HandleBiologicalZeros: Get dropout probabilities

In anacarolinaleote/ADImpute: Adaptive Dropout Imputer (ADImpute)

Description

GetDropoutProbabilities computes dropout probabilities (probability of being a dropout that should be imputed rather than a true biological zero) using an adaptation of scImpute's approach

Usage

HandleBiologicalZeros(data, imputed, thre = 0.5, cell.clusters,
labels = NULL, type = 'count', cores = BiocParallel::bpworkers(BPPARAM),
BPPARAM = BiocParallel::SnowParam(type = "SOCK"),
genelen = ADImpute::transcript_length, prob.mat = NULL)

Arguments

data	matrix; original data before imputation
imputed	list; imputation results for considered methods
thre	numeric; between 0 and 1 specifying the threshold to determine dropout values
cell.clusters	integer; number of cell subpopulations
labels	character; vector specifying the cell type of each column of data
type	A character specifying the type of values in the expression matrix. Can be 'count' or 'TPM'
cores	integer; number of cores used for paralell computation
BPPARAM	parallel back-end to be used during parallel computation. See BiocParallelParam-class.
genelen	matrix with at least 2 columns: 'hgnc_symbol' and 'transcript_length'
prob.mat	matrix with same dimensions as data containing the dropout probabilities for the corresponding entries

Details

This function follows scImpute's model to distinguish between true biological zeros and dropouts, and is based on adapted code from the scImpute R package.

Value

list with 2 components: zerofiltered, a list equivalent to imputed but with entries of imputed likely biological zeros set back to zero, and dropoutprobabilities matrix with same dimensions as data containing the dropout probabilities for the corresponding entries

anacarolinaleote/ADImpute documentation built on May 18, 2021, 10:11 p.m.