Returns boolean masks specifying MAR/MNAR imputation

Usage

create_imputation_mask(matrix, cutoff = "empirically")

Arguments

matrix: A matrix with log-transformed intensities
cutoff: The cutoff that specifies MAR vs MNAR. See details.

Value

A list with two matrices: one for MAR, one for MNAR.

Details

A value is considered as MNAR when the mean value of the replicates of a condition is below a threshold. This can be a fixed value between 0-1. E.g. '0.1' specifies that this threshold is the 10th percentile of all values over the replicates of a condition.

When set to 'empirically', this threshold is determined based on the data. In this case, all intensities that are the only non-missing value within replicates are collected, and the median value of this set is used as cutoff.

Examples

if (FALSE) { # \dontrun{
masks <- create_imputation_mask(data_matrix) # Default option.
masks <- create_imputation_mask(data_matrix, cutoff = 0.1) # Sets the
#MAR/MNAR cutoff at the 10th percentile of observed values
} # }