Align and validate sample correspondence between data and metadata
align_data_and_metadata.RdIn many quantitative proteomics workflows, mass spectrometry sample files are
named using numeric suffixes (e.g. Sample1, Sample2, ..., Sample10).
When read into R or other languages, these can be automatically sorted
lexicographically (Sample1, Sample10, Sample11, Sample2, ...),
causing incorrect sample-to-group mapping and misaligned metadata.
This helper ensures that the data and metadata objects are
correctly matched and ordered before creating a DAList. It automatically
fixes ordering issues, groups replicates of the same condition together, and
performs natural ("human") sorting of sample names containing numbers.
Usage
align_data_and_metadata(
data,
metadata,
sample_col = NULL,
group_col = NULL,
prefer_group_blocks = TRUE,
strict = FALSE
)Arguments
- data
A numeric matrix or data frame of protein intensities, where rows are proteins and columns are samples.
- metadata
A data frame of sample metadata, one row per sample.
- sample_col
Optional name of the column in
metadatathat contains the sample identifiers. If NULL, row names are used.- group_col
Optional column name in
metadataused to group replicates together (e.g. "group", "condition").- prefer_group_blocks
Logical; if TRUE (default) and
group_colis supplied, bothdataandmetadatawill be reordered to keep replicates of the same group contiguous and sorted naturally. If FALSE, only the metadata will be reordered to match data columns.- strict
Logical; if TRUE, stops with an error if sample orders differ. If FALSE (default), automatically reorders when sets match.
Value
A list with three elements:
- data
The reordered numeric intensity data.
- metadata
The reordered sample metadata.
- changes
A data frame summarizing the old vs new order of samples.
Details
The function performs three major checks:
Ensures the same sample IDs appear in both
dataandmetadata.Detects and removes any ordering mismatches.
(Optionally) groups replicates of the same experimental condition together using natural sorting.
If sample names differ only by order, the function reorders them safely. If samples are missing or duplicated, it stops with a descriptive error.
Examples
if (FALSE) { # \dontrun{
# Example data with misordered samples
dat <- data.frame(
Sample1 = c(10, 20, 30),
Sample10 = c(11, 21, 31),
Sample2 = c(12, 22, 32)
)
rownames(dat) <- paste0("Prot", 1:3)
meta <- data.frame(
sample = c("Sample1", "Sample2", "Sample10"),
group = c("A", "A", "B")
)
rownames(meta) <- meta$sample
# Align automatically, grouping by condition
aligned <- align_data_and_metadata(
data = dat,
metadata = meta,
sample_col = "sample",
group_col = "group",
prefer_group_blocks = TRUE
)
# Check results
aligned$changes
head(aligned$data)
head(aligned$metadata)
} # }