phenoFinder: Calculate pairwise similarities of phenoData between samples...

In doppelgangR: Identify likely duplicate samples from genomic or meta-data

Description

This function acts as a wrapper to phenoDist to handle cases of one ExpressionSet, a list of two identical ExpressionSets, or a list of two different ExpressionSets.

Usage

phenoFinder(eset.pair, separator = ":", ...)

Arguments

eset.pair	input: a list of ExpressionSets with two elements, or an ExpressionSet. If the two elements are identical, return the correlation matrix for pairs of samples in the first element. If not identical, return pairs between the two elements.
separator	a separator between dataset name (taken from the list names) and sample name (taken from sampleNames(eset), to keep track of which samples come from which dataset.
...	Extra arguments passed on to phenoDist

Value

A matrix of similarities between the phenotypes of pairs of samples.

Author(s)

Levi Waldron, Markus Riester, Marcel Ramos

Examples

library(curatedOvarianData)
data(GSE32063_eset)
data(GSE17260_eset)
esets2 <- list(JapaneseB=GSE32063_eset,
                Yoshihara2010=GSE17260_eset)

## standardize the sample ids to improve matching based on clinical annotation
esets2 <- lapply(esets2, function(X){
    X$alt_sample_name <- paste(X$sample_type, gsub("[^0-9]", "", X$alt_sample_name), sep="_")

## Removal of columns that cannot possibly match also helps duplicated patients to stand out
    pData(X) <- pData(X)[, !grepl("uncurated_author_metadata", colnames(pData(X)))]
    X <- X[, 1:20]  ##speed computations
    return(X) })

## See first six samples in both rows and columns
phenoFinder(esets2)[1:6, 1:6]

doppelgangR documentation built on Nov. 8, 2020, 6:36 p.m.