Impute missing values using zCompositions

Replaces missing values in an LFQData object using methods from the zCompositions package. The limit of detection (LOD) can be estimated globally or per-sample using quantiles.

Usage

impute_with_zcomp(
  lfqdata,
  method = c("multRepl", "GBM", "SQ", "BL", "CZM"),
  lod = c("global", "quantile")
)

Arguments

lfqdata

LFQData object containing the data to impute

method

imputation method: "multRepl" (multiplicative replacement), "GBM", "SQ", "BL", or "CZM" (passed to zCompositions)

lod

limit of detection strategy, either "global" or "quantile"

Value

the modified LFQData object (lfqdata), with imputed values

Note

This function assumes that missing values are Missing Completely At Random (MCAR) or Missing At Random (MAR). If missingness is abundance-dependent (MNAR, common in proteomics DDA), the imputed values and downstream statistics may be biased. For MNAR-aware analysis, consider packages such as proDA or DEqMS.

Examples

dd <- prolfqua::sim_lfq_data_peptide_config()
#> creating sampleName from fileName column
#> completing cases
#> completing cases done
#> setup done
lfqdata <- LFQData$new(dd$data, dd$config)
if (requireNamespace("zCompositions", quietly = TRUE)) {
  wide_before <- lfqdata$to_wide(as.matrix = TRUE)
  has_na_before <- any(is.na(wide_before$data))
  lfqdata <- impute_with_zcomp(lfqdata, method = "multRepl", lod = "global")
  wide_after <- lfqdata$to_wide(as.matrix = TRUE)
  has_na_after <- any(is.na(wide_after$data))
  stopifnot(has_na_before || !has_na_after)
  stopifnot(!has_na_after)
}
#> Warning: Expected 2 pieces. Missing pieces filled with `NA` in 336 rows [1, 2, 3, 4, 5,
#> 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, ...].
#> Joining with `by = join_by(sampleName, protein_Id, peptide_Id)`