Remove duplicated patient samples (samples from the same patient/column ID)

Function to keep only 1 sample per patient (column ID) in the data matrix. Keeps the sample that has the overall highest variance.

Usage

removeDuplicatedPatients(
  exprMatrix,
  outputFile = "duplicatedPatientsOutput.txt",
  varMetric = c("everything", "all.obs", "complete.obs", "na.or.complete",
    "pairwise.complete.obs")
)

Arguments

exprMatrix

Expression matrix, with features in the rows and samples in the columns.

outputFile

Output file for messages that print status of removing duplicated samples. Include full directory if file should not be printed to current working directory.

varMetric

Standard options taken from the base var() function. May be important if you have NA values in your data matrix; otherwise, "everything" is usually fine.

Value

exprMatrix: the final data matrix with only 1 sample per patient ID.

Note

Suggestions are welcome for further ways to pick the best sample from samples from the same patient. No curatedBreastData matrices currently have samples that share the same patient ID, but this function is especially useful for say TCGA data, where this is often the case.

It is suggseted one imputes missing values using the filterAndImpute function before running this function to avoid -Inf and NA values in the variance calculations.

Author

Katie Planey <katie.planey@gmail.com>

Examples

#No curatedBreastData has duplicated samples, 
#but we can still run this function on one of the datasets:
#load up our datasets
curatedBreastDataExprSetList <- getCuratedBreastDataExprSetList(test=TRUE)

#This dataset does not have NA values, which makes for a good example without
#extra pre-processing.
outputMatrix <- removeDuplicatedPatients(exprMatrix=
exprs(curatedBreastDataExprSetList[[1]]), 
outputFile = tempfile(), varMetric = c("everything"))
#> 
#> Starting with  30patients.
#> found no multiple samples from the same patient(s)
#final dimensions - unchanged in this case with 
#no samples sharing the same patient ID.
dim(outputMatrix)
#> [1] 500  30