addResultsToSingleCellExperiment: Add results to SingleCellExperiment object

In slalom: Factorial Latent Variable Modeling of Single-Cell RNA-Seq Data

Description

Add results to SingleCellExperiment object

Usage

addResultsToSingleCellExperiment(sce_object, slalom_object, n_active = 20,
  mad_filter = 0.4, annotated = TRUE, unannotated_dense = FALSE,
  unannotated_sparse = FALSE, add_loadings = TRUE, dimred = "slalom",
  check_convergence = TRUE)

Arguments

sce_object	an object of class SingleCellExperiment
slalom_object	an object of class Rcpp_SlalomModel
n_active	number of terms (factors) to be added (default is 20)
mad_filter	numeric(1), filter factors by this mean absolute deviation to ensure variability in the factor states. For large datasets this can be set to 0
annotated	logical(1), should annotated factors be included? Default is TRUE
unannotated_dense	logical(1), should dense unannotated factors be included? Default is FALSE
unannotated_sparse	logical(1), should sparse unannotated factors be included? Default is FALSE
add_loadings	logical(1), should gene/feature loadings be added to the rowData of the object?
dimred	character(1), name of the reduced-dimension slot to save the factor states to. Default is "slalom"
check_convergence	logical(1), check that model has converged before adding slalom results. If TRUE and model has not converged it throws an error.

Value

a SingleCellExperiment object with factor states (X) in a reduced-dimension slot, and gene loadings for factors added to rowData.

Examples

gmtfile <- system.file("extdata", "reactome_subset.gmt", package = "slalom")
genesets <- GSEABase::getGmt(gmtfile)
data("mesc")
model <- newSlalomModel(mesc, genesets, n_hidden = 5, min_genes = 10)
model <- initSlalom(model)
model <- trainSlalom(model, nIterations = 10)
mesc <- addResultsToSingleCellExperiment(mesc, model,
check_convergence = FALSE)

slalom documentation built on Nov. 8, 2020, 5:24 p.m.