multi.impute {mi4p}

Multiple imputation of quantitative proteomics datasets

Description

multi.impute performs multiple imputation on a given quantitative proteomics dataset.

Usage

multi.impute(data, conditions, nb.imp = NULL, method, parallel = FALSE)

Arguments

data

A quantitative matrix to be imputed, with proteins/peptides in rows and samples in columns.

conditions

A vector of length the number of samples where each element corresponds to the condition the sample belongs to.

nb.imp

The number of imputation to perform.

method

A single character string describing the imputation method to be used. See details.

parallel

Logical, whether or not use parallel computing (with foreach).

Details

Multiple imputation consists in imputing several times a given dataset using a given method. Here, imputation methods can be chosen either from mice, imp4p-package or impute.knn:

• "pmm", "midastouch", "sample", "cart", "rf","mean", "norm", "norm.nob", "norm.boot", "norm.predict": imputation methods as described in mice.

• "RF" imputes missing values using random forests algorithm as described in impute.RF.

• "MLE" imputes missing values using maximum likelihood estimation as described in impute.mle.

• "PCA" imputes missing values using principal component analysis as described in impute.PCA.

• "SLSA" imputes missing values using structured least squares algorithm as described in impute.slsa.

• "kNN" imputes missing values using k nearest neighbors as described in impute.knn.

Value

A numeric array of dimension c(dim(data),nb.imp).

References

M. Chion, Ch. Carapito and F. Bertrand (2021). Accounting for multiple imputation-induced variability for differential analysis in mass spectrometry-based label-free quantitative proteomics. arxiv:2108.07086. https://arxiv.org/abs/2108.07086.

Examples

library(mi4p)
data(datasim)
multi.impute(data = datasim[,-1], conditions = attr(datasim,"metadata")$Condition, method = "MLE")