dprobdropout: Density of the probabilistic dropout distribution

In const-ae/proDD: Identifying Differentially Abundant Proteins from Label-Free Mass Spectrometry Data

Description

Calculate the density under a censoring mechanism that probabilistically causes dropouts described with rho and zeta. The data is drawn from a normal with mu and sigma2 * nobs, but each value drops out with probability according to a inverse probit with its center at rho and width zeta:

p(x | μ, σ2, ρ, ζ) ~ Normal(x; μ, σ2) Prod_i Φ(x; ρ_i, ζ_i)

The distribution is related to the extended skewed normal distribution and formally belongs to the class of closed skewed normals.

Usage

dprobdropout(x, mu, sigma2, rho = numeric(0), zeta = numeric(0),
  nmis = length(rho), log = FALSE)

Arguments

x	vector of input positions.
mu	double. The mean of the observed values.
sigma2	double. The variance of the mu estimate.
rho	vector. The positions of the inflection points of the dropout curves for each sample. Can also be a single number that is repeated nmis times. Defaults to an empty vector.
zeta	vector. The scale of the dropout curves for each sample. Can also be a single number that is repeated nmis times. Defaults to an empty vector.
nmis	integer The number of missing values. Defaults to length(rho).
log	boolean if the log of the density is returned.

References

1. Azzalini, A. & Capitanio, A. The Skew-Normal and Related Families. (Cambridge University Press, 2013). doi:10.1017/CBO97811392488911.

2. González-Farías, G., Domínguez-Molina, A. & Gupta, A. K. Additive properties of skew normal random vectors. J. Stat. Plan. Inference 126, 521–534 (2004). doi:10.1016/j.jspi.2003.09.008

Examples

  xg <- seq(-5, 5, length.out=101)
  plot(xg, dprobdropout(xg, mu=0, sigma2=3, rho=0, zeta=1))

const-ae/proDD documentation built on Jan. 14, 2020, 9:34 a.m.