Package 'scds'

Find doublets/multiplets in UMI scRNA-seq data;

Description

Annotates doublets/multiplets using a binary classification approach to discriminate artificial doublets from original data.

Usage

bcds(sce, ntop = 500, srat = 1, verb = FALSE, retRes = FALSE,
  nmax = "tune", varImp = FALSE, estNdbl = FALSE)

Arguments

sce	single cell experiment (SingleCellExperiment) object to analyze; needs counts in assays slot.
ntop	integer, indicating number of top variance genes to consider. Default: 500
srat	numeric, indicating ratio between orginal number of "cells" and simulated doublets; Default: 1
verb	progress messages. Default: FALSE
retRes	logical, should the trained classifier be returned? Default: FALSE
nmax	maximum number of training rounds; integer or "tune". Default: "tune"
varImp	logical, should variable (i.e., gene) importance be returned? Default: FALSE
estNdbl	logical, should the numer of doublets be estimated from the data. Enables doublet calls. Default:FALSE. Use with caution.

Value

sce input sce object SingleCellExperiment with doublet scores added to colData as "bcds_score" column, and possibly more (details)

Examples

data("sce_chcl")
## create small data set using only 100 cells
sce_chcl_small = sce_chcl[, 1:100]
sce_chcl_small = bcds(sce_chcl_small)

---

Find doublets/multiplets in UMI scRNA-seq data;

Description

Annotates doublets/multiplets using co-expression based approach

Usage

cxds(sce, ntop = 500, binThresh = 0, verb = FALSE, retRes = FALSE,
  estNdbl = FALSE)

Arguments

sce	single cell experiment (SingleCellExperiment) object to analyze; needs counts in assays slot.
ntop	integer, indimessageing number of top variance genes to consider. Default: 500
binThresh	integer, minimum counts to consider a gene "present" in a cell. Default: 0
verb	progress messages. Default: FALSE
retRes	logical, whether to return gene pair scores & top-scoring gene pairs? Default: FALSE.
estNdbl	logical, should the numer of doublets be estimated from the data. Enables doublet calls. Default:FALSE. Use with caution.

Value

sce input sce object SingleCellExperiment with doublet scores added to colData as "cxds_score" column.

Examples

data("sce_chcl")
## create small data set using only 100 cells
sce_chcl_small = sce_chcl[, 1:100]
sce_chcl_small = cxds(sce_chcl_small)

---

Find doublets/multiples in UMI scRNA-seq data;

Description

Annotates doublets/multiplets using the hybrid approach

Usage

cxds_bcds_hybrid(sce, cxdsArgs = NULL, bcdsArgs = NULL, verb = FALSE,
  estNdbl = FALSE, force = FALSE)

Arguments

sce	single cell experiment (SingleCellExperiment) object to analyze; needs counts in assays slot.
cxdsArgs	list, arguments for cxds function in list form. Default: NULL
bcdsArgs	list, arguments for bcds function in list form. Default: NULL
verb	logical, switch on/off progress messages
estNdbl	logical, should the numer of doublets be estimated from the data. Enables doublet calls. Default:FALSE. Use with caution.
force	logical, force a (re)run of cxds and bcds. Default: FALSE

Value

sce input sce object SingleCellExperiment with doublet scores added to colData as "hybrid_score" column.

Examples

data("sce_chcl")
## create small data set using only 100 cells
sce_chcl_small = sce_chcl[, 1:100]
sce_chcl_small = cxds_bcds_hybrid(sce_chcl_small)

---

Extract top-scoring gene pairs from an SingleCellExperiment where cxds has been run

Description

Extract top-scoring gene pairs from an SingleCellExperiment where cxds has been run

Usage

cxds_getTopPairs(sce, n = 100)

Arguments

sce	single cell experiment to analyze; needs "counts" in assays slot.
n	integer. The number of gene pairs to extract. Default: 100

Value

matrix Matrix with two colulmns, each containing gene indexes for gene pairs (rows).

---

Wrapper for getting doublet calls

Description

Wrapper for getting doublet calls

Usage

get_dblCalls_ALL(scrs_real, scrs_sim, rel_loss = 1)

Arguments

scrs_real	numeric vector, the scores for the real/original data
scrs_sim	numeric vector, the scores for the artificial doublets
rel_loss	numeric scalar, relative weight of a false positive classification compared with a false negative. Default:1 (same loss for fp and fn).

Value

numeric, matrix containing the (estimated) number of doublets, the score threshold and the fraction of artificial doublets missed (false negative rate, of sorts) as columns and four types of estimating: "youden", "balanced" and a false negative rate of artificial doublets of 0.1 and 0.01, respecitvely.

---

Derive doublet calls from doublset scores

Description

Given score vectors for real data and artificial doubles, derive doublet calls based on determining doublet score cutoffs.

Usage

get_dblCalls_dist(scrs_real, scrs_sim, type = "balanced")

Arguments

scrs_real	numeric vector, the scores for the real/original data
scrs_sim	numeric vector, the scores for the artificial doublets
type	character or numeric, describes how the score threshold for calling doublets is determined. Either "balanced" or a number between zero and one that indicates the fraction of artificial doublets missed when making calls. Default: "balanced".

Value

numeric, vector containing the (estimated) number of doublets, the score threshold and the fraction of artificial doublets missed (false negative rate, of sorts)

---

Derive doublet calls from classification probabilities

Description

Given class probabilities (or scores) discriminating real data from artificial doublets, derive doublet calls. Based on selecting a ROC cutoff, see The Inconsistency of ‘‘Optimal’’ Cutpoints Obtained using Two Criteria basedon the Receiver Operating Characteristic Curve, (doi).

Usage

get_dblCalls_ROC(scrs_real, scrs_sim, rel_loss = 1)

Arguments

scrs_real	numeric vector, the scores for the real/original data
scrs_sim	numeric vector, the scores for the artificial doublets
rel_loss	numeric scalar, relative weight of a false positive classification compared with a false negative. Default:1 (same loss for fp and fn).

Value

numeric, vector containing the (estimated) number of doublets, the score threshold and the fraction of artificial doublets missed (false negative rate, of sorts)

---

Example single cell experiment (SingleCellExperiment) object

Description

Example data set, created by randomly sampling genes and cells from a real data set (ch_cl, i.e., the cell lines data from https://satijalab.org/seurat/hashing_vignette.html). Contains raw counts in the counts assay slot.

Usage

sce_chcl

Format

a single cell experiment object (SingleCellExperiment) with raw counts in the counts in assays, and colData with experimental annotations.

---