SparseSignatures: SparseSignatures

# Import point mutations data to build the count matrix to extract mutatational signatues.
#
### @examples
### data(ssm560_reduced)
### library("BSgenome.Hsapiens.1000genomes.hs37d5")
### bsg = BSgenome.Hsapiens.1000genomes.hs37d5
### data(mutation_categories)
### imported_data = import.counts.data(input=ssm560_reduced,bsg=bsg,mutation_categories=mutation_categories)
### data(imported_data)
### head(imported_data)
#
# @title import.counts.data
# @param input either a data.frame/data.table object or a file with 5 columns: sample name, chromosome, position, ref, alt.
# @param bsg a BSgenome object for the reference genome. Chromosome names have to match the input table.
# @param mutation_categories array with the 96 mutational categories to be considered. It is provided along with the package 
# by data(mutation_categories).
# @return A count matrix to extract mutatational signatues
# @export import.counts.data
# @importFrom data.table data.table as.data.table fread dcast .N
# @importFrom Biostrings DNAStringSet complement reverseComplement subseq
# @import GenomicRanges
# @import IRanges
# @import GenomeInfoDb
# @import BSgenome.Hsapiens.1000genomes.hs37d5
# @importFrom BSgenome getSeq
#
"import.counts.data" <- function(input, bsg, mutation_categories) {

    .Deprecated("import.trinucleotides.counts")

#  # check that input is a data frame or data table
#  if (!("data.frame" %in% class(input))) {
#    if (file.exists(input)) {
#      input <- fread(input, header = TRUE, check.names = FALSE)
#    } else {
#      stop("Input is neither a file nor a loaded data frame.")
#    }
#  }
#  
#  # set column names
#  colnames(input) <- c("sample", "chrom", "pos", "ref", "alt")
#  
#  # convert input to GRanges
#  inp <- GRanges(input$chrom, IRanges(start=input$pos-1, width=3), ref=DNAStringSet(input$ref), alt=DNAStringSet(input$alt), sample=input$sample)
#
#  # filter input: recognized bases only
#  bad <- which(!(inp$ref %in% c("A", "T", "C", "G") & inp$alt %in% c("A", "T", "C", "G")))
#  if(length(bad)>0){
#    warning(paste0("Removing ", length(bad), " entries containing nonstandard bases. \n"))
#    inp <- inp[inp$ref %in% c("A", "T", "C", "G") & inp$alt %in% c("A", "T", "C", "G")]
#  }
#  
#  # check that bsg is a bsgenome object
#  if(is.null(bsg)|class(bsg) != "BSgenome") {
#      stop("The bsg parameter needs to be a BSgenome object.")
#  }
#  
#  # check that all chromosomes match bsg
#  if(length(setdiff(seqnames(inp), GenomeInfoDb::seqnames(bsg)))>0) {
#        warning(paste0("Check chromosome names -- not all match ", bsg, " object.\n"))
#        inp[seqnames(inp) %in% GenomeInfoDb::seqnames(bsg)]
#  }
#
#  # find context for each mutation
#  inp$context <- getSeq(bsg, inp)
#
#  # check for mismatches with BSgenome context
#  if(any(subseq(inp$context,2,2)!=inp$ref)) {
#    warning("Check ref bases -- not all match context.\n ")
#  }
#  
#  # get complements and reverse complements
#  inp$cref <- complement(inp$ref)
#  inp$calt <- complement(inp$alt)
#  inp$rccontext <- reverseComplement(inp$context)
#  
#  # identify motif
#  inp$cat <- ifelse(inp$ref %in% c("C", "T"), 
#                     paste0(subseq(inp$context,1,1), "[", inp$ref, ">", inp$alt, "]", subseq(inp$context, 3, 3)),
#                     paste0(subseq(inp$rccontext,1,1), "[", inp$cref, ">", inp$calt, "]", subseq(inp$rccontext, 3, 3)))
#  
#  # count number of mutations per sample, category
#  counts <- merge(mutation_categories[, .(cat)], data.table(sample=inp$sample, cat=inp$cat)[, .N, by=.(sample, cat)], by="cat", all=TRUE)
#  counts <- dcast(counts, sample~cat, value.var = "N")
#  counts <- counts[!is.na(sample)]
#  counts[is.na(counts)] <- 0
#  
#  # make count marix
#  countMatrix <- as.matrix(counts[, 2:ncol(counts)])
#  rownames(countMatrix) <- counts$sample
#  
#  # order matrix columns alphabetically
#  countMatrix <- countMatrix[,order(colnames(countMatrix))]
#  
#  # identify samples with <100 mutations
#  if(!is.null(nrow(countMatrix))){
#    bad <- names(rowSums(countMatrix) <= 100)
#  }else{
#    bad <- (sum(countMatrix) <= 100)
#  }
#  if (length(bad) > 0) {
#    warning(paste("Some samples have fewer than 100 mutations:\n ", paste(bad, collapse = ", "), sep = " "))
#  }
#  
#  # return matrix
#  return(countMatrix)
#

}