Load 10X h5 count matrices from multiple directories — Read10X_h5_Multi

Enables easy loading of sparse data matrices provided by 10X genomics that are present in multiple subdirectories. Can function with either default output directory structure of Cell Ranger or custom directory structure.

Read10X_h5_Multi_Directory(
  base_path,
  secondary_path = NULL,
  default_10X_path = TRUE,
  cellranger_multi = FALSE,
  h5_filename = "filtered_feature_bc_matrix.h5",
  sample_list = NULL,
  sample_names = NULL,
  replace_suffix = FALSE,
  new_suffix_list = NULL,
  parallel = FALSE,
  num_cores = NULL,
  merge = FALSE,
  ...
)

Arguments

base_path: path to the parent directory which contains all of the subdirectories of interest.
secondary_path: path from the parent directory to count matrix files for each sample.
default_10X_path: logical (default TRUE) sets the secondary path variable to the default 10X directory structure.
cellranger_multi: logical, whether samples were processed with Cell Ranger multi, default is FALSE.
h5_filename: name of h5 file (including .h5 suffix). If all h5 files have same name (i.e. Cell Ranger output) then use full file name. By default function uses Cell Ranger name: "filtered_feature_bc_matrix.h5". If h5 files have sample specific prefixes (i.e. from Cell Bender) then use only the shared part of file name (e.g., "_filtered_out.h5").
sample_list: a vector of sample directory names if only specific samples are desired. If NULL will read in subdirectories in parent directory.
sample_names: a set of sample names to use for each sample entry in returned list. If NULL will set names to the subdirectory name of each sample.
replace_suffix: logical (default FALSE). Whether or not to replace the barcode suffixes of matrices using Replace_Suffix.
new_suffix_list: a vector of new suffixes to replace existing suffixes if replace_suffix = TRUE. See Replace_Suffix for more information. To remove all suffixes set new_suffix_list = "".
parallel: logical (default FALSE) whether or not to use multi core processing to read in matrices.
num_cores: how many cores to use for parallel processing.
merge: logical (default FALSE) whether or not to merge samples into a single matrix or return list of matrices. If TRUE each sample entry in list will have cell barcode prefix added. The prefix will be taken from sample_names.
...: Extra parameters passed to Read10X_h5.

Value

a list of sparse matrices (merge = FALSE) or a single sparse matrix (merge = TRUE).

Examples

if (FALSE) {
base_path <- 'path/to/data/directory'
expression_matrices <- Read10X_h5_Multi_Directory(base_path = base_path)
}