call_doublets: Automatically sets a threshold for calling doublets, but it's...

In iaaaka/Rscrublet: A tool for identifying doublets in single-cell RNA-seq data.

Description

Automatically sets a threshold for calling doublets, but it's best to check this by running plot_doublet_histogram() afterwards and adjusting threshold by call_doublets with user defined threshold parameter

Usage

call_doublets(nnc, threshold = NULL, verbose = TRUE)

Arguments

nnc	list, output of scrub_doublets function
threshold	Doublet score threshold for calling a transcriptome a doublet. If NULL, this is set automatically by looking for the minimum between the two modes of the nnc$doublet_scores_sim histogram. It is best practice to check the threshold visually using the nnc$doublet_scores_sim histogram and/or based on co-localization of predicted doublets in a 2-D embedding.
verbose	If TRUE, print summary statistics.

Value

nnc list with additional items: #'

• predicted_doublets - logical index of predicted doublet cells

• z_scores_ - Z-score conveying confidence in doublet calls.

• threshold - Doublet score threshold for calling a transcriptome a doublet.

• detected_doublet_rate - Fraction of observed transcriptomes that have been called doublets.

• detectable_doublet_fraction - Estimated fraction of doublets that are detectable, i.e., fraction of simulated doublets with doublet scores above threshold

• overall_doublet_rate - Estimated overall doublet rate, detected_doublet_rate/detectable_doublet_fraction. Should agree (roughly) with expected_doublet_rate.

Examples

# run rscrublet of 8k pbmc example dataset
scrr = scrub_doublets(E_obs = pbmc8k,expected_doublet_rate=0.06,min_counts=2, min_cells=3, min_gene_variability_pctl=85, n_prin_comps=30)
# set threshould automatically 
scrr=call_doublets(scrr)
# examine score distribution
plot_doublet_histogram(scrr)
# find predicted doublets
rownames(pbmc8k)[scrr$predicted_doublets]

iaaaka/Rscrublet documentation built on Dec. 20, 2021, 5:57 p.m.