Preprocessing: pp

The preprocessing module crested.pp hosts functions that prepare an AnnData object for CREsted training and evaluation.

Functions

change_regions_width(adata, width[, ...])	Change the widths of all regions in the adata object.
filter_regions_on_specificity(adata[, ...])	Filter bed regions & targets/predictions based on high Gini score.
normalize_peaks(adata[, peak_threshold, ...])	Normalize the adata.X based on variability of the top values per cell type.
sort_and_filter_regions_on_specificity(...)	Sort bed regions & targets/predictions based on high Gini or proportion score per colum while keeping the top k rows per column.
train_val_test_split(adata[, strategy, ...])	Add 'train/val/test' split column to AnnData object.