rPLN {PLNmodels}

PLN RNG

Description

Random generation for the PLN model with latent mean equal to mu, latent covariance matrix equal to Sigma and average depths (sum of counts in a sample) equal to depths

Usage

rPLN(
  n = 10,
  mu = rep(0, ncol(Sigma)),
  Sigma = diag(1, 5, 5),
  depths = rep(10000, n)
)

Arguments

n

the sample size

mu

vectors of means of the latent variable

Sigma

covariance matrix of the latent variable

depths

Numeric vector of target depths. The first is recycled if there are not n values

Details

The default value for mu and Sigma assume equal abundances and no correlation between the different species.

Value

a n * p count matrix, with row-sums close to depths, with an attribute "offsets" corresponding to the true generated offsets (in log-scale).

Examples

## 10 samples of 5 species with equal abundances, no covariance and target depths of 10,000
rPLN()
## 2 samples of 10 highly correlated species with target depths 1,000 and 100,000
## very different abundances
mu <- rep(c(1, -1), each = 5)
Sigma <- matrix(0.8, 10, 10); diag(Sigma) <- 1
rPLN(n=2, mu = mu, Sigma = Sigma, depths = c(1e3, 1e5))