Functional analysis of DE genes
Mapping gene identifiers
Most tools for functional analysis expect a specific type of gene IDs. The genes in our data set were represented by Ensembl IDs and we have added gene symbols in our R code. However, gene symbols are not always available. Note that there are also non-R alternatives to add info to a list of genes.
Check out the topic on gene ID mapping.
Functional enrichment analysis
Check the Functional analysis course for
- details on ORA (overrepresentation analysis) of lists of DE genes
- details on GSEA (gene set enrichment analysis) of RNASeq data
- When you try different methods you will see that they don’t agree. This is why we created a tutorial and code for decoupleR, an R package that supports 12 different methods (ORA, GSEA, GVA…) and lets you easily compare and combine them.
More details on simple ORA or GSEA in R can be found in the biological data mining textbook.
If you really just want to do GSEA, you can find the R code and tutorial for fGSEA (fast GSEA) in our course on the analysis of single cell RNASeq data. You can also use this code for bulk RNASeq data but you have skip the part with the presto package and the first aggregation of the counts for the heat map.
Below you can find the tutorial and accompanying R code for rROMA, a tool that is not in decoupleR and that performs functional enrichment analysis, similar to GSEA, on bulk RNASeq data. On the demo data set of the training it generates meaningful results.