query.evaluation: Evaluate multiple normalization methods and multiple DE...

In jhsiao999/ashbun: Methods and evaluation tools for single cell RNA-seq count data

Usage

query.evaluation(counts, condition, is_nullgene, methodsNormalize = c("LIB",
  "TMM", "RLE", "census", "SCnorm", "scran"),
  methodsMeanExpression = c("DESeq2", "limmaVoom"),
  report.control = list(fdr_cutoff = 0.05), nsim = NULL)

Arguments

thresholdDetection	minimum count per gene/sample. Default value = 1.
fractionExpressed	fraction of samples expressed (above thresholdDetection). Default value = .01.

Author(s)

Chiaowen Joyce Hsiao

Examples

ipsc_eset <- get(load(system.file("testdata", "HumanTungiPSC.rda", package = "ashbun")))
counts <- exprs(ipsc_eset)[sample(nrow(exprs(ipsc_eset)), ), ]

#---- generat simulated datasets
library(ashbun)
simdata_list <- simulationWrapper(counts, Nsim = 2,
                                 Ngenes = 100,
                                 Nsam = 20,
                                 sample_method = "all_genes",
                                 pi0 = .5,
                                 beta_args = args.big_normal(betapi = 1,
                                                             betamu = 0, betasd = .8))

#---- extract a single dataset as an example
#---- take pi0 = .9, the first simulated data
simdata <- simdata_list[[1]]

# ---- gather evaluation results
eval_ouptut <- query.evaluation(counts = simdata$counts,
                                condition = simdata$condition,
                                is_nullgene = simdata$is_nullgene,
                                methodsNormalize = c("TMM", "RLE"),
                                methodsMeanExpression = c("DESeq2", "limmaVoom"),
                                report = "fdr_cutoff_summary")

jhsiao999/ashbun documentation built on May 8, 2019, 11:17 p.m.