R/methods-AssayData.R

In MinimumDistance: A Package for De Novo CNV Detection in Case-Parent Trios

AssayDataList <- function(storage.mode = c("lockedEnvironment", "environment", "list"), ...) {
  storage.mode <- match.arg(storage.mode) ## defaults to "lockedEnvironment"
  assayData <- switch(storage.mode,
                      lockedEnvironment =,
                      environment = new.env(parent=emptyenv()),
                      list = list())
  arglist <- list(...)
  for (nm in names(arglist)) assayData[[nm]] <- arglist[[nm]]
  ##if (storage.mode == "lockedEnvironment") Biobase:::assayDataEnvLock(assayData)
  storageMode(assayData) <- storage.mode
  assayData
}

assayDataListDims <- function(object) {
  nms <- if (storageMode(object) == "list") names(object) else ls(object)
  if (length(nms) == 0)
    return(matrix(integer(0), nrow = 2, ncol = 0,
                  dimnames = list(c("Features", "Samples"), character(0))))
  d <- lapply(nms, function(i) lapply(object[[i]], dim)) ##dim(object[[i]]))
  ##rownames(d) <- c("Features", "Samples", rep("...", nrow(d)-2))
  names(d) <- nms
  ##colnames(d) <- nms
  ##d[,order(colnames(d)), drop=FALSE]
  return(d)
}


validAssayDataDims <- function(object){
	msg <- NULL
	d <- assayDataListDims(object)
	firstElement <- d[[1]]
	d <- d[-1]
	res <- sapply(d, function(i) identical(i, firstElement))
	## check that the 3rd dimension is 3
	if(!all(res)){
		msg <- "Assay data elements must have the same dimension"
	}
	if(firstElement[[1]][3] != 3){
		msg <- c(msg, "third dimension of assayData elements must be 3")
	}
	if(is.null(msg)) return(TRUE) else return(msg)
}

##assayDataListLD <- function(path="", ext="", pedigree, featureData, ffprefix=""){
##  filenames <- paste(originalNames(allNames(pedigree)), ext, sep="")
##  fnames <- file.path(path, filenames)
##  stopifnot(all(file.exists(fnames)))
##  if(missing(featureData)) stop("featureData can not be missing")
##  if(missing(pedigree)) stop("pedigree can not be missing")
##  index <- split(seq_len(nrow(featureData)), chromosome(featureData))
##  useff <- isPackageLoaded("ff")
##  if(useff){
##    message("Initializing ff arrays for BAFs and LRRs  ...")
##  } else {
##    message("ff package not loaded -- initializing arrays for BAFs and LRRs ...")
##  }
##  pkgs <- if(useff) c("ff", neededPkgs()) else neededPkgs()
##  outdir <- ldPath()
##  i <- NULL
##  bafAndLrrList <- foreach(i=index, .packages=pkgs) %dopar% {
##    initializeLrrAndBafArrays(dims=c(length(i), nrow(pedigree), 3),
##                              outdir=outdir,
##                              col.names=sampleNames(pedigree),
##                              name=ffprefix)
##  }
##  baflist <- lapply(bafAndLrrList, "[[", 1)
##  lrrlist <- lapply(bafAndLrrList, "[[", 2)
##  ##	} else {
##  ##		baflist <- lapply(index, function(x) initializeBigArray("baf", dim=c(length(x), nrow(pedigree), 3), vmode="integer"))
##  ##		lrrlist <- lapply(index, function(x) initializeBigArray("lrr", dim=c(length(x), nrow(pedigree), 3), vmode="integer"))
##  ##		baflist <- lapply(baflist, function(x, sampleNames) {
##  ##			colnames(x) <- sampleNames
##  ##			return(x)
##  ##		}, sampleNames=sampleNames(pedigree))
##  ##		lrrlist <- lapply(lrrlist, function(x, sampleNames){
##  ##			colnames(x) <- sampleNames
##  ##			return(x)
##  ##		}, sampleNames=sampleNames(pedigree))
##  ##	}
##  if(useff){
##    message("Reading ", length(fnames),
##            " files and writing ff files to directory ", ldPath())
##  }
##  fathers <- paste(fatherNames(pedigree), ext, sep="")
##  mothers <- paste(motherNames(pedigree), ext, sep="")
##  offsprg <- paste(offspringNames(pedigree), ext, sep="")
##  trioindex <- seq_len(nrow(pedigree))
##  ## want to avoid reading in more than 10 files per node -- would
##  ## require a lot of total ram, depending on how many nodes are avail.
##  ##
##  ## for reading in data, we can't split by chromosome (all
##  ## markers are read in at once) So, we split by samples.
##  if(parStatus()){
##    ## e.g., for 900 trios and 3 workers,
##    ## each worker reads in 300 trios
##    ##  --- below we read 100 fathers, 100 mothers, 100 offspring...
##    ##ilist <- splitIndicesByLength(trioindex, getCluster())
##    ilist <- splitIndicesByNode(trioindex)
##    ilist <- ilist[sapply(ilist, function(x) length(x) > 0)]
##  } else {
##    ## execution is sequential.
##    ilist <- list(trioindex)
##  }
##  if(useff){
##    ## pass the ff object / array to each worker
##    ## read in the files and assign the results to column z
##    ## workers read in different sets of files and assign to the baflist and lrrlist ff objects
##    ## read one file.
##    if(!getDoParWorkers()) registerDoSEQ()
##    fns <- featureNames(featureData)
##    res <- foreach(i=ilist, .packages=pkgs) %dopar% {
##      read.bsfiles2(path=path,
##                           filenames=originalNames(fathers[i]),
##                           sampleNames=sampleNames(pedigree)[i],
##                           marker.index=index,
##                           z=1,
##                           baflist=baflist,
##                           lrrlist=lrrlist,
##                           featureNames=fns)
##    }
##    res <- foreach(i=ilist, .packages=pkgs) %dopar% {
##      read.bsfiles2(path=path,
##                    filenames=originalNames(mothers[i]),
##                    sampleNames=sampleNames(pedigree)[i],
##                    marker.index=index,
##                    z=2,
##                    baflist=baflist,
##                    lrrlist=lrrlist,
##                    featureNames=fns)
##    }
##    res <- foreach(i=ilist, .packages=pkgs) %dopar% {
##      read.bsfiles2(path=path,
##                    filenames=offsprg[i],
##                    sampleNames=sampleNames(pedigree)[i],
##                    marker.index=index,
##                    z=3,
##                    baflist=baflist,
##                    lrrlist=lrrlist,
##                    featureNames=fns)
##    }
##    message("Finished reading/writing processed data.")
##    gc()
##  } else {
##    F <- read.bsfiles2(path=path, filenames=originalNames(fathers), sampleNames=sampleNames(pedigree))
##    M <- read.bsfiles2(path=path, filenames=originalNames(mothers), sampleNames=sampleNames(pedigree))
##    O <- read.bsfiles2(path=path, filenames=offsprg, sampleNames=sampleNames(pedigree))
##    ## the featureData is ordered by chromosome and physical position
##    ##		.i <- rownames(F) %in% featureNames(featureData)
##    ##		F <- F[.i,,,drop=FALSE]
##    ##		M <- M[.i,,,drop=FALSE]
##    ##		O <- O[.i,,,drop=FALSE]
##    mindex <- match(featureNames(featureData), rownames(F))
##    F <- F[mindex, , , drop=FALSE]
##    M <- M[mindex, , , drop=FALSE]
##    O <- O[mindex, , , drop=FALSE]
##    ##featureData <- featureData[mindex, ]
##    ##if(!identical(featureNames(featureData), rownames(F))) stop("feature names in featureData do not match rownames in assay data array")
##    ##index <- split(seq_len(nrow(featureData)), chromosome(featureData))
##    for(j in seq_along(index)){
##      k <- index[[j]]
##      baflist[[j]][, , 1] <- integerMatrix(F[k, 2, ], scale=1)
##      baflist[[j]][, , 2] <- integerMatrix(M[k, 2, ], scale=1)
##      baflist[[j]][, , 3] <- integerMatrix(O[k, 2, ], scale=1)
##      lrrlist[[j]][, , 1] <- integerMatrix(F[k, 1, ], scale=1)
##      lrrlist[[j]][, , 2] <- integerMatrix(M[k, 1, ], scale=1)
##      lrrlist[[j]][, , 3] <- integerMatrix(O[k, 1, ], scale=1)
##    }
##  }
##  ad <- AssayDataList(logRRatio=lrrlist,
##                      BAF=baflist)
##  return(ad)
##}