Simulate Single Cell Analysis Pipeline Expression

Source: R/Flow.R

Blabla

simulateFlow(n1=c(20,50,100), n2=c(30,60,120),
sim.settings,
DEmethod,
normalisation,
Preclust=FALSE,
Prefilter = NULL,
Impute = NULL,
DEFilter = FALSE,
GeneSelect=NULL,
DimReduce=NULL,
ClustMethod=NULL,
Pipeline=NULL,
spikeIns=FALSE,
NCores=NULL,
verbose=TRUE)

Arguments

n1, n2

Integer vectors specifying the number of biological replicates in each group. Default values are n1=c(20,50,100) and n2=c(30,60,120).
sim.settings

This object specifies the simulation setup. This must be the return object from SimSetup.
DEmethod

A character vector specifying the DE detection method to be used. Please consult the Details section for available options.
normalisation

Normalisation method to use. Please consult the Details section for available options.
Preclust

A logical vector indicating whether to run a hierarchical clustering prior to normalisation. This is implemented for scran only. Default is FALSE. For details, see quickCluster.
Prefilter

A character vector specifying the gene expression filtering method to be used prior to normalisation (and possibly imputation). Default is NULL, i.e. no filtering. Please consult the Details section for available options.
Impute

A character vector specifying the gene expression imputation method to be used prior to normalisation. Default is NULL, i.e. no imputation. Please consult the Details section for available options.
DEFilter

A logical vector indicating whether to run DE testing on filtered and/or imputed count data. Default is FALSE.
GeneSelect

A character vector specifying the gene selection method. Default is NULL, i.e. no gene selection. Availabe options are: HVG, Gini-Index.
DimReduce

A character vector specifying the dimension reduction method. Default is NULL, i.e. no dimension reduction. Availabe options are: Euclidean, Spearman, PCA, t-SNE-expr, t-SNE-Euclidean, PCA+t-SNE, CIDR.
ClustMethod

A character vector specifying the clustering method. Default is NULL, i.e. no clustering. Availabe options are: kmeans, PAM.
Pipeline

A character vector specifying out-of-the-box single cell clustering pipelines. Available options are: CIDR_free and CIDR_bound.
spikeIns

Logical value to indicate whether to simulate spike-ins. Default is FALSE.
NCores

integer positive number of cores for parallel processing, default is NULL, ie 1 core.
verbose

Logical value to indicate whether to show progress report of simulations. Default is TRUE.

Value

A list with the following fields:

pvalue, fdr

3D array (ngenes * N * nsims) for p-values and FDR from each simulation. Note that FDR values will be empty and the calculation will be done by evaluateDE whenever applicable.

mu,disp,dropout

3D (ngenes * N * nsims) array for mean, dispersion and dropout of library size factor normalized read counts.

elfc,rlfc

3D array (ngenes * N * nsims) for log2 fold changes (LFC): elfc is for the DE tool estimated LFC; rlfc is for the LFC estimated from the normalised read counts.

sf.values,gsf.values

3D array (ngenes * N * nsims) for size factor estimates. Global estimates per sample in sf.values; Gene- and sample-wise estimates in gsf.values only for SCnorm normalisation.

true.designs,def.designs

3D array (ngenes * N * nsims) for group assignment specifications. true.designs for the simulated group assignment; def.designs for the group assignment determined after clustering.

sim.settings

The input sim.settings to which the specifications of simulateDE is added.

time.taken

The time taken for each simulation, given for preprocessing, normalisation, clustering, differential expression testing and moment estimation.

See also

estimateParam, insilicoNBParam for negative binomial parameter specifications;
DESetup, SimSetup for simulation setup

Examples

# NOT RUN {
## not yet
# }