msImpute: Function for imputing missing values in data.

Description Usage Arguments Value References Examples

View source: R/impute.R

Description

Replaces missing values with non-zero estimates calculated using a selected method.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
msImpute(
  data,
  imputeMethod = c("halfmin", "bpca", "knn"),
  kKnn = 5,
  nPcs = 3,
  compoundsAsNeighbors = FALSE,
  compVars = c("mz", "rt"),
  sampleVars = c("subject_id"),
  colExtraText = NULL,
  separator = NULL,
  missingValue = NA,
  returnToSE = FALSE,
  returnToDF = FALSE
)

Arguments

data

Data set as either a data frame or 'SummarizedExperiement'.

imputeMethod

String specifying imputation method. Options are "halfmin" (half the minimum value), "bpca" (Bayesian PCA), and "knn" (k-nearest neighbors).

kKnn

Number of clusters for 'knn' method.

nPcs

Number of principle components used for re-estimation for 'bpca' method.

compoundsAsNeighbors

For KNN imputation. If TRUE, compounds will be used as neighbors rather than samples. Note that using compounds as neighbors is significantly slower than using samples.

compVars

Vector of the columns which identify compounds. If a 'SummarizedExperiment' is used for 'data', row variables will be used.

sampleVars

Vector of the ordered sample variables found in each sample column.

colExtraText

Any extra text to ignore at the beginning of the sample columns names. Unused for 'SummarizedExperiments'.

separator

Character or text separating each sample variable in sample columns. Unused for 'SummarizedExperiment'.

missingValue

Specifies the abundance value which indicates missing data. May be a numeric or 'NA'.

returnToSE

Logical value indicating whether to return as 'SummarizedExperiment'

returnToDF

Logical value indicating whether to return as data frame.

Value

A data frame or 'SummarizedExperiment' with missing data imputed. Default return type is set to match the data input but may be altered with the 'returnToSE' or 'returnToDF' arguments.

References

Oba, S.et al.(2003) A Bayesian missing value estimation for gene expression profile data. Bioinformatics, 19, 2088-2096

Stacklies, W.et al.(2007) pcaMethods A bioconductor package providing PCA methods for incomplete data. Bioinformatics, 23, 1164-1167.

A. Kowarik, M. Templ (2016) Imputation with R package VIM. Journal of Statistical Software, 74(7), 1-16.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
# Load, tidy, summarize, and filter example dataset
data(msquant)

summarizedDF <- msSummarize(msquant,
                            compVars = c("mz", "rt"),
                            sampleVars = c("spike", "batch", "replicate", 
                            "subject_id"),
                            cvMax = 0.50,
                            minPropPresent = 1/3,
                            colExtraText = "Neutral_Operator_Dif_Pos_",
                            separator = "_",
                            missingValue = 1)
                            
filteredDF <- msFilter(summarizedDF,
                       filterPercent = 0.8,
                       compVars = c("mz", "rt"),
                       sampleVars = c("spike", "batch", "subject_id"),
                       separator = "_")

                           
# Impute dataset using 3 possible options
hmImputedDF <- msImpute(filteredDF, imputeMethod = "halfmin",
                        compVars = c("mz", "rt"),
                        sampleVars = c("spike", "batch", "subject_id"),
                        separator = "_",
                        missingValue = 0)

bpcaImputedDF <- msImpute(filteredDF, imputeMethod = "bpca",
                          nPcs = 3,
                          compVars = c("mz", "rt"),
                          sampleVars = c("spike", "batch", "subject_id"),
                          separator = "_",
                          missingValue = 0)

knnImputedDF <- msImpute(filteredDF, imputeMethod = "knn",
                         kKnn = 5,
                         compVars = c("mz", "rt"),
                         sampleVars = c("spike", "batch", "subject_id"),
                         separator = "_",
                         missingValue = 0)                                

MSPrep documentation built on Nov. 8, 2020, 5:07 p.m.