impute: Quantitative proteomics data imputation

In lgatto/MSnbase: Base Functions and Classes for Mass Spectrometry and Proteomics

Quantitative proteomics data imputation

Description

The impute method performs data imputation on MSnSet instances using a variety of methods.

Users should proceed with care when imputing data and take precautions to assure that the imputation produce valid results, in particular with naive imputations such as replacing missing values with 0.

See MsCoreUtils::impute_matrix() for details on the different imputation methods available and strategies.

Usage

## S4 method for signature 'MSnSet'
impute(object, method, ...)

Arguments

object	An MSnSet object with missing values to be imputed.
method	character(1) defining the imputation method. See MsCoreUtils::imputeMethods() for available ones. See MsCoreUtils::impute_matrix() for details.
...	Additional parameters passed to the inner imputation function. See MsCoreUtils::impute_matrix() for details.

Examples

data(naset)

## table of missing values along the rows
table(fData(naset)$nNA)

## table of missing values along the columns
pData(naset)$nNA

## non-random missing values
notna <- which(!fData(naset)$randna)
length(notna)
notna

impute(naset, method = "min")

if (require("imputeLCMD")) {
    impute(naset, method = "QRILC")
    impute(naset, method = "MinDet")
}

if (require("norm"))
    impute(naset, method = "MLE")

impute(naset, "mixed",
       randna = fData(naset)$randna,
       mar = "knn", mnar = "QRILC")


## neighbour averaging
x <- naset[1:4, 1:6]

exprs(x)[1, 1] <- NA ## min value
exprs(x)[2, 3] <- NA ## average
exprs(x)[3, 1:2] <- NA ## min value and average
## 4th row: no imputation
exprs(x)

exprs(impute(x, "nbavg"))

lgatto/MSnbase documentation built on Nov. 12, 2024, 10:58 a.m.