Generates simulated bulk or spatial transcriptomics mixtures from a single-cell reference, following the simulation logic used in the original DISSECT workflow. Returned proportions are stored as samples x cell types.
Usage
dissect_simulate(
sc_data,
celltype_col = "celltype",
batch_col = NULL,
type = "bulk",
n_samples = NULL,
cells_per_sample = 500,
prop_sparse = 0.5,
concentration = NULL,
save_expr = TRUE,
min_genes = 200,
min_cells = 3,
mt_cutoff = 5,
min_expr = 0,
downsample = NULL,
seed = 42
)Arguments
- sc_data
A `Seurat` or `SingleCellExperiment` object containing single-cell counts.
- celltype_col
Character scalar. Column name in cell metadata containing cell-type labels.
- batch_col
Character scalar or `NULL`. Optional column name in cell metadata containing batch labels. If multiple batches are present, simulation is performed per batch and concatenated.
- type
Character scalar. Either `"bulk"` or `"st"`.
- n_samples
Integer scalar or `NULL`. Number of simulated samples. If `NULL`, defaults to `1000 * n_celltypes`.
- cells_per_sample
Integer scalar. Number of cells per simulated sample for bulk simulation.
- prop_sparse
Numeric scalar in `[0, 1]`. Fraction of sparse simulated bulk samples.
- concentration
Numeric vector or `NULL`. Dirichlet concentration parameter for bulk simulation. If `NULL`, a vector of ones is used.
- save_expr
Logical scalar. If `TRUE`, stores per-cell-type expression contributions for each simulated sample.
- min_genes
Integer scalar. Minimum number of detected genes required for a cell to be kept during preprocessing.
- min_cells
Integer scalar. Minimum number of cells required for a gene to be kept during preprocessing.
- mt_cutoff
Numeric scalar. Maximum allowed mitochondrial percentage per cell.
- min_expr
Numeric scalar. Minimum mean `log1p` expression threshold
- downsample
Numeric scalar or `NULL`. Optional downsampling factor for spatial transcriptomics simulation.
- seed
Integer scalar. Random seed used for simulation.
Value
A named list with components:
- X
A simulated expression matrix with samples in rows and genes in columns.
- props
A proportions matrix with samples in rows and cell types in columns.
- cells
An integer matrix of sampled cell counts with samples in rows and cell types in columns.
- gene_names
Character vector of gene names.
- celltype_names
Character vector of cell-type names.
- layers
Named list of per-cell-type expression matrices, or `NULL` if `save_expr = FALSE`.
- batch
Character vector giving the simulated batch assignment of each sample.
- type
Simulation type, either `"bulk"` or `"st"`.
Details
The preprocessing and simulation logic are designed to follow the original DISSECT implementation as closely as possible while using in-memory R data structures instead of file-based Python objects.
References
Khatri, R., Machart, P., & Bonn, S. (2024). DISSECT: deep semi-supervised consistency regularization for accurate cell type fraction and gene expression estimation. Genome Biology, 25(1), 112.
Original DISSECT software repository: https://github.com/imsb-uke/DISSECT