Check a wide-format MHC haplotype designation file
Source:R/checkMhcHaplotypeFile.R
checkMhcHaplotypeFile.RdValidates the structure of a wide per-animal MHC haplotype designation
table: exactly three columns – id plus the animal's two named
haplotype designations, one row per animal. This is the explicit input
designation for MHC haplotype reporting (issue #148): a curator
designates data as MHC by supplying it through this format, and no MHC
semantics are ever inferred from locus or column names (the format has
no locus names). A new, sibling input family: the long-format marker
validators (checkMarkerGenotypeFile,
checkLinkageMarkerGenotypeFile,
checkSequenceGenotypeFile) and the single-locus
checkGenotypeFile path are untouched by this one.
Value
The genotype dataframe, checked to ensure the column count,
first-column identity, and row uniqueness are all valid. The returned
dataframe has its column names forced to c("id", "haplotype1",
"haplotype2").
Details
Column names are not prescribed: the first column must be id-like
(case-insensitive match on "id") and is forced to id; the
second and third are forced to haplotype1 and haplotype2,
so a colony file with, e.g., first_name/second_name
headers (the bundled rhesusGenotypes shape) loads
unchanged. NA/empty cells pass validation – a missing call is
the statistics layer's concern, not a structural error. A trailing
? on a designation (an uncertain, provisional call) is likewise
not validation's business; see the issue #148 design plan's D3 parse
rule. Haplotype designations are otherwise opaque labels: never parsed,
split, or matched against MHC region names.
Examples
library(nprcgenekeepr)
checked <- checkMhcHaplotypeFile(rhesusGenotypes)
head(checked)
#> id haplotype1 haplotype2
#> 1 I67LRJ A004_B002 A004_B048a
#> 2 K0M2RD A004_B012b A008_B017a
#> 3 ZPVN1V A008_B015b A002a_B069a
#> 4 0F4FY1 A004_B012b A008_B015b?
#> 5 ZHVYYN A007_B008 A019_B017a
#> 6 K1HTZ8 A004_B012b A006_B024a