filterCells: filterCells

In greenelab/miQC: Flexible, probabilistic metrics for quality control of scRNA-seq data

filterCells

Description

Find those cells probabilistically determined to be compromised by the mixture model and remove them from the dataset.

Usage

filterCells(
sce,
model = NULL,
posterior_cutoff = 0.75,
keep_all_below_boundary = TRUE,
enforce_left_cutoff = TRUE,
verbose = TRUE
)

Arguments

sce	(SingleCellExperiment) Input data object.
model	(flexmix) Output of mixtureModel function, which should be explicitly called first to ensure stability of model parameters. Default = NULL.
posterior_cutoff	(numeric) The posterior probability of a cell being part of the compromised distribution, a number between 0 and 1. Any cells below the appointed cutoff will be marked to keep. Default = 0.75
keep_all_below_boundary	(boolean) Ensures that no cells below the intact cell distribution are removed. This should almost always be set to true. Default = TRUE
enforce_left_cutoff	(boolean) Prevents a U-shape in the filtering plot. Identifies the cell with the lowest mitochondrial fraction that is set to be discarded, it ensures that no cells with lower library complexity (further left) and higher mitochondrial percentage (further up) than it are kept. Default = TRUE
verbose	(boolean) Whether to report how many cells (columns) are being removed from the SingleCellExperiment object. Default = TRUE

Value

Returns a SingleCellExperiment object, the same as the input except with a new column in colData, prob_compromised, and all cells with greater than the set posterior probability removed from the dataset.

Examples

library(scRNAseq)
library(scater)
sce <- ZeiselBrainData()
mt_genes <- grepl("^mt-",  rownames(sce))
feature_ctrls <- list(mito = rownames(sce)[mt_genes])
sce <- addPerCellQC(sce, subsets = feature_ctrls)
model <- mixtureModel(sce)
sce <- filterCells(sce, model)

greenelab/miQC documentation built on Jan. 6, 2023, 12:16 a.m.