Skip to content

Latest commit

 

History

History
41 lines (29 loc) · 2.34 KB

v2_0.md

File metadata and controls

41 lines (29 loc) · 2.34 KB
Raw

Overview

We expect a custom_selections HDF5 group at the root of the file, containing information about the custom selections. The group itself contains the parameters and results subgroups.

Definitions:

  • modalities: an array of names of modalities with available embeddings. This is based on the availability of embeddings in the pca and adt_pca steps
  • num_cells: number of cells remaining after QC filtering. This is typically determined from the cell_filtering step.
  • num_features: an array of length equal to modalities, containing the number of features present in each modality. This is typically determined from the inputs step.

Parameters

parameters should contain:

  • selections: a group defining the custom selections. Each child is named after a user-created selection. Each child is an integer dataset of arbitrary length containing the indices of the selected cells. Note that indices refer to the dataset after QC filtering and should be less than num_cells.

Results

results should contain per_selection, a group containing the marker results for each selection after a comparison to a group containing all other cells. Each child of per_selection is named after its selection and is itself a group containing children named after modalities. Each modality-specific child is yet another group containing the statistics for that modality:

  • means: a float dataset of length equal to the number of features in that modality (as determined from num_features), containing the mean expression of each gene in the selection.
  • detected: a float dataset of length equal to the number of features in that modality, containing the proportion of cells with detected expression of each gene in the selection.
  • lfc: a float dataset of length equal to the number of features in that modality, containing the log-fold change in the selection compared to all other cells.
  • delta_detected: same as lfc, but for the delta-detected (i.e., difference in the percentage of detected expression).
  • cohen: same as lfc, but for Cohen's d.
  • auc: same as lfc, but for the AUCs.

History

Updated in version 2.0, with the following changes from the previous version:

  • Support multiple modalities in the results.