Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feature: updates for cellranger 9.0 #9541

Open
wants to merge 3 commits into
base: develop
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
132 changes: 132 additions & 0 deletions R/preprocessing.R
Original file line number Diff line number Diff line change
Expand Up @@ -645,6 +645,138 @@ Load10X_Spatial <- function (

return(object)
}
#' Add 10X Cell Types to a Seurat Object
#'
#' This function reads cell type annotations from a CSV file and adds them to the metadata of a Seurat object.
#' If the cell type file does not exist, the original Seurat object is returned unchanged.
#'
#' @param data.dir A string specifying the directory containing the "cell_types" folder with the "cell_types.csv" file.
#' @param object A Seurat object to which the cell type annotations will be added.
#'
#' @return A Seurat object with updated metadata including cell type annotations if the file is found.
#'
#' @details
#' The function searches for a CSV file named "cell_types.csv" in the "cell_types" subdirectory within `data.dir`.
#' The CSV file should contain at least a "barcode" column that matches the cell barcodes in the Seurat object.
#' Additional columns in the CSV file will be merged into the Seurat object's metadata.
#'
#' @importFrom utils read.csv
#' @importFrom tibble rownames_to_column column_to_rownames
#' @importFrom base file.path file.exists merge
#'
#' @examples
#' \dontrun{
#' # Specify the data directory containing the "cell_types" folder
#' data.dir <- "/path/to/data"
#'
#' # Create a Seurat object (example)
#' seurat_obj <- CreateSeuratObject(counts = some_counts_matrix)
#'
#' # Add cell type annotations to the Seurat object
#' seurat_obj <- Add_10X_CellTypes(data.dir, seurat_obj)
#' }

Add_10X_CellTypes <- function(data.dir, object) {
cell_types_path <- file.path(data.dir, "cell_types", "cell_types.csv")
if (file.exists(cell_types_path)) {
cell.types <- read.csv(cell_types_path)
meta_data_with_barcodes <- tibble::rownames_to_column([email protected], "barcode")
merged_meta_data <- merge(
x = meta_data_with_barcodes,
y = cell.types,
by = "barcode",
all.x = TRUE
)
[email protected] <- tibble::column_to_rownames(merged_meta_data, "barcode")
return(object)
} else {
return(object)
}
}

#' Load a 10x Genomics Single Cell Experiment into a \code{Seurat} object
#'
#' @inheritParams Read10X
#' @inheritParams SeuratObject::CreateSeuratObject If multiome 10x data the
#' assay param will not be used. The names of each assay contained in the matrix are used.
#' @param data.dir Directory containing the H5 file specified by \code{filename}
#' @param filename Name of H5 file containing the feature barcode matrix
#' @param to.upper Converts all feature names to upper case. This can provide an
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(Preemptive-- this parameter doesn't seem to be implemented yet)
In my opinion, it would be better not to suggest that blind uppercasing is ever an appropriate way of finding mouse-human orthologs, even in an exploratory analysis.
If the user has some good reason to convert all the symbols to uppercase, they can simply use toupper and replace the rownames of the object.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This option was put in originally by the seurat team. I'll let them address the usefulness here.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The parameter was introduced well before my time but I tend to agree with @rharao.

#' approximate conversion of mouse to human gene names which can be useful in an
#' explorative analysis. For cross-species comparisons, orthologous genes should
#' be identified across species and used instead.
#' @param ... Arguments passed to \code{\link{Read10X_h5}}
#'
#' @return A \code{Seurat} object
#'
#'
#' @export
#' @concept preprocessing
#'
#' @examples
#' \dontrun{
#' data_dir <- 'path/to/data/directory'
#' list.files(data_dir) # Should show filtered_feature_bc_matrix.h5
#' Load10X(data.dir = data_dir)
#' }
#'
Load10X <- function(data.dir, filename = "filtered_feature_bc_matrix.h5",
assay = "RNA", to.upper = FALSE, ...) {

if (length(data.dir) > 1) {
stop("`data.dir` expects a single directory path but received multiple values.")
}
if (!file.exists(data.dir)) {
stop("No such file or directory: '", data.dir, "'")
}


filename <- list.files(data.dir, filename, full.names = FALSE, recursive = FALSE)
counts.path <- file.path(data.dir, filename)
if (!file.exists(counts.path)) {
stop("File not found: '", counts.path, "'")
}

counts <- Read10X_h5(counts.path, ...)

if (to.upper) {
counts <- imap(counts, ~{
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nit: because this formula operates on the counts matrices only and doesn't use the list's names, I think the imap call could be replaced by base::sapply.
Frankly, I'm not sure why this package imports purrr::imap anyway, as I can't find it used anywhere in the codebase, not even in Load10X_Spatial. But that's an issue for another time.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

personally, I like the simplicity and readability of tidy functions. I'll leave it up to the seruat developers to decide if they mind the extra import. I haven't heard any complaints since we wrote this function a few years ago.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's not really an issue; the choice one way or another doesn't affect me. (Personally, I agree; I always prefer the purrr functions in my own scripts.) I only wanted to mention it for the attention of the dev team.
Thanks for your receptiveness 😄

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 to replacing the imap call with a sapply to match the current implementation of Load10X_Spatial.

Frankly, I'm not sure why this package imports purrr::imap anyway, as I can't find it used anywhere in the codebase, not even in Load10X_Spatial. But that's an issue for another time.

This is also a good point, @importFrom purrr imap should be dropped from Load10X_Spatial and as a suggested dependency. Now that you mention it, @importFrom grid rasterGrob could also be dropped from the function. I'll include these tweaks in #9556 👌

rownames(.x) <- toupper(rownames(.x))
.x
})
}

if (is.list(counts)) {
seurat.list <- lapply(names(counts), function(name) {
CreateSeuratObject(
counts = counts[[name]],
assay = name,
project = name
)
})

for (i in seq_along(seurat.list)) {
if (Assays(seurat.list[[i]]) %in% c("Gene Expression", "RNA")) {
seurat.list[[i]] <- Add_10X_CellTypes(data.dir, seurat.list[[i]])
}
}

merged.object <- merge(
x = seurat.list[[1]],
y = seurat.list[-1],
add.cell.ids = names(counts),
merge.data = FALSE
)
return(merged.object)

} else {
object <- CreateSeuratObject(counts, assay = assay)
if (Assays(object) %in% c("Gene Expression", "RNA")) {
object <- Add_10X_CellTypes(data.dir, object)
}
Comment on lines +772 to +776
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not sure that is.list(counts) conditional is necessary. Instead, you can just force counts to always be a list, and then you can do without the else block.

It looks like there's just one CellTypes file per data.dir? If that's the case, this could be simplified even more—the call to Add_10X_CellTypes could be made on the merged.object.

return(object)
}
}


#' Read10x Probe Metadata
Expand Down
Loading