get_gse_data.r

library(GEOquery)
# Set libcurl as the prefered method to use when downloading files
options(download.file.method.GEOquery='libcurl')

# Create function to get expression set of given gse
# along with the gene symbols
get_gse_expression_set <- function(gse_name, symbol_col_name) {
  # Get the GSE by name and create a dataframe out of the expression set
  gse <- getGEO(gse_name, GSEMatrix = TRUE)[[1]]
  gse_df <- as.data.frame(exprs(gse))
  gse_df$probe_id <- rownames(gse_df)

  # Extract Gene Symbols and associated Probe IDs
  feature_data <- attr(gse, "featureData")
  symbols <- attr(feature_data, "data")[ , symbol_col_name]
  probe_ids <- attr(feature_data, "data")$Probe_Id

  # Create a dataframe to hold the symbols and probe IDs
  symbol_df <- data.frame(symbol=numeric(length(symbols)),
                          probe_id=character(length(probe_ids)) )
  symbol_df$symbol <- symbols
  symbol_df$probe_id <- probe_ids

  # Merge the dataframes together
  complete_data <- merge(x=gse_df, y=symbol_df, by='probe_id', all.x = TRUE)
}

base_path <- "./"  # where to put the resulting file
gse_name <- 'GSE52253'
symbol_col_name <- 'Symbol'
my_data <- get_gse_expression_set(gse_name, 'Symo')
write.table(my_data, file=paste(base_path, gse_name, ".csv", sep=""), sep = '\t', row.names=FALSE)