Prerequisite

To preprocess scFlowExamples dataset, the following packages should be installed.

if (!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}
BiocManager::install("DropletUtils")

install.packages("ids")

devtools::install_github("NathanSkene/One2One")

devtools::install_github(repo = "hhoeflin/hdf5r")

devtools::install_github(repo = "mojaveazure/loomR", ref = "develop")

Install the dataset repository

devtools::install_github("neurogenomics/scFlowExample")

Setup the dataset

The current dataset uses TEINH15, TEINH19, MGL1, MOL1 cells. For a list of available cell types please visit this link. Use the following codes only if you want to use different cell types.

# create a temporary directory
td <- tempdir()

# create the placeholder file
tf <- tempfile(tmpdir = td, fileext = ".loom")

# download into the placeholder file
download.file("https://storage.googleapis.com/linnarsson-lab-loom/l5_all.loom", tf) # tf="~/l5_all.loom"
unzip(tf)

allExp <- prep_zeisel2018(path = tf)
keptExp <- merge_zeisel_celltypes(allExp, useCells = c("TEINH15", "TEINH19", "MGL1", "MOL1"))
indvExp <- split_celltypes_byIndv(keptExp, joinCells = c("TEINH15", "TEINH19"), 
                                  nCases = 3, jointName = "TEINH")

# Save dataset so that it can be used easily
usethis::use_data(indvExp, overwrite = TRUE)

Downsampling the dataset

The user can downsample the dataset by reducing the cell number and gene number using the following commands. To downsample the dataset use the keptExp object created in the previous step.

To downsample cells only

keptExp_ds <- downsample_cells(keptExp = keptExp,
                               prop_cell = c(0.5,0.5,0.05,0.02))

indvExp_ds <- split_celltypes_byIndv(keptExp_ds, joinCells = c("TEINH15", "TEINH19"), 
                                              nCases = 3, jointName = "TEINH")

# Save dataset so that it can be used easily
usethis::use_data(indvExp_ds, overwrite = TRUE)

To downsample both cells and genes

keptExp_ds_4K <- downsample_cells(keptExp = keptExp,
                                    prop_cell = c(0.5,0.5,0.05,0.02),
                                    n_top_genes = 4000)
indvExp_ds_4K <- split_celltypes_byIndv(keptExp_ds_4K, joinCells = c("TEINH15", "TEINH19"), 
                                                        nCases = 3, jointName = "TEINH")

# Save dataset so that it can be used easily
usethis::use_data(indvExp_ds_4K, overwrite = TRUE)

Save the data to file

You could just run the following codes and continue from here. The following codes will generate scFlowExample dataset in 10x genomics Cellranger output format, a Manifest.txt file containing data path for individual samples and a SampleSheet.tsv containing sample metadata.

library(scFlowExamples)

#To use the full size dataset
data("indvExp", package = "scFlowExamples")  

#To use a downsampled dataset
data("indvExp_ds", package = "scFlowExamples") #This dataset contains all genes (~29000)

#To use a downsampled dataset with 4000 genes
data("indvExp_ds_4K", package = "scFlowExamples") 

#To write out the data in 10X genomics format
write_data(indvExp, output_dir = "full/path/to/output/dir")
write_scflow_manifest(indvExp, output_dir = "full/path/to/output/dir")
write_scflow_samplesheet(indvExp, output_dir = "full/path/to/output/dir")