Generate data

Usage

generate_data(
  nat_mat,
  family,
  nuisance_param_vec = NA,
  library_size_vec = 1,
  tol = 0.001
)

Arguments

nat_mat: An \(n\times p\) matrix of natural parameters, where \(n\) rows represent cells and \(p\) columns represent genes.
family: A character string, one of "gaussian", "exponential", "poisson", "neg_binom", "neg_binom2", "curved_gaussian", and "bernoulli".
nuisance_param_vec: Either NA or a single numeric or a length-\(p\) vector of numerics representing nuisance parameters. It is required if family is "gaussian", "curved_gaussian", "neg_binom" or "neg_binom2", and ignored otherwise.
library_size_vec: A single positive numeric (recycled to every cell; default 1) or a length-\(n\) vector of positive numerics.
tol: Small positive value to determine the smallest possible value in the output matrix, applied only when family is "gaussian" or "curved_gaussian".

Value

The generated data matrix, of the same dimension as nat_mat

Examples

set.seed(10)
x_mat <- matrix(stats::rnorm(20 * 2), nrow = 20, ncol = 2)
y_mat <- matrix(stats::rnorm(10 * 2), nrow = 10, ncol = 2)
nat_mat <- tcrossprod(x_mat, y_mat)
dat <- generate_data(nat_mat = nat_mat, family = "poisson")
dim(dat)
#> [1] 20 10