Sequencing: Bulk RNA-seq homework

1 Default edgeR analysis

Analyze the data using edgeR, by using the code in the RNA-seq analysis intro lecture. In all analyses, focus on the contrast comparing DPN treatment to control at 48h.

2 Impact of blocking

Assess the difference in number of DE genes when not blocking on patient, i.e., removing the patient effect of the model. Compare the p-value distributions between these two models (i.e., with and without blocking).

3 Analyze dataset using full-quantile normalization

3.1 Implement and apply full-quantile normalization

### implement FQ normalization
FQnorm <- function(counts){
  ...
}

### normalize the data using FQ

3.2 Visualize effect of FQ normalization

Visualize the distributions of log1p-transformed counts (use the density function) to compare sample-specific count distributions before and after FQ normalization. What’s the impact of FQ normalization on the differences in distribution between samples?

3.3 edgeR analysis using full-quantile normalized data

Don’t forget to remove the calcNormFactors step in the edgeR analysis as data have already been normalized when using FQ-normalized counts as input!

### use FQ-normalized data as input to the edgeR analysis

3.4 Compare DE genes at 5% FDR

### Get list of DE genes using TMM and FQ normalization

### Compare list using, e.g., a Venn diagram