scmorph.qc.filter_outliers

scmorph.qc.filter_outliers(adata, outliers=0.05, fraction=None, n_obs=None, detect_only=False, n_cores=1)

Filter outlier observations from an AnnData object.

Note

The outliers argument determines how many cells will be classified as outlier cells. Since it is an arbitrary threshold this will depend on your dataset and downstream analysis. We encourage you to try different values and see which one works best for your dataset.

Parameters:

adata AnnData

Annotated data matrix.

outliers float (default: 0.05)

Expected fraction of outlier cells.

fraction float

During training, subsample to this fraction of the number of observations.

n_obs int | None (default: None)

During training, subsample to this number of observations. We recommend 10,000 or fewer, as this results in faster training with adequate accuracy.

detect_only bool (default: False)

Whether to only detect outliers but not filter them.

n_cores int (default: 1)

Number of cores to use for parallelization. -1 for all cores.

Return type:

AnnData

Returns:

adata