# Match array probes with gene IDs from biomaRt
# efg, Stowers Institute, 3 Feb 2006

# UNIX:  change "BasePath" based on Windows/UNIX platform
switch(.Platform$OS.type,
         windows = BasePath <- "U:",
         unix    = BasePath <- "/n/projects",
         stop("unsupported OS platform")
      )

filename <- paste(BasePath, "/efg/camda/2006/arrayprobes.csv", sep="")
array.probes <- read.csv(filename, as.is=TRUE)
probe.list <- array.probes$probe
length(probe.list)
length(unique(probe.list))

GOinfo <- NULL

library(biomaRt)
mart <- martConnect()

# Write GO information to disk file, one line at a time
outGONums <- file("ArrayProbeGONums.txt", "w")

for (i in 1:length(probe.list))
{
  probe <- probe.list[i]
  cat(i, probe, "\n")  # show progress
  flush.console()

  # Assume embl ID
  x <- getGO(id=probe,type="embl",species="hsapiens",mart=mart)
  if ( (length(x@table$GOID) == 1) & is.na(x@table$GOID[1]) )
  {
    # IF embl ID fails, try as refseq
    x <- getGO(id=probe,type="refseq",species="hsapiens",mart=mart)
  }

  cat(x@id[1], x@table$GOID, "\n", file=outGONums)

  d <- data.frame(x@table)
  d <- cbind(x@id, d)
  names(d)[1] <- "id"
  GOinfo <- rbind(GOinfo, d)
}

close(outGONums)

martDisconnect(mart)

write.csv(GOinfo, row.names=FALSE, file="ArrayProbeGOInfo.csv")