R/generatingCDSaaFile.R

In geno2proteo: Finding the DNA and Protein Sequences of Any Genomic or Proteomic Loci

generatingCDSaaFile <- function(geneticCodeFile_line, gtfFile, DNAfastaFile, 
                                outputFolder='./', perlExec='perl') {

    scriptFolder= system.file("exec", package="geno2proteo")
    # the perl script used to get the DNA for the CDS from the fasta file.
    perlScript_getDNAseq=file.path(scriptFolder, 
        "getDNAseqsFromGenome_loci.perl")
    # the perl script for translating a DNA sequence into a AA sequence, 
    # using the genetic codes provided.
    perlScript_getAAseq=file.path(scriptFolder, "DNAseqTranslation_CDS.perl") 
    filenamePrefix = sub('.*/', '', gtfFile)
    # read the CDS information from the GTF file 
    kk21 = read.table(gtfFile, sep='\t', stringsAsFactors=FALSE)
    kk = c(1:22, 'X', 'Y', 'MT')
    kk = kk21[,1] %in% kk
    kk21 = kk21[kk,]
    kk = kk21[,3] =='CDS' # CDS are those coding regions
    kk22 = kk21[kk,]
    kkt = kk22[,9]
    # get all the transcript ids
    kk = sub('.*transcript_id ', '',kkt); kk=sub(';.*', '', kk)
    # for each transcipt, get all its CDS (indexes of them)
    kk9991 = tapply(1:length(kk), kk, sort ) 

    # get all the protein ids
    kkt1 = sub('.*protein_id ', '',kkt); kkt1=sub(';.*', '', kkt1)
    # get all the protein ids
    kkt2 = sub('.*gene_name ', '',kkt); kkt2=sub(';.*', '', kkt2)

    pIDgene = cbind(proteinId=kkt1, geneName=kkt2)
    rownames(pIDgene) = kk
    # get the genomic position information of each CDS, and frame for 
    # protein translation
    kk9992 = kk22[,c(1, 4,5, 7, 8)] 
    kk9993 = lapply(kk9991, function(a) kk9992[a,] )
    # order the CDS of each transcript by their start
    kk9994 = lapply(kk9993, function(a) a[order(a[,2]),] )

    kk99941 = lapply(kk9994, as.matrix)

    kk = names(kk99941)

    kk9995 = lapply(1:length(kk), function(n) {
        kkt=kk99941[[n]]; kkt1=1:nrow(kkt); 
        if(kkt[1,4]=='-') kkt1=rev(kkt1); 
        cbind(paste(kk[n], kkt1, sep='_'), kkt) 
    } )

    kk99951 = lapply(kk9995, t)
    kk = unlist(kk99951)
    kk = t(matrix(kk, nrow=6))

    # the 6th column is for the frame of translation
    kk9996 = kk[,c(2:5, 1, 6)] 

    kk = strsplit(kk9996[,5], split='_')
    kk = t(matrix(unlist(kk), nrow=2))
    # matrix containing the ordered CDS for each transcript
    kk9996 = cbind(kk9996[,1:4], kk, kk9996[,6]) 

    ## get a proper bed format
    #kk9996 = kk9996[,c(1:3, 5:6, 4, 7)] # now using the

    #kk9996[,2] = as.numeric(kk9996[,2])
    
    filenameAllCDS = paste(outputFolder, '/', filenamePrefix, 
                            '_CDS_transcript.txt', sep='');
    
    write.table(kk9996, file=filenameAllCDS, sep='\t', 
                            quote=FALSE, row.names=FALSE, col.names=FALSE)
    
    # then run the perl script to get the DNA sequence of the CDS regions
    
    filenameAllCDSdnaSeq = paste(outputFolder, '/', 
                    filenamePrefix, '_CDS_transcript_DNAseq.txt', sep='');
    DNAfastaFile00 = DNAfastaFile
    # uncompress the fasta file, not remove the compressed
    if(grepl('\\.gz$', DNAfastaFile)[1]) { 
        #R.utils::gunzip(DNAfastaFile, remove=FALSE)
        #DNAfastaFile = sub('\\.gz$','', DNAfastaFile)
        destName=paste(outputFolder, '/', 
                    sub('.gz$', '', sub('.*\\/', '', DNAfastaFile)), sep='')
        R.utils::decompressFile(filename=DNAfastaFile, destname=destName, 
                                ext='gz', FUN=gzfile, remove=FALSE)
        DNAfastaFile = destName;
        
    }
    # using the double quotation to address the problem associate with blank
    # spaces in the command.
    command = paste('"', perlExec, '" "', perlScript_getDNAseq, '" "', 
                    filenameAllCDS, '" "', DNAfastaFile, '" "', 
                    filenameAllCDSdnaSeq, '"', sep='')
    
    system(command)
    
    if(grepl('\\.gz$', DNAfastaFile00)[1]) { # if uncompressed the fasta file 
        if (file.exists(DNAfastaFile)) { 
            invisible(file.remove(DNAfastaFile)) 
        }
        DNAfastaFile= DNAfastaFile00
    }
    
    # then run the perl script to get the AA sequences from the DNA sequence 
    # of a transcipts from its CDS.
    
    filenameAllCDSdnaSeqAA = paste(outputFolder, '/', 
                                    filenamePrefix, '_AAseq.txt', sep='');
    
    command = paste('"', perlExec, '" "', perlScript_getAAseq, '" "', 
        filenameAllCDSdnaSeq, '" "', geneticCodeFile_line, '" "', 
        filenameAllCDSdnaSeqAA, '"', sep='')
    
    system(command)

    resultsNow = read.table(filenameAllCDSdnaSeqAA, stringsAsFactors=FALSE,
                        sep='\t')
    
    kk = pIDgene[resultsNow[,5],]
    rownames(kk) = NULL
    
    # append the protein ID and gene symbol to the table.
    resultsNow = cbind(resultsNow, kk) 
    
    write.table(resultsNow, file=filenameAllCDSdnaSeqAA, sep='\t', 
                    row.names=FALSE, col.names=FALSE, quote=FALSE)
    
    R.utils::gzip(filenameAllCDSdnaSeqAA, overwrite=TRUE) # compress the file
    
    if (file.exists(filenameAllCDSdnaSeq)) { 
        invisible(file.remove(filenameAllCDSdnaSeq)) 
    }

    if ( file.exists(filenameAllCDS) ) { 
        invisible(file.remove(filenameAllCDS)) 
    }
}