impute: Impute missing values

In DEP: Differential Enrichment analysis of Proteomics data

Description

impute imputes missing values in a proteomics dataset.

Usage

impute(se, fun = c("bpca", "knn", "QRILC", "MLE", "MinDet", "MinProb",
  "man", "min", "zero", "mixed", "nbavg"), ...)

Arguments

se	SummarizedExperiment, Proteomics data (output from make_se() or make_se_parse()). It is adviced to first remove proteins with too many missing values using filter_missval() and normalize the data using normalize_vsn().
fun	"bpca", "knn", "QRILC", "MLE", "MinDet", "MinProb", "man", "min", "zero", "mixed" or "nbavg", Function used for data imputation based on manual_impute and impute.
...	Additional arguments for imputation functions as depicted in manual_impute and impute.

Value

An imputed SummarizedExperiment object.

Examples

# Load example
data <- UbiLength
data <- data[data$Reverse != "+" & data$Potential.contaminant != "+",]
data_unique <- make_unique(data, "Gene.names", "Protein.IDs", delim = ";")

# Make SummarizedExperiment
columns <- grep("LFQ.", colnames(data_unique))
exp_design <- UbiLength_ExpDesign
se <- make_se(data_unique, columns, exp_design)

# Filter and normalize
filt <- filter_missval(se, thr = 0)
norm <- normalize_vsn(filt)

# Impute missing values using different functions
imputed_MinProb <- impute(norm, fun = "MinProb", q = 0.05)
imputed_QRILC <- impute(norm, fun = "QRILC")

imputed_knn <- impute(norm, fun = "knn", k = 10, rowmax = 0.9)
imputed_MLE <- impute(norm, fun = "MLE")

imputed_manual <- impute(norm, fun = "man", shift = 1.8, scale = 0.3)

DEP documentation built on Nov. 8, 2020, 7:49 p.m.