DTU analysis

Below you find the material for doing differential transcript usage analysis: identifying isoforms that are used in one group of samples and not in another group. You can do this starting from:

• short reads generated by bulk RNASeq analysis: you need the output of Salmon or Kallisto for this
• long reads generated by bulk RNASeq analysis: you need the output of GenomeComb for this
• long reads generated by sc RNASeq analysis: you need the output of scywalker for this, it performs pseudobulking on transcript-level counts by for each gene taking the sum of the counts per set of cells that you want consider as one sample, e.g. all cells from one sample that belong to a specific cluster or all cells from one sample.

In the tutorial we use transcript-level counts obtained by pseudobulking on the full sample level of ONT-based sc RNASeq data.

• Tutorial on DTU analysis of transcript-level counts
• R markdown for generating the tutorial
• Simple R script for DTU analysis of transcript-level counts
• Transcript-level count matrix to use as input for DTU analysis

Note that this analysis assumes you have biological replicates: multiple samples in each group you want to compare!