This function reads the standard CosMX output into an SFE object, as in "Basic Data Files" on the Nanostring website.
Arguments
- data_dir
Top level output directory.
- z
Integer z index or "all" to indicate which z-planes to read for the transcript spots.
- sample_id
A
character
sample identifier, which matches thesample_id
inimgData
. Thesample_id
will also be stored in a new column incolData
, if not already present. Default =sample01
.- add_molecules
Logical, whether to add transcripts coordinates to an object.
- split_cell_comps
Logical, whether to split transcript spot geometries by cell compartment. Only relevant when `add_molecules = TRUE`.
- BPPARAM
A
BiocParallelParam
object specifying parallel processing backend and number of threads to use for parallelizable tasks:To load cell segmentation from HDF5 files from different fields of view (FOVs) with multiple cores. A progress bar can be configured in the
BiocParallelParam
object. When there are numerous FOVs, reading in the geometries can be time consuming, so we recommend using a server and larger number of threads. This argument is not used ifuse_cellpose = TRUE
and the parquet file is present.To get the largest piece and see if it's larger than
min_area
when there are multiple pieces in the cell segmentation for one cell.
- file_out
Name of file to save the geometry or raster to disk. Especially when the geometries are so large that it's unwieldy to load everything into memory. If this file (or directory for multiple files) already exists, then the existing file(s) will be read, skipping the processing. When writing the file, extensions supplied are ignored and extensions are determined based on `dest`.
- z_option
What to do with z coordinates. "3d" is to construct 3D geometries. "split" is to create a separate 2D geometry for each z-plane so geometric operations are fully supported but some data wrangling is required to perform 3D analyses. When the z coordinates are not integers, 3D geometries will always be constructed since there are no z-planes to speak of. This argument does not apply when `spatialCoordsNames` has length 2.
Value
An SFE object. Cell polygons are written to `cell_boundaries_sf.parquet` in `data_dir`. If reading transcript spots (`add_molecules = TRUE`), then the reformatted transcript spots are saved to file specified in the `file_out` argument, which is by default `tx_spots.parquet` in the same directory as the rest of the data.
Examples
fp <- tempdir()
dir_use <- SFEData::CosMXOutput(file_path = file.path(fp, "cosmx_test"))
#> see ?SFEData and browseVignettes('SFEData') for documentation
#> loading from cache
#> The downloaded files are in /tmp/RtmpRqe7ue/cosmx_test/cosmx
sfe <- readCosMX(dir_use, z = "all", add_molecules = TRUE)
#> >>> Constructing cell polygons
#> >>> Reading transcript coordinates
#> >>> Converting transcript spots to geometry
#> >>> Writing reformatted transcript spots to disk
# Clean up
unlink(dir_use, recursive = TRUE)