I have log2 fold changes from a microarray experiment in which the expression of 768 genes were measured in two groups of patients. Both groups only have 4 patients in them. Additionally, the specific gene expression assay used was the NanoString nCounter Metabolic Pathways Panel. I was wondering if it's ok to run fgsea with this data, even though it's from a small number of patients and only a (relatively) small number of genes were measured, all related to metabolism. I recognize that I'll probably miss enriched gene sets that are not related to metabolism, but in terms of detecting enriched gene sets related to metabolism, will the analysis still work?