Perform mixed imputation on a data matrix
Arguments
- data
A matrix with intensity values
- ed
The experimental design data frame.
- cutoff
The cutoff used for MAR/MNAR classification. See create_imputation_mask
Examples
se <- prepare_se(report.pg_matrix, expDesign, report.pr_matrix, impute = 'none')
data_missing <- as.matrix(SummarizedExperiment::assay(se)) # Intensity matrix with missing values
ed <- as.data.frame(SummarizedExperiment::colData(se)) # The experimental conditions
data_full <- mixed_imputation_matrix(data_missing, ed, cutoff = 'empirically')
#> Imputing along margin 2 (samples/columns).
#> [1] 0.3061269
#> Imputing along margin 1 (features/rows).
#> Warning: 36 rows with more than 50 % entries missing;
#> mean imputation used for these rows
#> Cluster size 5548 broken into 2203 3345
#> Cluster size 2203 broken into 1569 634
#> Cluster size 1569 broken into 1504 65
#> Cluster size 1504 broken into 645 859
#> Done cluster 645
#> Done cluster 859
#> Done cluster 1504
#> Done cluster 65
#> Done cluster 1569
#> Done cluster 634
#> Done cluster 2203
#> Cluster size 3345 broken into 2136 1209
#> Cluster size 2136 broken into 1087 1049
#> Done cluster 1087
#> Done cluster 1049
#> Done cluster 2136
#> Done cluster 1209
#> Done cluster 3345