DirichletMultinomial (version 1.14.0)

dmn: Fit Dirichlet-Multinomial models to count data.

Description

Fit Dirichlet-Multinomial models to a sample x taxon count matrix.

Usage

dmn(count, k, verbose = FALSE, seed = runif(1, 0, .Machine$integer.max))

Arguments

count

matrix() of sample x taxon counts.

k

integer(1), the number of Dirichlet components to fit.

verbose

logical(1) indicating whether progress in fit should be reported.

seed

numeric(1) random number seed.

Value

An object of class dmn, with elements (elements are usually retrieved via functions defined in the package, not directly).

GoodnessOfFit

NLE, LogDet, Laplace, AIC, and BIC criteria assessing goodness-of-fit.

Group

matrix of dimension samples x k, providing the Dirichlet parameter vectors.

Mixture

Weight

numeric() of length k, with relative weight of each component.

Fit

Lower

matrix() of dimension taxa x k with 95% lower bounds on Dirichlet component vector estimates.

Estimate

matrix() of dimension taxa x k with Dirichlet component vector estimates.

Upper

matrix() of dimension taxa x k with 95% upper bounds on Dirichlet component vector estimates.

Details

This implements Dirichlet-multinomial mixture models describe in the package help page, DirichletMultinomial-package.

References

Holmes I, Harris K, Quince C, 2012 Dirichlet Multinomial Mixtures: Generative Models for Microbial Metagenomics. PLoS ONE 7(2): e30126. doi:10.1371/journal.pone.0030126.

See Also

DirichletMultinomial-package, vignette("DirichletMultinomial")

Examples

data(fit)
## k = 1:7; full example in vignette
lplc <- sapply(fit, laplace)
plot(lplc, type="b")
fit[[which.min(lplc)]]