MarrFilterData: Filter by Maximum Rank Reproducibility

In Ghoshlab/marr: Maximum rank reproducibility

Description

Filters Marr object according to the Maximum Rank Reproducibility of the features, samples pairs, or both. Features are removed if their reproducibility per sample pair is less than pFeatures. Samples are removed if their sample pair reproducibility per feature is less than pSamplepairs for all pairings of that sample and the other samples in the set.

Usage

MarrFilterData(object, by = c("both", "features", "samplePairs"))

Arguments

object	a Marr object from Marr
by	String specifying which reproducibility values to filter by. Options include "features" to filter features according to their reproducibility, "samplePairs" to filter samples according to the reproducibility of sample pairs, or "both" to filter both features and sample pairs according to their respective reproducibility. Default is "both".

Value

A list of data.frame's or a SummarizedExperiment. If a data.frame was originally input into the Marr function, a list with three elements, filteredData, removedSamples, and removedFeatures, will be returned. If a SummarizedExperiment was originally input, output will be a SummarizedExperiment with the assay filtered and with two metadata objects, removedSamples and removedFeatures added.

Examples

data <- matrix(rnorm(2400), nrow=200, ncol=12)
data_Marr <- Marr(object = data, pSamplepairs=0.75,
                  pFeatures=0.75, alpha=0.05)
MarrFilterData(data_Marr, by = "both")

Ghoshlab/marr documentation built on Oct. 3, 2021, 3:20 p.m.