Facile (Gene) Set Expression Analysis

Under Construction

The facile feature set enrichment analysis leans on the sparrow framework for running GSEA methods, as well as providing gene set collections via sparrow::GeneSetDb objects.

This vignette assumes that you’ve got a GeneSetDb object in hand named gdb.

library(FacileData)
library(FacileAnalysis)
samples <- exampleFacileDataSet() %>% 
  filter_samples(indication == "BLCA")

For now, I imagine (gene) set enerichment analyses will be performed over the result of other analyses. This means that we can perform (G)SEA on:

The result of a differential expression analysis (fdge) using:

Pre-ranked methods like cameraPR and fgsea
Enrichment methods (goseq, hypergeometric testing)

or some combination of the two.

gdb <- sparrow::exampleGeneSetDb()

dge.gsea <- samples %>% 
  flm_def(covariate = "sample_type", 
             numer = "tumor", denom = "normal",
             batch = "sex") %>% 
  fdge() %>% 
  ffsea(gdb, method = c("cameraPR", "ora"), rank_by = "t")

The “highly loaded” genes from the first prinicpal component of a principal components analysis (fpca). Deafult GSEA methods are set for each analysis result type, too, if you don’t want to specify them.
```
pc1.gsea <- samples %>% 
  fpca() %>% 
  ffsea(gdb, pc = 1)
```

In the immediate term, the GSEA functionality will lean heavily on the sparrow package, including the bits that enable interactivity over the results (ie. the sparrow.shiny package).

Comparing (Gene) Set Enrichment Results

To compare GSEA results, Thomas’ idea of overlaying enrichment maps from the two results on top of each other in some dynamic way may be the first thing we try.

Steve Lianoglou

2023-04-03