Data export

Data generated by steinbock can be exported to various formats for downstream data analysis.

OME-TIFF

To export images to OME-TIFF, with channel names determined by the panel file:

steinbock export ome

The exported OME-TIFF files are generated by xtiff; the default destination directory is ome.

histoCAT

To export images and masks to a folder structure compatible with histoCAT for MATLAB:

steinbock export histocat

This will create a histoCAT-compatible folder structure (defaults to histocat), with one subfolder per image, where each subfolder contains one image file per channel. Additionally, if masks are available, each image subfolder contains a single mask file.

CSV

To export specified object data from all images as a single .csv file:

steinbock export csv intensities regionprops

This will collect object data from the intensities and regionprops directories and create a single object data table in object data format, with an additional first column indicating the source image. The default destination file is objects.csv.

FCS

To export specified object data from all images as a single .fcs file:

steinbock export fcs intensities regionprops

This will collect object data from the intensities and regionprops directories and create a single object data table in FCS format using the fcswrite package. The default destination file is objects.fcs.

AnnData

To export specified object data to AnnData:

steinbock export anndata --x intensities --obs regionprops

By default, this will generate one .h5ad file per image, with object intensities as main data and object regionprops as observations annotation. The default destination directory is anndata.

AnnData file format

To export the data as .loom or .zarr, specify --format loom or --format zarr, respectively.

Multiple observations annotation sources

The --obs option can be specified multiple times to include different object data as observations annotations.

Graphs

To export neighbors as spatial object graphs, with object data as node attributes:

steinbock export graphs --data intensities

By default, this will generate one .graphml file per graph using the networkx Python package, with object intensities as node attributes. The default destination directory is graphs.

NetworkX file format

To export the graphs as .gexf or .gml, specify --format gexf or --format gml, respectively.

Multiple graph attributes sources

The --data option can be specified multiple times to include different object data as graph attributes.