R/combineHapState.R

In ruqianl/comapr: Crossover analysis and genetic map construction

#' combineHapState
#'
#' combine two `RangedSummarizedExperiment` objects, each contains the haplotype
#' state for a list of SNPs across a set of cells. The combined result will have
#' cells from two individuals and merged list of SNPs from the two.
#'
#' @importFrom  Matrix Matrix
#' @importFrom IRanges findOverlaps IRanges
#' @importFrom SummarizedExperiment rowRanges assay assay<- rowRanges<-
#' @importFrom SummarizedExperiment assays assays<-
#' @importFrom S4Vectors merge from to 
#' @param rse1, the first `RangedSummarizedExperiment`
#' @param rse2, the second `RangedSummarizedExperiment`
#' @param groupName, a character vector of length 2 that contains the first and
#' the second group's names
#'
#' @examples
#' BiocParallel::register(BiocParallel::SnowParam(workers = 1))
#' demo_path <- paste0(system.file("extdata",package = "comapr"),"/")
#' s1_rse_state <- readHapState("s1",chroms=c("chr1"),
#'                              path=demo_path,barcodeFile=NULL,minSNP = 0,
#'                              minlogllRatio = 50,
#'                              bpDist = 100,maxRawCO=10,
#'                              minCellSNP = 1)
#'
#' s2_rse_state <- readHapState("s2",chroms=c("chr1"),
#'                              path=demo_path,
#'                              barcodeFile=paste0(demo_path,"s2_barcodes.txt"),
#'                              minSNP = 0,
#'                              minlogllRatio = 50,
#'                              bpDist = 100,maxRawCO=10,
#'                              minCellSNP = 1)
#' sb <- combineHapState(s1_rse_state,s2_rse_state)
#'
#' @return A `RangedSummarizedExperiment` that contains the cells and SNPs in
#' both `rse`
#'
#' @export
#'
#' @author Ruqian Lyu
#'
combineHapState <- function(rse1,rse2,
                             groupName=c("Sample1","Sample2")){

  stopifnot(length(groupName)==2)

  merged_anno <- S4Vectors::merge(rowRanges(rse1),rowRanges(rse2),all=TRUE)

  stopifnot(!is.data.frame(merged_anno))
  t <- Matrix(data=0,nrow=length(merged_anno),ncol=(ncol(rse1)+ncol(rse2)),
              sparse = TRUE)
  se1_rows <- findOverlaps(merged_anno,rowRanges(rse1),select = "all")
  se1_rows <- from(se1_rows)
  t[se1_rows,seq_len(ncol(rse1))] <- assays(rse1)[["vi_state"]]
  se2_rows <- findOverlaps(merged_anno,rowRanges(rse2),select = "all")
  se2_rows <- from(se2_rows)
  t[se2_rows,(ncol(rse1)+1):ncol(t)] <- assay(rse2)
  combinedCols <- rbind(colData(rse1),colData(rse2))


  SummarizedExperiment(assays=list(vi_state=t),
                         colData = combinedCols,
                         rowRanges = merged_anno)



}