Currently, the module tags each column with a suffix that specifies which mask was used during quantification, e.g., Hoechst0_cellMask, CD20_nucleiMask, etc. This creates a mismatch between column names and marker names in markers.csv, requiring all downstream modules to perform additional processing.
TODO: Add a command-line option that moves the mask suffix from column names to filenames. In other words, instead of producing a single exemplar-001.csv that contains columns CD45_cellMask, CD45_nucleiMask, etc., have the tool output multiple files, exemplar-001_cellMask.csv, exemplar-001_nucleiMask.csv, etc., with each file containing the column CD45, which will be a direct match against the names in markers.csv.
As a bonus, multiple files will provide morphological features for the different masks, allowing one to compare, e.g., the ratio of nucleus and cytoplasm areas.
ArtemSokolov
commented
3 years ago
Currently, the module tags each column with a suffix that specifies which mask was used during quantification, e.g.,
Hoechst0_cellMask,CD20_nucleiMask, etc. This creates a mismatch between column names and marker names inmarkers.csv, requiring all downstream modules to perform additional processing.TODO: Add a command-line option that moves the mask suffix from column names to filenames. In other words, instead of producing a single
exemplar-001.csvthat contains columnsCD45_cellMask,CD45_nucleiMask, etc., have the tool output multiple files,exemplar-001_cellMask.csv,exemplar-001_nucleiMask.csv, etc., with each file containing the columnCD45, which will be a direct match against the names inmarkers.csv.As a bonus, multiple files will provide morphological features for the different masks, allowing one to compare, e.g., the ratio of nucleus and cytoplasm areas.