R/splitBAM_byRepindex.R

In vivekbhr/icetea: Integrating Cap Enrichment with Transcript Expression Analysis

#' Split the composite BAM file using replicate indexes (MAPCap data)
#'
#' @param bamFile character. Path to a mapped BAM file
#' @param outfile_prefix character. prefix for output file (replicates IDs will be added as RR/YY)
#' @param ncores integer. Number of cores to use for parallel processing
#'
#' @return Filtered files by replicate Index
#' @export
#'
#' @examples
#'
#' bam <- system.file("extdata", "bam/embryo1.bam", package = "icetea")
#' splitBAM_byRepindex(bamFile = bam, outfile_prefix = "testSplit", ncores = 1)
#'

splitBAM_byRepindex <-
    function(bamFile, outfile_prefix, ncores = 1) {
        ## Write a closure that return the function to search idx in readname
        message("Creating Filtering Rules")
        make_FilterFunc <- function(rep_name) {
            function(df) {
                df$qname <- as.character(df$qname)
                df_sep <- data.frame(
                    idx = vapply(strsplit(df$qname, "#"), "[[",
                                    character(1), 2),
                    stringsAsFactors = FALSE
                )
                df_sep3 <-
                    data.frame(
                        idx = vapply(strsplit(df_sep$idx, ":"), "[[",
                                        character(1), 3),
                        stringsAsFactors = FALSE
                    )

                return(grepl(rep_name, df_sep3$idx))
            }

        }

        ## A function to create filterRules argument
        make_FilterRules <- function(FilterFunc) {
            return(S4Vectors::FilterRules(list(FilterFunc)))
        }

        ## Splitting by PURINES (RR) or PYRIMIDINES (YY)
        repindex_list <-
            list(
                RR = paste("AA", "GG", "GA", "AG", sep = "|"),
                YY = paste("CC", "TT", "CT", "TC", sep = "|")
            )
        ## Now put them together, get lists back

        filtfuncs <- lapply(repindex_list, make_FilterFunc)
        filtrules <- lapply(filtfuncs, make_FilterRules)

        ## Filter the files in parallel
        message("Filtering the BAM file")
        destinations <-
            paste0(outfile_prefix, "_", names(repindex_list), ".bam")

        param <- getMCparams(ncores)
        # register parallel backend
        if (!BiocParallel::bpisup(param)) {
            BiocParallel::bpstart(param)
            on.exit(BiocParallel::bpstop(param))
        }
        BiocParallel::bplapply(seq_along(destinations),
                               function(i, file, destinations, filtrules) {
                                   Rsamtools::filterBam(file, destinations[i],
                                                        filter = filtrules[[i]])
                               }, bamFile, destinations, filtrules,
                               BPPARAM = param)

        ## Files written
        message("Done!")
    }