zinbModel: Initialize an object of class ZinbModel

In zinbwave: Zero-Inflated Negative Binomial Model for RNA-Seq Data

Description

Initialize an object of class ZinbModel

Usage

zinbModel(
  X,
  V,
  O_mu,
  O_pi,
  which_X_mu,
  which_X_pi,
  which_V_mu,
  which_V_pi,
  W,
  beta_mu,
  beta_pi,
  gamma_mu,
  gamma_pi,
  alpha_mu,
  alpha_pi,
  zeta,
  epsilon,
  epsilon_beta_mu,
  epsilon_gamma_mu,
  epsilon_beta_pi,
  epsilon_gamma_pi,
  epsilon_W,
  epsilon_alpha,
  epsilon_zeta,
  epsilon_min_logit,
  n,
  J,
  K
)

Arguments

X	matrix. The design matrix containing sample-level covariates, one sample per row.
V	matrix. The design matrix containing gene-level covariates, one gene per row.
O_mu	matrix. The offset matrix for mu.
O_pi	matrix. The offset matrix for pi.
which_X_mu	integer. Indeces of which columns of X to use in the regression of mu.
which_X_pi	integer. Indeces of which columns of X to use in the regression of pi.
which_V_mu	integer. Indeces of which columns of V to use in the regression of mu.
which_V_pi	integer. Indeces of which columns of V to use in the regression of pi.
W	matrix. The factors of sample-level latent factors.
beta_mu	matrix or NULL. The coefficients of X in the regression of mu.
beta_pi	matrix or NULL. The coefficients of X in the regression of pi.
gamma_mu	matrix or NULL. The coefficients of V in the regression of mu.
gamma_pi	matrix or NULL. The coefficients of V in the regression of pi.
alpha_mu	matrix or NULL. The coefficients of W in the regression of mu.
alpha_pi	matrix or NULL. The coefficients of W in the regression of pi.
zeta	numeric. A vector of log of inverse dispersion parameters.
epsilon	nonnegative scalar. Regularization parameter.
epsilon_beta_mu	nonnegative scalar. Regularization parameter for beta_mu.
epsilon_gamma_mu	nonnegative scalar. Regularization parameter for gamma_mu.
epsilon_beta_pi	nonnegative scalar. Regularization parameter for beta_pi.
epsilon_gamma_pi	nonnegative scalar. Regularization parameter for gamma_pi.
epsilon_W	nonnegative scalar. Regularization parameter for W.
epsilon_alpha	nonnegative scalar. Regularization parameter for alpha (both alpha_mu and alpha_pi).
epsilon_zeta	nonnegative scalar. Regularization parameter for zeta.
epsilon_min_logit	scalar. Minimum regularization parameter for parameters of the logit model, including the intercept.
n	integer. Number of samples.
J	integer. Number of genes.
K	integer. Number of latent factors.

Details

This is a wrapper around the new() function to create an instance of class ZinbModel. Rarely, the user will need to create a ZinbModel object from scratch, as tipically this is the result of zinbFit.

If any of X, V, W matrices are passed, n, J, and K are inferred. Alternatively, the user can specify one or more of n, J, and K.

The regularization parameters can be set by a unique parameter epsilon or specific values for the different regularization parameters can also be provided. If only epsilon is specified, the other parameters take the following values:

• epsilon_beta = epsilon/J

• epsilon_gamma = epsilon/n

• epsilon_W = epsilon/n

• epsilon_alpha = epsilon/J

• epsilon_zeta = epsilon

We empirically found that large values of epsilon provide a more stable estimation of W.

A call with no argument has the following default values: n = 50, J = 100, K = 0, epsilon=J.

Although it is possible to create new instances of the class by calling this function, this is not the most common way of creating ZinbModel objects. The main use of the class is within the zinbFit function.

Value

an object of class ZinbModel.

Examples

a <- zinbModel()
nSamples(a)
nFeatures(a)
nFactors(a)

zinbwave documentation built on Nov. 8, 2020, 8:11 p.m.