Marr: Marr
In Ghoshlab/marr: Maximum rank reproducibility

Description Usage Arguments Details Value References Examples

This function applies an Rcpp-based implementation of a computationally efficient method for assessing reproducibility in high-throughput experiments, called the the Marr procedure. This function also defines the Marr class and constructor.

Marr(
  object,
  pSamplepairs = 0.75,
  pFeatures = 0.75,
  alpha = 0.05,
  featureVars = NULL
)

`object`	an object which is a `matrix` or `data.frame` with features (e.g. metabolites or genes) on the rows and samples as the columns. Alternatively, a user can provide a `SummarizedExperiment` object and the `assay(object)` will be used as input for the Marr procedure.
`pSamplepairs`	(Optional) a threshold value that lies between 0 and 1, used to assign a feature to be reproducible based on the reproducibility output of the sample pairs per feature. Default is 0.75.
`pFeatures`	(Optional) a threshold value that lies between 0 and 1, used to assign a sample pair to be reproducible based on the reproducibility output of the features per sample pair. Default is 0.75.
`alpha`	(Optional) level of significance to control the False Discovery Rate (FDR). Default is 0.05.
`featureVars`	(Optional) Vector of the columns which identify features. If a 'SummarizedExperiment' is used for 'data', row variables will be used.

marr (Maximum Rank Reproducibility) is a nonparametric approach, which assesses reproducibility in high-dimensional biological replicate experiments. Although it was originally developed for RNASeq data it can be applied across many different high-dimensional biological data including MassSpectrometry based Metabolomics and ChIPSeq. The Marr procedure uses a maximum rank statistic to identify reproducible signals from noise without making any distributional assumptions of reproducible signals. This procedure can be easily applied to a variety of measurement types since it employs a rank scale.

This function computes the distributions of percent reproducible sample pairs (row-wise) per feature and percent reproducible features (column-wise) per sample pair, respectively. Additionally, it also computes the percent of reproducible sample pairs and features based on a threshold value. See the vignette for more details.

A object of the class Marr that contains a numeric vector of the Marr sample pairs in the MarrSamplepairs slot, a numeric vector of the Marr features in the MarrFeatures slot, a numeric value of the Marr filtered features in the MarrSamplepairsfiltered slot, and a numeric value of the Marr filtered sample pairs in the MarrFeaturesfiltered slot.

Philtron, D., Lyu, Y., Li, Q. and Ghosh, D., 2018. Maximum Rank Reproducibility: A Nonparametric Approach to Assessing Reproducibility in Replicate Experiments. Journal of the American Statistical Association, 113(523), pp.1028-1039.

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