Use DEGenes to identify clusters with few differentially expressed genes

Source: R/find_noise_cl.R

To use this, you'll need to perform pairwise gene expression tests with de_score(), and a set of clusters with high confidence (cl.good).

find_low_quality_cl(cl.df, cl.good, de.score.mat = NULL, de.genes)

Arguments

cl.df

A data.frame with cluster annotations that includes a "cluster_label" column

cl.good

a cluster factor object for cells from high-quality clusters.

de.genes

differential gene expression results like those generated by de_score().

Value

a data.frame with the closest "good" cluster for each non-good/low cluster.

Details

Low-quality clusters will be from all clusters in cl.df that are not in cl.good.