Run DIALOGUE over meta cells — rs_mc

The meta cell entry point into DIALOGUE, see rs_dialogue_sc() for what the method does. This is a shim rather than a second implementation: everything DIALOGUE asks of the expression matrix is per-gene, so the in-memory matrix is wrapped as a gene-major reader and the same core runs.

Only the normalised layer is ever read, so sparse_data has to carry the normalised counts. The data layer is cast to integers on the way in and then goes unused.

Meta cells are already aggregates, so the sample a meta cell belongs to has to be unambiguous: build them within samples, not across them. The random intercept in stage two is over samples, and a meta cell straddling two of them has no well-defined level.

Usage

rs_mc_dialogue(
  sparse_data,
  cell_type_indices,
  features,
  sample_ids,
  cell_quality,
  gene_indices,
  dialogue_params,
  verbose
)

Arguments

sparse_data: A named list that needs to have data, indptr, indices, cs_type, nrow and ncol, holding the normalised meta cell counts with shape (metacells, genes).
cell_type_indices: List of integer vectors. 0-indexed(!) positions of the meta cells belonging to each cell type. At least two cell types are needed.
features: List of numeric matrices, one per cell type, shaped n_metacells_in_type x k_i with rows aligned to cell_type_indices. Needs at least two columns per cell type.
sample_ids: Integer vector. 0-indexed(!) sample code per meta cell, over all meta cells rather than per cell type.
cell_quality: Numeric vector. Quality covariate per meta cell, indexed the same way as sample_ids.
gene_indices: Integer vector. 0-indexed(!) positions of the genes to consider when building signatures.
dialogue_params: Named list. Contains the DIALOGUE parameters across all three stages, see params_dialogue_pmd(), params_dialogue_hlm() and params_dialogue_refine(). The three blocks share one flat list.
verbose: Integer. 0L - quiet; 1L - normal verbosity; 2L - detailed verbosity.

Value

A list, identical in shape to rs_dialogue_sc().

References

Jerby-Arnon & Regev, Nature Biotechnology, 2022