Get the module membership from a BulkModuleResult
get_modules.RdReturns the data.table of gene to module_id assignments. The exact
columns depend on the method that produced the result (CoReMo adds
sign and stability; NMF/ICA/DGRDL add loading, sign and the
thresholding score; Leiden adds only module_id).
gene is not a unique key for the matrix factorisation methods. ICA, NMF
and DGRDL assign membership by keeping the tails of each component's loading
distribution, so a gene loading strongly on three components appears in three
rows, and a gene in no tail appears in none. That is the point of a
factorisation. The partition-based methods (CoReMo, Leiden) do emit one row
per gene. Do not assume uniqueness without checking method.
Examples
# gene to module assignments from an NMF fit
syn <- synthetic_bulk_cor_matrix()
mat <- log1p(t(syn$counts))
meta <- data.table::data.table(sample_id = rownames(mat))
object <- BulkCoExp(raw_data = mat, meta_data = meta)
object <- preprocess_bulk_coexp(object, hvg = 500L, .verbose = FALSE)
object <- nmf_bulk(object, k = 3L, .verbose = FALSE)
res <- S7::prop(object, "final_results")
head(get_modules(res))
#> gene module_id loading sign z
#> <char> <char> <num> <char> <num>
#> 1: gene_107 comp_01 42.79332 pos 3.774586
#> 2: gene_240 comp_02 29.67022 pos 9.529616
#> 3: gene_201 comp_02 27.39015 pos 8.528474
#> 4: gene_269 comp_02 26.93528 pos 8.328749
#> 5: gene_291 comp_02 26.75724 pos 8.250575
#> 6: gene_288 comp_02 26.25955 pos 8.032049