Exercice 3

Read metadata

• Go into the data/ directory and read in R the Toy_Metadata.txt file

Show answer !

Many comments are present in the file, they are all skipped by the default option comment.char=“#”

metadata<-read.table("data/Toy_Metadata.txt",header=T)

metadata

Read ReadCounts data

• Control that all in metadata are present in data/Toy_STARcounts/ directory (using code !)

• Based on comments found in metadata, write a function that:

  • Take sample name as argument
  • Read STAR gene count file
  • skip first 4 lines
  • return counts from the 4th column

• Read all Sample file, to have at the end a matrix of count, matrix col.names should be Sample ID and matrix row.names should be Gene

Show answer !

First we control that all files in Toy_STARcounts directory can be found in metadata

# We list all file in data/Toy_STARcounts/
ReadCountsFiles<-list.files("data/Toy_STARcounts/",pattern = ".out.tab")
# We remove "ReadsPerGene.out.tab" from the name to only have sampleID
ReadCountsFiles<-gsub("ReadsPerGene.out.tab","",ReadCountsFiles)
# We control that all file in metadata are present:
metadata$SampleID %in% ReadCountsFiles

## [1] TRUE TRUE TRUE TRUE TRUE TRUE

# We could test in we find a FALSE results:
any((metadata$SampleID %in% ReadCountsFiles)== FALSE)

## [1] FALSE

Then we write our function the read data from a sample ID

ReadCountFunction<-function(SampleID){
  # skip = 4 allows to skip 4 rows
  STARcounts<-read.table(paste("data/Toy_STARcounts/",SampleID,"ReadsPerGene.out.tab",sep=""),skip=4)
  # We take the column 4
  GeneCount<-STARcounts$V4
  # We name our vector of counts
  names(GeneCount)<-STARcounts$V1
  # We return our vector of counts
  return(GeneCount)
}

We then apply our function to all sampleID

counts<-sapply(metadata$SampleID,ReadCountFunction)

We have indeed a matrix of counts

class(counts)

## [1] "matrix" "array"

head(counts)

##                    T6N2 T6N3 T6N4 T6S1 T6S3 T6S4
## ENSMUSG00000102693    0    0    0    0    0    0
## ENSMUSG00000064842    0    0    0    0    0    0
## ENSMUSG00000051951 1436 1389 2166 1434 1865 1514
## ENSMUSG00000102851    0    0    0    0    0    0
## ENSMUSG00000103377    0    5    1    2    0    0
## ENSMUSG00000104017    0    3    0    1    0    0

Read Log files

• Do the same as above for STAR Log file name Log.final.out (The file is a bit harder to read, you can use the option “fill” in the read.table() function)

• Plot the number of Uniquely mapped reads

Show answer !

ReadLogFunction<-function(SampleID){
  # use fill to avoid error when data don't have 2 values
  STAR_Log_Data<-read.table(paste("data/Toy_STARcounts/Log/",SampleID,"Log.final.out",sep=""),sep="\t",header=F,fill=T)
  STAR_Log<-STAR_Log_Data$V2
  names(STAR_Log)<-STAR_Log_Data$V1
  # For better reading, we transform space and | character
  names(STAR_Log)<-gsub(" ","",names(STAR_Log))
  names(STAR_Log)<-gsub("\\|","",names(STAR_Log))
  return(STAR_Log)
}

Get Log for all files

Logs<-sapply(metadata$SampleID,ReadLogFunction)

Plot data

plot(log10(as.numeric(Logs["Uniquelymappedreadsnumber",])),ylab="log10 Uniquely Mapped Reads")