Restrict imputed values to specific conditions — restrict

This function takes a Qfeatures object with unimputed data and imputed data in separate assays and creates a new assay where the imputed values are only used in specified circumstances. Note that this restriction occurs post imputation, which may not be suitable for imputation methods where the imputed values are not independently derived

Usage

restrict_imputation(
  obj,
  i_unimputed,
  i_imputed,
  i_restricted_imputed,
  use_imputed_df,
  verbose = TRUE
)

Arguments

obj: QFeatures. Proteomics dataset
i_unimputed: string. Index for the SummarizedExperiment with data without imputation
i_imputed: string. Index for the SummarizedExperiment with data with imputation
i_restricted_imputed: string. Index for the output assay with restricted imputation
use_imputed_df: data.frame with conditions where imputation should be used. see @details for more information
verbose: logical Describe how many missing values in input and output assays

Value

Returns a QFeatures with restricted imputation with specified assay name

Details

The data.frame provided to use_imputed_df should specify the conditions under which imputation should be performed, with respect to the experimental conditions and the number of missing values. data.frame must contain a column called n_finite and all other columns must be columns in colData(obj)

For example, in an enrichment vs control experiment, to only input in control samples where zero or one replicate are quantified, this should be specified thusly

condition n_finite control 0 control 1