Skip to contents

De-identify a markerLdBlock() result table

Source: R/obfuscateLdBlocks.R
obfuscateLdBlocks.Rd

Remaps the idsUsed column of a markerLdBlock result table through the same alias vector obfuscatePed(..., map = TRUE) already returns, mirroring obfuscateTwinRelations's pattern. markerLdBlock's output is otherwise a locus-pair-level population statistic table with no per-individual ids – idsUsed (populated only when markerLdBlock was called with founderIds) is the only place a real id can appear, and any exported block/LD statistic table must route through the same curator-controlled de-identification gate issue #150 established (D9): a joint, multi-locus statistic carries more identifying power than a single-locus one, not less.

Usage

obfuscateLdBlocks(ldBlockResult, map)

Arguments

ldBlockResult

data.frame as returned by markerLdBlock: at least an idsUsed column (comma-joined ids, or NA).

map

named character vector of aliases, keyed by the original id – the map element of obfuscatePed(..., map = TRUE)'s return value.

Value

ldBlockResult with idsUsed ids replaced by their aliases (comma-joined, same order); a row whose idsUsed is NA is returned unchanged. Every other column is unchanged.

Details

A row whose idsUsed contains an id absent from map stop()s rather than silently dropping or leaking the real id – markerLdBlock's own founderIds contract should already guarantee every id is a valid pedigree id, so this is a defensive check, not the primary validation path.

See also

Other obfuscation: mapIdsToObfuscated(), obfuscateDate(), obfuscateGenomicROH(), obfuscateGenotypeMatrix(), obfuscateId(), obfuscatePed(), obfuscateTwinRelations()

Examples

library(nprcgenekeepr)
ped <- data.frame(
  id = c("F1", "F2", "S1", "S2"),
  sire = c(NA, NA, "F1", "F1"),
  dam = c(NA, NA, "F2", "F2"),
  sex = c("M", "F", "F", "F"),
  stringsAsFactors = FALSE
)
ldBlockResult <- data.frame(
  locus1 = "L1", locus2 = "L2", chrom = "1", Dprime = 0.5, r2 = 0.3,
  nUsed = 2L, idsUsed = "F1,F2", caveat = "Descriptive statistic only.",
  stringsAsFactors = FALSE
)
obfuscated <- obfuscatePed(ped, map = TRUE)
obfuscateLdBlocks(ldBlockResult, obfuscated$map)
#>   locus1 locus2 chrom Dprime  r2 nUsed       idsUsed
#> 1     L1     L2     1    0.5 0.3     2 XNZZUQ,QKV6EP
#>                        caveat
#> 1 Descriptive statistic only.