In bschilder/scKirby: Automated ingestion and conversion of various single-cell data formats.

library(scKirby)

Intro

ingest_data() runs three main steps:

1. Read: Automatically infers the file/object type and loads it (sub-function: read_data()).
2. Convert: Converts it to the desired file/object type (sub-function: to_<format>).
3. Save: Saves the converted file/object (sub-function: save_data()).

Examples

seurat <- example_obj("seurat")
h5seurat <- example_obj("h5seurat")
sce <- example_obj("sce")
se <- example_obj("se")
h5 <- example_obj("h5")
cds <- example_obj("cds")
ctd <- example_obj("ctd")
adat <- example_obj("anndata")
library(Seurat)
loom <- example_obj("loom")

Ingest expression matrix

As SingleCellExperiment

sce2 <- ingest_data(obj = seurat, 
                    output_type = "SingleCellExperiment")

As Seurat

seurat <- ingest_data(obj=seurat@assays$RNA@counts, 
                      output_type = "Seurat") 

Ingest list

scKirby can ingest a named list (i.e. list(exp=..., annot=...)) with the following items:

• exp: Expression matrix with rows/genes x cols/cells. Can be a variety of matrix classes, including dense or sparse.

• annot: Cell annotation data.frame with one cell per row. rownames(annot) should be the same as colnames(exp).

This happens to be the format that the example data in EWCE uses, but any user-supplied data will work.

As SingleCellExperiment

sce <- ingest_data(obj=ctd)

As Seurat

seurat.2 <- ingest_data(obj = ctd, 
                        output_type = "seurat")

Ingest Seurat

As SingleCellExperiment

In-memory

sce.2 <- ingest_data(obj=seurat)

Ingest H5Seurat

On-disk

seurat.2 <- ingest_data(obj = h5seurat, 
                        output_class = "seurat")

Ingest HDF5 SingleCellExperiment

As SingleCellExperiment

## Read in the sce object directly
h5.2 <- ingest_data(obj=h5)

## Read it from disk
sce_dir <- dirname(se_filepath(h5.2))
h5.3 <- ingest_data(obj=sce_dir)

As Seurat

## Read in the sce object directly
h5.4 <- ingest_data(obj = h5,
                    output_type = "Seurat") 

Ingest CellDataSet

seurat.2 <- ingest_data(obj=cds,
                        output_type = "Seurat") 

Ingest AnnData

As SingleCellExperiment

# ### Set condaenv= to the name of a conda env you've made 
# reticulate::use_condaenv(condaenv = "echoR") 
## In memory
sce.2 <- ingest_data(obj=adat)

## On disk
path <- save_data(obj = adat,
                 filetype = "anndata")
sce.3 <- ingest_data(obj = path)

As Seurat

seurat.2 <- ingest_data(obj=adat, 
                      output_type = "Seurat")

Ingest loom

From loomR package

As SingleCellExperiment

## In memory 
sce2 <- ingest_data(obj=loom) 

## From disk 
sce3 <- ingest_data(obj=loom$filename)

As Seurat

## In memory 
seurat2 <- ingest_data(obj=loom, 
                       output_type = "Seurat") 

## From disk
print(loom$filename)
seurat3 <- ingest_data(obj=loom$filename,
                       output_type = "Seurat")

Cleanup

Remove example files

try({ file.remove("./pbmc_small.h5Seurat", showWarnings=FALSE) })
try({ unlink("./pbmc_small_h5/", recursive = TRUE) })
try({ file.remove("./pbmc_small.h5ad") })
try({ file.remove("./pbmc_small.loom")  })

Session Info

utils::sessionInfo()

---

bschilder/scKirby documentation built on April 14, 2025, 7 p.m.