Nextflow workflow specific to outlining approaches to building and testing dynamic ML classifiers
- Nextflow 23.04.2 (requires: bash, java 11 [or later, up to 21] git and docker)
- Python 3.10+
- fpdf 1.7.2
- numpy 1.23.5
- matplotlib 3.8.0
- dataframe-image 0.2.3
- pandas 2.2.0
- seaborn 0.13.1
- xgboost 1.6.2
- scipy 1.12.0
- scikit-learn 1.4.0
Note: This pipeline requires exported QuPath (0.5+) measurement tables (quantification files) generated from segmented single cell MxIF images. Those exported files need to include some annotated classification lables.
| Field | Description |
|---|---|
| bit_depth | Original Image Capture quality: 8-bit (pixel values will be 0-255) or 16-bit (pixel values will be 0-65,535) |
| qupath_object_type | "CellObject" has two ROIs, jointly and 4 components [Cell, Cytoplasm, Membrane, Nucleus] from QuPath; 'DetectionObject' is Single Whole Cell or Nucleus only |
| classifed_column_name | |
| exclude_markers | |
| nucleus_marker | |
| override_normalization | |
| downsample_normalization_plots | |
| holdout_fraction | |
| filter_out_junk_celltype_labels | |
| minimum_label_count | |
| max_xgb_cv |