Fisher's exact for Miller et al.

fishers.R

# Probability of GO term (GeneSet) in gene hits (Hits) from a screen 

# Initialize our relevant variables
nBackground     <- 407 # Total genes screened
nGeneSet        <- 18  # Genes in GO term that are in the bachground
nHits           <- 60  # Genes enriched in screen in background
nOutsideGeneSet <- nBackground - nGeneSet  # Genes not in GeneSet but in background 
nQueryPathway   <- 0:nGeneSet # Overlap between query genes ('Hits') and GO term (in background) 

# Use the dhyper built-in function for hypergeometric probability
probability <- dhyper(nQueryPathway, nGeneSet, nOutsideGeneSet, nHits, log = FALSE)
data <- data.frame( x = nQueryPathway, y = probability )

# Bar plot
library(ggplot2)
ggplot(data, aes(x=factor(x), y=y)) +
  theme(axis.text=element_text(size=14),
    axis.title=element_text(size=18,face="bold"),
    axis.title.x=element_text(margin=margin(20,0,0,0)),
    axis.title.y=element_text(margin=margin(0,20,0,0))
  ) +
  geom_bar(stat="identity", fill= "#2c3e50", colour="black") +
  labs(x = "Hits in GeneSet", y = "Probability")

pval <- sum(probability)