getData.R

# install.packages("data.table", "dplyr")
library(data.table)
library(dplyr)


parseProperly <- function(x){
  x <- x[,-c(ncol(x)-1,ncol(x))]
  x
}

# Read data drectly from GEO
m1 <- "https://ftp.ncbi.nlm.nih.gov/geo/series/GSE124nnn/GSE124167/suppl/GSE124167_FINAL_master_list_of_gene_counts_MIN.V4.txt.gz"
m2 <- "https://ftp.ncbi.nlm.nih.gov/geo/series/GSE124nnn/GSE124167/suppl/GSE124167_FINAL_master_list_of_gene_counts_MIN.sense.Pico.R.txt.gz"
m3 <- "https://ftp.ncbi.nlm.nih.gov/geo/series/GSE124nnn/GSE124167/suppl/GSE124167_FINAL_master_list_of_gene_counts_MIN.sense.Truseq.txt.gz"

method1 <- parseProperly(fread(m1, data.table=FALSE))
method2 <- parseProperly(fread(m2, data.table=FALSE))
method3 <- parseProperly(fread(m3, data.table=FALSE))

# Combine into a single table
counts <- Reduce(function(x, y) full_join(x, y, by="id"), list(method1, method2, method3))

# Replace NAs by zero
counts[is.na(counts)] <- 0

# Move "ids" to rownames to have a ure count matrix:
rownames(counts) <- counts$id
counts$id <- NULL