QTL Dataset (individual-level data for one study)

S4 container for a single QTL study's regional data, extending MultiAssayExperiment. Every QTL context is one RangedSummarizedExperiment: rows are molecular traits positioned by rowRanges, columns are samples, and per-context phenotype covariates sit in that experiment's colData. Alongside them a genotype experiment carries variants in its rowRanges, dosages in a lazily-read DelayedArray assay, and genotype-derived covariates (e.g., ancestry PCs) in its own colData. The sampleMap records which samples each experiment observes, so contexts need not share a sample set.

Extending MultiAssayExperiment means the multi-assay surface applies directly: experiments(), colData(), sampleMap(), and two-dimensional subsetting (x[, samples, contexts]), all of which preserve the class and its own slots. The genotype assay stays unread until an operation touches it.

One caveat comes with that laziness: longForm() and wideFormat() fail on a delayed assay, with MultiAssayExperiment reshaping it to zero rows and reporting "replacement has 0 rows". Subsetting does not sidestep it, because selecting contexts keeps the genotype experiment. Reshape a context experiment on its own, or drop the genotype one by coercing first: longForm(as(x, "MultiAssayExperiment")[, , "brain"]). Reshaping genotypes is variants x samples and enormous on real data, so it is rarely what you want regardless; anything that materialises a tidy view of that experiment reads its dosages.

Usage

# S4 method for class 'QtlDataset,ANY,ANY,ANY'
x[i, j, k, ..., drop = FALSE]

# S4 method for class 'QtlDataset'
longForm(object, ..., genotype = FALSE)

Arguments

x: A QtlDataset.
i, j, k: Feature, sample and experiment subscripts, as for MultiAssayExperiment.
...: Passed on to row subsetting.
drop: Passed through to MultiAssayExperiment: when TRUE, experiments the subset leaves empty are removed.
object: A QtlDataset.
genotype: Logical (length 1), default FALSE. Include the genotype experiment, reading its dosages into memory to do so.

Value

A QtlDataset narrowed to the requested features, samples and contexts.

A long-format DataFrame, one row per (assay, primary, rowname) observation.

Functions

x = QtlDataset[i = ANY, j = ANY, drop = ANY]: Subset by feature, sample and experiment. Selecting experiments selects among contexts: the genotype experiment is the substrate every context is interpreted against rather than one of the things being chosen between, so it is always retained.
longForm(QtlDataset): Reshape the measurements into one long DataFrame. Only the QTL contexts are reshaped: the genotype experiment is the substrate they are interpreted against rather than a measurement, and being variants x samples it would dwarf them. Pass genotype = TRUE to include it, which reads the dosages into memory – MultiAssayExperiment cannot reshape a delayed assay, and fails with "replacement has 0 rows" if asked to.

wideFormat() has no such method: it is a plain function rather than a generic, and it reshapes the ExperimentList directly, so there is nothing to dispatch on. Coerce first – wideFormat(as(x, "MultiAssayExperiment")[, , contexts]).

Slots

study: Character (length 1). Study identifier; used in collection classes to tag downstream FineMappingResult / TwasWeights entries. The genotype experiment's assay reads through this handle; the extraction accessors read it directly, so that QC can be applied per block.
scaleResiduals: Logical (length 1). Whether residualization accessors scale residuals to unit variance.
mafCutoff: Numeric (length 1). Minor allele frequency threshold; variants with MAF < mafCutoff are dropped at extraction time inside getGenotypes() / getResidualizedGenotypes(). Default 0 (no filter).
macCutoff: Numeric (length 1). Minor allele count threshold; converted to a MAF threshold using max(mafCutoff, macCutoff / (2 * n)) where n is the post-narrowing sample count of the extracted block. Default 0 (no filter).
xvarCutoff: Numeric (length 1). Per-variant genotype variance threshold; variants with column variance below this are dropped at extraction time. Default 0 (no filter).
imissCutoff: Numeric (length 1). Per-sample genotype-missingness threshold; samples with a missing-genotype rate above this are dropped at extraction time. Default 0 (no filter).
keepVariants: Character vector of variant identifiers to retain prior to per-block QC. Length 0 means no restriction.
keepIndel: Logical (length 1). When FALSE, indel variants (alleles that are not single nucleotides) are dropped at extraction time. Default TRUE.