Patient-level test of the between-sample composition association

For each index cell type \(k\) and niche cell type \(n\), regresses the per-sample pseudobulk log2-CPM of every gene in the type-\(k\) cells on the per-sample mean log1p niche density of type \(n\) around those cells, across samples, with a limma moderated \(t\)-test. With a condition the design is ~ niche * condition + covariates and two terms are reported: "niche", the association pooled across conditions, and "condition:niche", its difference between conditions – the patient-level counterpart of the CellType:condition:niche term that fitSpiDE() tests within samples. Without a condition only "niche" is reported.

Usage

compositionTest(spe, ...)

# S4 method for class 'ANY'
compositionTest(
  spe,
  condition = NULL,
  sigma,
  index = NULL,
  niche = NULL,
  covariates = character(),
  assay = "counts",
  cell_type = "cell_type",
  sample_id = "sample_id",
  name = "Niche",
  min.cells = 10L,
  prior.count = 1,
  verbose = TRUE
)

Arguments

spe

a SpatialExperiment with niche reducedDims (see buildNiches()).

...

further arguments passed to the method.

condition

a character, the colData column of the condition (constant within sample), or NULL.

sigma

a numeric, the bandwidth (one value).

index, niche

character vectors restricting the index / niche cell types (NULL = all). An index type is never tested against its own niche.

covariates

a character vector of sample-level colData columns to adjust for (constant within sample; the first non-missing value per sample is used).

assay

a character, the counts assay.

cell_type, sample_id

the colData columns of cell type and sample.

name

the niche reducedDim prefix.

min.cells

minimum cells of the index type a sample must contribute.

prior.count

the pseudocount in the log2-CPM.

verbose

report progress.

Value

a data.frame with one row per (gene, index type, niche type, term): gene, ct_index, ct_niche, term, coef (log2-CPM per unit log1p density), t, p, fdr (BH within each (index, niche, term) over genes), fdr.global (BH over every row), and n_samples.

Details

This is the association that fitSpiDE()'s nested (sample x cell type) intercept (re.celltype = TRUE) deliberately absorbs. It is a between-patient effect with \(S\) experimental units; it is not neighbourhood-dependent differential expression, and the two must not be conflated (see the model vignette). Samples contributing fewer than min.cells cells of the index type are dropped for that index type, and an index type with fewer than three remaining samples is skipped.

Examples

data(toySpiDE)
spe <- buildNiches(toySpiDE, sigma = 20)
ct <- compositionTest(spe, condition = "condition", sigma = 20)
#> compositionTest: A x B (6 samples)
#> compositionTest: A x C (6 samples)
#> compositionTest: B x A (6 samples)
#> compositionTest: B x C (6 samples)
#> compositionTest: C x A (6 samples)
#> compositionTest: C x B (6 samples)
head(ct[order(ct$p), ])
#>     gene ct_index ct_niche            term       coef         t           p
#> 84    G4        B        A           niche -133.40784 -3.273716 0.002604489
#> 205   G5        C        B           niche  101.13068  2.985133 0.004817548
#> 91   G11        B        A           niche -130.08615 -3.053832 0.004876956
#> 67    G7        A        C condition:niche   43.63017  4.685958 0.006620484
#> 111  G11        B        A condition:niche  127.12975  2.896307 0.007200383
#> 165   G5        C        A           niche  -16.41544 -2.775646 0.008338056
#>     n_samples        fdr fdr.global
#> 84          6 0.04876956   0.288528
#> 205         6 0.09635095   0.288528
#> 91          6 0.04876956   0.288528
#> 67          6 0.08415400   0.288528
#> 111         6 0.14400766   0.288528
#> 165         6 0.16676113   0.288528