Reads in single-cell data generated by the ImcSegmentationPipeline

Source: R/read_cpout.R

Reader function to generate a SpatialExperiment or SingleCellExperiment object from single-cell data obtained by the ImcSegmentationPipeline pipeline.

read_cpout(
  path,
  object_file = "cell.csv",
  image_file = "Image.csv",
  panel_file = "panel.csv",
  graph_file = "Object relationships.csv",
  object_feature_file = "var_cell.csv",
  intensities = "Intensity_MeanIntensity_FullStack",
  extract_imgid_from = "ImageNumber",
  extract_cellid_from = "ObjectNumber",
  extract_coords_from = c("Location_Center_X", "Location_Center_Y"),
  extract_cellmetadata_from = c("AreaShape_Area", "Neighbors_NumberOfNeighbors_8",
    "AreaShape_Eccentricity", "AreaShape_MajorAxisLength", "AreaShape_MinorAxisLength",
    "AreaShape_MeanRadius"),
  extract_imagemetadata_from = c("Metadata_acname", "Metadata_acid",
    "Metadata_description"),
  extract_graphimageid_from = "First Image Number",
  extract_graphcellids_from = c("First Object Number", "Second Object Number"),
  extract_metal_from = "Metal Tag",
  scale_intensities = TRUE,
  extract_scalingfactor_from = "Scaling_FullStack",
  return_as = c("spe", "sce")
)

Arguments

path

full path to the CellProfiler output folder.

object_file

single character indicating the file name storing the object/cell-specific intensities and metadata.

image_file

single character indicating the file name storing meta data per image (can be NULL).

panel_file

single character indicating the file name storing the panel information (can be NULL).

graph_file

single character indicating the file name storing the object/cell interaction information (can be NULL).

object_feature_file

single character indicating the file name storing object/cell feature information.

intensities

single character indicating which column entries of the object_file contain the intensity features of interest. See details.

extract_imgid_from

single character indicating which column entries of the object_file and image_file contain the image integer ID.

extract_cellid_from

single character indicating which column entry of the object_file contains the object/cell integer ID.

extract_coords_from

character vector indicating which column entries of the object_file contain the x and y location of the objects/cells.

extract_cellmetadata_from

character vector indicating which additional object/cell specific metadata to extract from the object_file.

extract_imagemetadata_from

character vector indicating which additional image specific metadata to extract from the image_file. These will be stored in the colData(x) slot as object/cell-specific entries.

extract_graphimageid_from

single character indicating which column entries of the graph_file contain the image integer ID.

extract_graphcellids_from

character vector indicating which column entries of the graph_file contain the first and second object/cell integer IDs. These will be stored as the from and to entry of the SelfHits object in colPair(x, "neighborhood").

extract_metal_from

single character indicating which column entry of the panel_file contains the metal isotopes of the used antibodies. This entry is used to match the panel information to the acquired channel information.

scale_intensities

single logical. Should the measured intensity features be scaled by extract_scalingfactor_from.

extract_scalingfactor_from

single character indicating which column entries of the image_file contain the image specific scaling factor.

return_as

should the object be returned as SpatialExperiment (return_as = "spe") or SingleCellExperiment (return_as = "sce").

Value

returns a SpatialExperiment or SingleCellExperiment

object with markers in rows and cells in columns.

The returned data container

In the case of both containers x, intensity features (as selected by the intensities parameter) are stored in the counts(x) slot. Cell metadata (e.g morphological features) are stored in the colData(x) slot. The interaction graphs are stored as SelfHits object in the colPair(x, "neighborhood") slot.

Intensity features are extracted via partial string matching. Internally, the read_cpout function checks if per channel a single intensity feature is read in (by checking the _cXY ending where XY is the channel number).

In the case of a returned SpatialExperiment object, the cell coordinates are stored in the spatialCoords(x) slot.

In the case of a returned SingleCellExperiment object, the cell coordinates are stored in the colData(x) slot named as Pos_X and Pos_Y.

See also

https://github.com/BodenmillerGroup/ImcSegmentationPipeline for the pipeline

read_steinbock for reading in single-cell data as produced by the steinbock pipeline

colPair for information on how to work with the cell-cell interaction graphs

Author

Tobias Hoch

Nils Eling (nils.eling@dqbm.uzh.ch)

Examples

path <- system.file("extdata/mockData/cpout", package = "imcRtools")

# Read in as SpatialExperiment object
x <- read_cpout(path, graph_file = "Object_relationships.csv")
x
#> class: SpatialExperiment 
#> dim: 5 239 
#> metadata(0):
#> assays(1): counts
#> rownames(5): Ag107 Pr141 Sm147 Eu153 Yb172
#> rowData names(7): Tube.Number Metal.Tag ... deepcell cellpose
#> colnames(239): 1_1 1_2 ... 15_29 15_30
#> colData names(12): sample_id ObjectNumber ... Metadata_acid
#>   Metadata_description
#> reducedDimNames(0):
#> mainExpName: NULL
#> altExpNames(0):
#> spatialCoords names(2) : Pos_X Pos_Y
#> imgData names(1): sample_id

# Read in as SingleCellExperiment object
x <- read_cpout(path, graph_file = "Object_relationships.csv", 
                return_as = "sce")
x
#> class: SingleCellExperiment 
#> dim: 5 239 
#> metadata(0):
#> assays(1): counts
#> rownames(5): Ag107 Pr141 Sm147 Eu153 Yb172
#> rowData names(7): Tube.Number Metal.Tag ... deepcell cellpose
#> colnames(239): 1_1 1_2 ... 15_29 15_30
#> colData names(14): sample_id ObjectNumber ... Metadata_acid
#>   Metadata_description
#> reducedDimNames(0):
#> mainExpName: NULL
#> altExpNames(0):