divnet: divnet
In adw96/DivNet: Diversity Estimation in Networked Ecological Communities

divnet

R Documentation

divnet

Description

divnet

Usage

divnet(
  W,
  X = NULL,
  fitted_model = NULL,
  tuning = NULL,
  perturbation = NULL,
  network = NULL,
  base = NULL,
  ncores = NULL,
  variance = "parametric",
  B = 5,
  nsub = NULL,
  formula = NULL,
  ...
)

Arguments

`W`	An abundance table with taxa as columns and samples as rows; or a phyloseq object.
`X`	The covariate matrix, with samples as rows and variables as columns. Defaults to NULL (sample_names are the covariates). Instead of specifying `X`, you can specify this information using the argument `formula`. If you specify a `formula` and `W` is a phyloseq object, this argument will be ignored.
`fitted_model`	object produced by fit_aitchison. Defaults to NULL.
`tuning`	settings for tuning the MC-MH algorithm. Options include NULL (defaults to "fast"), "fast", "careful" or a named list with components EMiter (number of EM iterations; 6 for fast, 10 for careful), EMburn (number of EM iterations to burn; 3 for fast, 5 for careful), MCiter (number of MC iterations; 500 for fast, 1000 for careful), MCburn (number of MC iterations to burn; 250 for fast, 500 for careful) and stepsize (variance used for MH samples; 0.01 for both fast and careful)
`perturbation`	Perturbation magnitude for zero values when calculating logratios.
`network`	How to estimate network. Defaults to NULL (the default), "default" (generalised inverse, aka naive). Other options include "diagonal", "stars" (requires glasso and SpiecEasi to be installed), or a function that you want to use to estimate the network
`base`	The column index of the base taxon in the columns of W, or the name of the taxon (must be a column name of W, or a taxon name if W is a phyloseq object). If NULL, will use 'pick_base' to choose a taxon. If no taxa are observed in all samples, an error will be thrown. In that case, we recommend trying a number of different highly abundant taxa to confirm the results are robust to the taxon choice.
`ncores`	Number of cores to use for parallelization
`variance`	method to get variance of estimates. Current options are "parametric" for parametric bootstrap, "nonparametric" for nonparametric bootstrap, and "none" for no variance estimates
`B`	Number of bootstrap iterations for estimating the variance.
`nsub`	Number of subsamples for nonparametric bootstrap. Defaults to half the number of observed samples.
`formula`	an object of class `formula`: a symbolic description of the model to be fitted; a means of constructing `X` via `stats::model.matrix`. If `W` is a phyloseq object, the formula should refer to variables stored in sample_data. If `W` is not a phyloseq object, `X` should be a data frame containing columns referred to in your formula. Formula references must match column names found in the sample data from `W` or `X`. Optional, defaults to `NULL`.
`...`	Additional parameters to be passed to the network function