Generate null data — generate

This function does not allow much flexibility, and is meant primarily for simple null simulations. Here, `n` (the number of cells) is equal to `cell_per_person*num_individuals`.

Usage

generate_null(cell_per_person = 100, num_genes = 1000, num_individuals = 20)

Arguments

cell_per_person: The number of cells per individual
num_genes: The number of genes in the simulation
num_individuals: The number of individuals

Value

a list with covariates (a `matrix` with `n` rows (cells) and 5 columns, named `"Intercept"`, `"Log_UMI"`, `"CC"`, `"Sex"` and `"Age"`, ready for `initialize_esvd`; `"CC"` is the 0/1 case-control indicator), df (a `data.frame` with `num_individuals` rows and the 6 columns `"Intercept"`, `"Log_UMI"`, `"CC"`, `"Sex"`, `"Age"` and `"Individual"`), metadata_individual (a `factor` of length `n` naming each cell's individual), nat_mat (a `n` by `p` `matrix` of natural parameters), and obs_mat (a `n` by `p` `dgCMatrix` of observed counts). The first 10 genes carry a case-control effect; the rest are null.

Examples

set.seed(10)
sim <- generate_null(cell_per_person = 15, num_genes = 40,
                     num_individuals = 8)
dim(sim$obs_mat)
#> [1] 120  40
table(sim$metadata_individual)
#> 
#> indiv_1 indiv_2 indiv_3 indiv_4 indiv_5 indiv_6 indiv_7 indiv_8 
#>      15      15      15      15      15      15      15      15