edge (version 2.4.2)

nbinomDeviance: Negative Binomial Deviance

Description

Fit the same log-link negative binomial or Poisson generalized linear model (GLM) to each row of a matrix of counts.

Usage

nbinomUnitDeviance(y, mean, dispersion=0)
nbinomDeviance(y, mean, dispersion=0, weights=NULL)

Arguments

y

numeric vector or matrix containing the negative binomial counts. If a matrix, then rows for genes and columns for libraries. nbinomDeviance treats a vector as a matrix with one row.

mean

numeric vector matrix of expected values, of same dimension as y.

dispersion

numeric vector or matrix of negative binomial dispersions. Can be a scalar, or a vector of length equal to the number of genes, or a matrix of same dimensions as y.

weights

numeric vector or matrix of non-negative weights, as for glmFit.

Value

• nbinomUnitDeviance returns a numeric vector or matrix of the same size as y.

  nbinomDeviance returns a numeric vector of length equal to the number of rows of y.

Details

nbinomUnitDeviance computes the unit deviance for each y observation. nbinomDeviance computes the total residual deviance for each row of y observation, i.e., weighted row sums of the unit deviances.

Care is taken to ensure accurate computation for small dispersion values.

References

Jorgensen, B. (2006). Generalized linear models. Encyclopedia of Environmetrics, Wiley. http://onlinelibrary.wiley.com/doi/10.1002/9780470057339.vag010/full.

McCarthy, DJ, Chen, Y, Smyth, GK (2012). Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Research 40, 4288-4297. http://nar.oxfordjournals.org/content/40/10/4288

Examples

y <- matrix(1:6,3,2)
mu <- matrix(3,3,2)
nbinomUnitDeviance(y,mu,dispersion=0.2)
nbinomDeviance(y,mu,dispersion=0.2)