@@ -36,4 +36,5 @@ Collate:

     'TENxFile-class.R'

     'TENxFileList-class.R'

     'TENxFile-methods.R'

+    'TENxH5-class.R'

     'utils.R'

@@ -23,13 +23,6 @@

 S4Vectors::setValidity2("TENxFile", .validTENxFile)

-#' @exportClass TENxH5

-.TENxH5 <- setClass(

-    Class = "TENxH5",

-    contains = "TENxFile",

-    slots = c(version = "character", group = "character")

-)

-

 .TENxMTX <- setClass(

     Class = "TENxMTX",

     contains = "TENxFile"

@@ -70,36 +63,3 @@ TENxFile <- function(resource, ...) {

     TENxFUN(resource = resource,  extension = ext, ...)

 }

-.get_h5_group <- function(fpath) {

-    if (!requireNamespace("rhdf5", quietly = TRUE))

-        stop("Install 'rhdf5' to work with TENxH5")

-    l1 <- rhdf5::h5ls(fpath, recursive = FALSE)

-    l1[l1$otype == "H5I_GROUP", "name"]

-}

-

-.KNOWN_H5_GROUPS <- c("matrix", "outs")

-

-.check_group_ok <- function(fpath) {

-    gname <- .get_h5_group(fpath)

-    if (!gname %in% .KNOWN_H5_GROUPS)

-        stop("'group' not recognized")

-    gname

-}

-

-#' @rdname TENxH5

-#' @title Import H5 files from 10X

-#'

-#' @examples

-#'

-#' h5f <- "~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_filtered_feature_bc_matrix.h5"

-#' con <- TENxFile(h5f)

-#' import(con)

-#'

-#' @export

-TENxH5 <-

-    function(resource, version = c("3", "2"), ...)

-{

-    version <- match.arg(version)

-    group <- .check_group_ok(resource)

-    .TENxH5(resource = resource, group = group, version = version, ...)

-}

@@ -1,215 +1,6 @@

-#' @importFrom MatrixGenerics rowRanges

 #' @include TENxFileList-class.R

 NULL

-gene.meta <- data.frame(

-    version = c("3", "2"),

-    ID = c("/features/id", "/genes"),

-    Symbol = c("/features/name", "/gene_names"),

-    Type = c("/features/feature_type", NA_character_)

-)

-

-.getRowDat <- function(con) {

-    gm <- gene.meta[

-        con@version == gene.meta[["version"]],

-        !names(gene.meta) %in% c("group", "version")

-    ]

-    gm[] <- Filter(Negate(is.na), gm)

-    res <- lapply(gm, function(colval) {

-        readname <- paste0(con@group, colval)

-        rhdf5::h5read(path(con), readname)

-    })

-    as.data.frame(res)

-}

-

-.getGenome <- function(con) {

-    gloc <- "matrix/features/genome"

-    gen <- unique(rhdf5::h5read(path(con), gloc))

-    if (length(gen) != 1L)

-        stop("The genome build in ", gloc, " is not consistent")

-    gen

-}

-

-#' @importFrom GenomeInfoDb genome genome<-

-#' @importFrom S4Vectors mcols<-

-.getRowRanges <- function(con) {

-    interval <- rhdf5::h5read(path(con), "matrix/features/interval")

-    interval[interval == "NA"] <- "NA_character_:0"

-    gr <- as(as.character(interval), "GRanges")

-    mcols(gr) <- .getRowDat(con)

-    genbuild <- rep(.getGenome(con), length(genome(gr)))

-    genome(gr) <- genbuild

-    gr

-}

-

-#' @import SingleCellExperiment

-#' @export

-setMethod("import", "TENxH5", function(con, format, text, ...) {

-    if (!requireNamespace("HDF5Array", quietly = TRUE))

-        stop("Install 'HDF5Array' to import TENxH5 files")

-    matrixdata <- HDF5Array::TENxMatrix(path(con), con@group)

-    if (identical(con@version, "3")) {

-        sce <- SingleCellExperiment::SingleCellExperiment(

-            assays = list(counts = matrixdata), rowRanges = .getRowRanges(con)

-        )

-        rownames(sce) <- mcols(sce)[["ID"]]

-        splitAltExps(sce, rowData(sce)[["Type"]], ref = "Gene Expression")

-    } else {

-        stop("Version 2 not supported yet.")

-    }

-})

-

-#' @import SummarizedExperiment

-#' @export

-setMethod("import", "TENxMTX", function(con, format, text, ...) {

-    mtxf <- SingleCellMultiModal:::.read_mtx(path(con))

-    ## TODO: make use of other files

-    SummarizedExperiment::SummarizedExperiment(

-        assays = SimpleList(counts = mtxf)

-    )

-})

-

-

-#' @importFrom utils untar tail

-.TENxUntar <- function(con) {

-    dir.create(tempdir <- tempfile())

-    untar(path(con), exdir = tempdir)

-    tempdir

-}

-

-.readInFuns <- function(files) {

-    file_exts <- .get_ext(files)

-    lapply(.setNames(file_exts, basename(files)), function(ext) {

-        switch(

-            ext,

-            mtx.gz = Matrix::readMM,

-            tsv.gz = function(...)

-                readr::read_tsv(col_names = FALSE, show_col_types = FALSE, ...)

-        )

-    })

-}

-

-.cleanUpFuns <- function(datalist) {

-    if (is.null(names(datalist)))

-        stop("'datalist' names must correspond to originating file names")

-    lapply(.setNames(nm = names(datalist)), function(fname) {

-        switch(

-            fname,

-            features.tsv.gz = function(df) {

-                names(df) <- c("ID", "Symbol", "Type", "Chr", "Start", "End")

-                df

-            },

-            barcodes.tsv.gz = function(df) {

-                names(df) <- "barcode"

-                df

-            },

-            matrix.mtx.gz = function(mat) {

-                as(mat, "dgCMatrix")

-            },

-        )

-    })

-}

-

-.TENxDecompress <- function(con) {

-    res_ext <- .get_ext(path(con))

-    if (identical(res_ext, "tar.gz")) {

-        tenfolder <- .TENxUntar(con)

-        gfolder <- list.files(tenfolder, full.names = TRUE)

-        if (file.info(gfolder)$isdir)

-            gfiles <- list.files(gfolder, recursive = TRUE, full.names = TRUE)

-        else

-            gfiles <- gfolder

-        gdata <- Map(

-            f = function(reader, x) {

-                reader(x)

-            }, reader = .readInFuns(gfiles), x = gfiles

-        )

-        Map(f = function(cleaner, x) {

-                cleaner(x)

-            }, cleaner = .cleanUpFuns(gdata), x = gdata

-        )

-    } else {

-        stop("Extension type: ", res_ext, " not supported")

-    }

-}

-

-#' @export

-setMethod("import", "TENxFileList", function(con, format, text, ...) {

-    if (con@compressed)

-        fdata <- .TENxDecompress(con)

-    else

-        fdata <- con@listData

-    mat <- fdata[["matrix.mtx.gz"]]

-    colnames(mat) <- unlist(fdata[["barcodes.tsv.gz"]])

-    warning("Matrix of mixed types; see in rowData(x)")

-    SingleCellExperiment::SingleCellExperiment(

-        SimpleList(counts = mat),

-        rowData = fdata[["features.tsv.gz"]]

-    )

-})

-

-#' @importFrom MatrixGenerics rowRanges

-#' @include TENxFileList-class.R

-NULL

-

-gene.meta <- data.frame(

-    version = c("3", "2"),

-    ID = c("/features/id", "/genes"),

-    Symbol = c("/features/name", "/gene_names"),

-    Type = c("/features/feature_type", NA_character_)

-)

-

-.getRowDat <- function(con) {

-    gm <- gene.meta[

-        con@version == gene.meta[["version"]],

-        !names(gene.meta) %in% c("group", "version")

-    ]

-    gm[] <- Filter(Negate(is.na), gm)

-    res <- lapply(gm, function(colval) {

-        readname <- paste0(con@group, colval)

-        rhdf5::h5read(path(con), readname)

-    })

-    as.data.frame(res)

-}

-

-.getGenome <- function(con) {

-    gloc <- "matrix/features/genome"

-    gen <- unique(rhdf5::h5read(path(con), gloc))

-    if (length(gen) != 1L)

-        stop("The genome build in ", gloc, " is not consistent")

-    gen

-}

-

-#' @importFrom GenomeInfoDb genome genome<-

-#' @importFrom S4Vectors mcols<-

-.getRowRanges <- function(con) {

-    interval <- rhdf5::h5read(path(con), "matrix/features/interval")

-    interval[interval == "NA"] <- "NA_character_:0"

-    gr <- as(as.character(interval), "GRanges")

-    mcols(gr) <- .getRowDat(con)

-    genbuild <- rep(.getGenome(con), length(genome(gr)))

-    genome(gr) <- genbuild

-    gr

-}

-

-#' @import SingleCellExperiment

-#' @export

-setMethod("import", "TENxH5", function(con, format, text, ...) {

-    if (!requireNamespace("HDF5Array", quietly = TRUE))

-        stop("Install 'HDF5Array' to import TENxH5 files")

-    matrixdata <- HDF5Array::TENxMatrix(path(con), con@group)

-    if (identical(con@version, "3")) {

-        sce <- SingleCellExperiment::SingleCellExperiment(

-            assays = list(counts = matrixdata), rowRanges = .getRowRanges(con)

-        )

-        rownames(sce) <- mcols(sce)[["ID"]]

-        sce <- sce[seqnames(rowRanges(sce)) != "NA_character_", ]

-        splitAltExps(sce, rowData(sce)[["Type"]], ref = "Gene Expression")

-    } else {

-        stop("Version 2 not supported yet.")

-    }

-})

-

 #' @import SummarizedExperiment

 #' @export

 setMethod("import", "TENxMTX", function(con, format, text, ...) {

@@ -299,20 +90,32 @@ setMethod("import", "TENxFileList", function(con, format, text, ...) {

     )

 })

+#' Importing peak-annotation files

+#'

+#' @keywords internal

+#'

 #' @examples

+#'

 #' fi <- "~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_atac_peak_annotation.tsv"

 #' pa <- .import_peak_anno(fi)

+#'

 .import_peak_anno <- function(file) {

     stopifnot(endsWith(file, "peak_annotation.tsv"))

     panno <- readr::read_tsv(file, col_types = c("c", "n", "n", "c", "n", "c"))

     makeGRangesFromDataFrame(panno, keep.extra.columns = TRUE)

 }

+#' Importing fragment files

+#'

+#' @keywords internal

+#'

 #' @examples

+#'

 #' con <- TENxFile("~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_filtered_feature_bc_matrix.h5")

 #' sce <- import(con)

 #' fr <- "~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_atac_fragments.tsv.gz"

 #' fra <- .import_fragments(fr)

+#'

 .import_fragments <- function(file, withIndex = TRUE) {

     ## read tabix format with Rsamtools

     stopifnot(endsWith(file, "atac_fragments.tsv.gz"))

@@ -328,7 +131,9 @@ setMethod("import", "TENxFileList", function(con, format, text, ...) {

     ## names from table in the link below

     ## https://support.10xgenomics.com/single-cell-atac/software/pipelines/latest/output/fragments

     names(ex) <- c("chrom", "chromStart", "chromEnd", "barcode", "readSupport")

-    ggr <- makeGRangesFromDataFrame(ex, keep.extra.columns = TRUE, starts.in.df.are.0based = TRUE)

+    ggr <- makeGRangesFromDataFrame(

+        ex, keep.extra.columns = TRUE, starts.in.df.are.0based = TRUE

+    )

     ggrl <- as(splitAsList(ggr, mcols(ggr)$barcode), "GRangesList")

     RaggedExperiment(ggrl)

     # compactAssay(erg, i = "readSupport", sparse = TRUE)

@@ -32,3 +32,81 @@ TENxFileList <- function(..., compressed = FALSE) {

     asl <- SimpleList(dots[[1]])

     .TENxFileList(asl, compressed = compressed)

 }

+

+#' @importFrom utils untar tail

+.TENxUntar <- function(con) {

+    dir.create(tempdir <- tempfile())

+    untar(path(con), exdir = tempdir)

+    tempdir

+}

+

+.readInFuns <- function(files) {

+    file_exts <- .get_ext(files)

+    lapply(.setNames(file_exts, basename(files)), function(ext) {

+        switch(

+            ext,

+            mtx.gz = Matrix::readMM,

+            tsv.gz = function(...)

+                readr::read_tsv(col_names = FALSE, show_col_types = FALSE, ...)

+        )

+    })

+}

+

+.cleanUpFuns <- function(datalist) {

+    if (is.null(names(datalist)))

+        stop("'datalist' names must correspond to originating file names")

+    lapply(.setNames(nm = names(datalist)), function(fname) {

+        switch(

+            fname,

+            features.tsv.gz = function(df) {

+                names(df) <- c("ID", "Symbol", "Type", "Chr", "Start", "End")

+                df

+            },

+            barcodes.tsv.gz = function(df) {

+                names(df) <- "barcode"

+                df

+            },

+            matrix.mtx.gz = function(mat) {

+                as(mat, "dgCMatrix")

+            },

+        )

+    })

+}

+

+.TENxDecompress <- function(con) {

+    res_ext <- .get_ext(path(con))

+    if (identical(res_ext, "tar.gz")) {

+        tenfolder <- .TENxUntar(con)

+        gfolder <- list.files(tenfolder, full.names = TRUE)

+        if (file.info(gfolder)$isdir)

+            gfiles <- list.files(gfolder, recursive = TRUE, full.names = TRUE)

+        else

+            gfiles <- gfolder

+        gdata <- Map(

+            f = function(reader, x) {

+                reader(x)

+            }, reader = .readInFuns(gfiles), x = gfiles

+        )

+        Map(f = function(cleaner, x) {

+                cleaner(x)

+            }, cleaner = .cleanUpFuns(gdata), x = gdata

+        )

+    } else {

+        stop("Extension type: ", res_ext, " not supported")

+    }

+}

+

+#' @export

+setMethod("import", "TENxFileList", function(con, format, text, ...) {

+    if (con@compressed)

+        fdata <- .TENxDecompress(con)

+    else

+        fdata <- con@listData

+    mat <- fdata[["matrix.mtx.gz"]]

+    colnames(mat) <- unlist(fdata[["barcodes.tsv.gz"]])

+    warning("Matrix of mixed types; see in rowData(x)")

+    SingleCellExperiment::SingleCellExperiment(

+        SimpleList(counts = mat),

+        rowData = fdata[["features.tsv.gz"]]

+    )

+})

new file mode 100644

@@ -0,0 +1,100 @@

+#' @importFrom MatrixGenerics rowRanges

+NULL

+

+#' @exportClass TENxH5

+.TENxH5 <- setClass(

+    Class = "TENxH5",

+    contains = "TENxFile",

+    slots = c(version = "character", group = "character")

+)

+

+.get_h5_group <- function(fpath) {

+    if (!requireNamespace("rhdf5", quietly = TRUE))

+        stop("Install 'rhdf5' to work with TENxH5")

+    l1 <- rhdf5::h5ls(fpath, recursive = FALSE)

+    l1[l1$otype == "H5I_GROUP", "name"]

+}

+

+.KNOWN_H5_GROUPS <- c("matrix", "outs")

+

+.check_group_ok <- function(fpath) {

+    gname <- .get_h5_group(fpath)

+    if (!gname %in% .KNOWN_H5_GROUPS)

+        stop("'group' not recognized")

+    gname

+}

+

+#' @rdname TENxH5

+#' @title Import H5 files from 10X

+#'

+#' @examples

+#'

+#' h5f <- "~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_filtered_feature_bc_matrix.h5"

+#' con <- TENxFile(h5f)

+#' import(con)

+#'

+#' @export

+TENxH5 <-

+    function(resource, version = c("3", "2"), ...)

+{

+    version <- match.arg(version)

+    group <- .check_group_ok(resource)

+    .TENxH5(resource = resource, group = group, version = version, ...)

+}

+

+gene.meta <- data.frame(

+    version = c("3", "2"),

+    ID = c("/features/id", "/genes"),

+    Symbol = c("/features/name", "/gene_names"),

+    Type = c("/features/feature_type", NA_character_)

+)

+

+.getRowDat <- function(con) {

+    gm <- gene.meta[

+        con@version == gene.meta[["version"]],

+        !names(gene.meta) %in% c("group", "version")

+    ]

+    gm[] <- Filter(Negate(is.na), gm)

+    res <- lapply(gm, function(colval) {

+        readname <- paste0(con@group, colval)

+        rhdf5::h5read(path(con), readname)

+    })

+    as.data.frame(res)

+}

+

+.getGenome <- function(con) {

+    gloc <- "matrix/features/genome"

+    gen <- unique(rhdf5::h5read(path(con), gloc))

+    if (length(gen) != 1L)

+        stop("The genome build in ", gloc, " is not consistent")

+    gen

+}

+

+#' @importFrom GenomeInfoDb genome genome<-

+#' @importFrom S4Vectors mcols<-

+.getRowRanges <- function(con) {

+    interval <- rhdf5::h5read(path(con), "matrix/features/interval")

+    interval[interval == "NA"] <- "NA_character_:0"

+    gr <- as(as.character(interval), "GRanges")

+    mcols(gr) <- .getRowDat(con)

+    genbuild <- rep(.getGenome(con), length(genome(gr)))

+    genome(gr) <- genbuild

+    gr

+}

+

+#' @import SingleCellExperiment

+#' @export

+setMethod("import", "TENxH5", function(con, format, text, ...) {

+    if (!requireNamespace("HDF5Array", quietly = TRUE))

+        stop("Install 'HDF5Array' to import TENxH5 files")

+    matrixdata <- HDF5Array::TENxMatrix(path(con), con@group)

+    if (identical(con@version, "3")) {

+        sce <- SingleCellExperiment::SingleCellExperiment(

+            assays = list(counts = matrixdata), rowRanges = .getRowRanges(con)

+        )

+        rownames(sce) <- mcols(sce)[["ID"]]

+        splitAltExps(sce, rowData(sce)[["Type"]], ref = "Gene Expression")

+    } else {

+        stop("Version 2 not supported yet.")

+    }

+})

@@ -1,5 +1,5 @@

 % Generated by roxygen2: do not edit by hand

-% Please edit documentation in R/TENxFile-class.R

+% Please edit documentation in R/TENxH5-class.R

 \name{TENxH5}

 \alias{TENxH5}

 \title{Import H5 files from 10X}

new file mode 100644

@@ -0,0 +1,20 @@

+% Generated by roxygen2: do not edit by hand

+% Please edit documentation in R/TENxFile-methods.R

+\name{.import_fragments}

+\alias{.import_fragments}

+\title{Importing fragment files}

+\usage{

+.import_fragments(file, withIndex = TRUE)

+}

+\description{

+Importing fragment files

+}

+\examples{

+

+con <- TENxFile("~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_filtered_feature_bc_matrix.h5")

+sce <- import(con)

+fr <- "~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_atac_fragments.tsv.gz"

+fra <- .import_fragments(fr)

+

+}

+\keyword{internal}

new file mode 100644

@@ -0,0 +1,18 @@

+% Generated by roxygen2: do not edit by hand

+% Please edit documentation in R/TENxFile-methods.R

+\name{.import_peak_anno}

+\alias{.import_peak_anno}

+\title{Importing peak-annotation files}

+\usage{

+.import_peak_anno(file)

+}

+\description{

+Importing peak-annotation files

+}

+\examples{

+

+fi <- "~/data/10x/pbmc_3k/pbmc_granulocyte_sorted_3k_atac_peak_annotation.tsv"

+pa <- .import_peak_anno(fi)

+

+}

+\keyword{internal}