assumptions of bulk methods might not be valid
  • → New methods are designed for scDE, and they are typically part of pipelines.

5.4.9. Gene set analysis

Problem:

• Many DEG, but what do they really mean?
• Is there a common function?
• GO-term analysis can help summarising the results.

Each gene is associated with GO-terms

• Calculate enrichment.

$$Terms.in.your.genes \over Terms.in.ALL.genes$$

GO-term annotation & protein interaction networks

• Open up STRING website for interactive exploration