Skip to contents

Getting Niche-DE Genes

Getting_Niche_DE_Genes.Rmd

Extract Niche-DE Genes For a Specific (Index,Niche) Pair

After calculating pvalues, we can find which genes are (index,niche)(index,niche) niche genes at varying resolutions and FDR control levels. The function to do so is called ‘get_niche_DE_genes’ and contains 6 arguments

Arguments
  • object: A niche-DE object
  • test.level: The resolution at which to return genes. There are three choices for test.level;gene level (‘G’), cell type level (‘CT’), and interaction level (‘I’).
  • index: The index cell type of interest
  • niche: The niche cell type of interest
  • positive: Boolean of whether to return positive/upregulated niche-DE genes (True) or negative/downweighted niche-DE genes (False). Default value is True.
  • alpha: The level at which to perform the benjamini-hochberg procedure at each resolution level. Default value is 0.05.

Below, we find interaction level (fibroblast,tumor)+ niche genes.

DE_genes = get_niche_DE_genes(NDE_obj,'I',index='stromal',niche = 'tumor_epithelial',positive = T,alpha = 0.05)
head(DE_genes)

The output is a matrix of genes and their corresponding pvalues at the resolution specified.

Interpretting Your Results

Assume that the direction parameter is set to ‘positive’. The interpretation of your output will differ based on the resolution chosen.

  • Test.level = gene: Genes outputted show some sign of being a niche gene for some (index,niche)(index,niche) pair.

  • Test.level = cell type: Genes outputted are significantly niche up or down regulated in the index cell. The niche cell type is unknown.

  • Test.level = interaction: Genes outputted are significantly upregulated in the index cell type when in the presence of the niche cell type. If ‘pos’ = ‘False’ then Genes outputted are significantly downregulated in the index cell type when in the presence of the niche cell type.

Pathway Enrichement Analysis With Your Niche Genes

Interaction level (index,niche)+(index,niche)+ and (index,niche)−(index,niche)- genes can be input into a pathway enrichment analysis to determine what processes are being up and downregulated in the index cell type when in the presence of the niche cell type. Here we perform pathway enrichment analysis of (fibroblast,tumor)+ genes using the enrichR package.

#Load enichr package
library(enrichR)
#get fibroblast tumor niche genes
fibro_tum_pos = get_niche_DE_genes(NDE_obj,'I',index='stromal',niche = 'tumor_epithelial',positive = T,alpha = 0.05)
#run pathway enrichment analysis
fibro_tum_processes = enrichr(fibro_tum_pos[,1],databases = 'Reactome_2016')
#View processes in a table
View(fibro_tum_processes$Reactome_2016)