estDisp: Estimate the overdispersion

In RCM: Fit row-column association models with the negative binomial distribution for the microbiome

Description

Estimate the overdispersion

Usage

estDisp(
  X,
  cMat = NULL,
  rMat = NULL,
  muMarg,
  psis,
  trended.dispersion = NULL,
  prior.df = 10,
  dispWeights = NULL,
  rowMat = NULL,
  allowMissingness = FALSE,
  naId
)

Arguments

X	the data matrix of dimensions nxp
cMat	a 1xp colum scores matrix
rMat	a nx1 rowscores matrix, if unconstrained
muMarg	an nxp offset matrix
psis	a scalar, the current psi estimate
trended.dispersion	a vector of length p with pre-calculated trended.dispersion estimates. They do not vary in function of the offset anyway
prior.df	an integer, number of degrees of freedom of the prior for the Bayesian shrinkage
dispWeights	Weights for estimating the dispersion in a zero-inflated model
rowMat	matrix of row scores in case of constrained ordination
allowMissingness	A boolean, are missing values present
naId	The numeric index of the missing values in X

Details

Information between taxa is shared with empirical Bayes using the edgeR pacakage, where the time-limiting steps are programmed in C.

Value

A vector of length p with dispersion estimates

RCM documentation built on Nov. 8, 2020, 5:22 p.m.