gistfile1.r

library(simpleaffy)
library(mouse430a2.db)
library(arrayQualityMetrics)
library(RColorBrewer)
library(pheatmap)

#library(pathprint)
#library(ggplot2)
#library(ggdendro)
#library("AnnotationDbi")
#library("annotationTools")

# Parse array information, normalize with RMA
data.affy <- read.affy(covdesc='covdesc.txt', 
                       path='./data', 
                       verbose=T)
data.affy
pData(data.affy)
data.rma <- call.exprs(data.affy, 'rma')

#
# QC
#
arrayQualityMetrics(data.rma)
 
# Get gene identifiers
entrez <- as.vector(unlist(mget(rownames(all.exprs), 
                                mouse430a2ENTREZID, 
                                ifnotfound=NA)))
length(entrez) == length(rownames(all.exprs))

symbols <- as.vector(unlist(mget(rownames(all.exprs),
                                 mouse430a2SYMBOL,
                                 ifnotfound=NA)))
length(symbols) == length(entrez)

# Extract expression matrix
all.exprs <- exprs(data.rma)
all.unlogged <- 2^all.exprs

# Get mean, standard deviation
rowMean <- apply(all.unlogged, 1, mean)
rowSD <- apply(all.unlogged, 1, sd)

# Create a data frame to hold information
df <- data.frame(all.unlogged)
df$Entrez <- entrez
df$Symbol <- symbols
df$Mean <- rowMean
df$SD <- rowSD

# Sort by mean
dfSort <- df[with(df, order(-Mean)), ]
head(dfSort)
write.table(dfSort, file='unloggedExpression.txt',
            quote=F, sep='\t', row.names=T,
            col.names=T)

# Explore percentiles
quantile(df$Mean)
quantile(df$Mean, probs=seq(0.95, 1, 0.01))