StatModel: The StatModel class for msqrob

In statOmics/msqrob2: Robust statistical inference for quantitative LC-MS proteomics

The StatModel class for msqrob

Description

The StatModel class contains a statistical model as applied on a feature.

Models are created by the dedicated user-level functions (msqrob(), mqrobAggregate()) or manually, using the StatModel() constructor. In the former case, each quantitative feature is assigned its statistical model and the models are stored as a variable in a DataFrame object, as illustred in the example below.

Function for constructing a new StatModel object.

Usage

## S4 method for signature 'StatModel'
show(object)

StatModel(
  type = "fitError",
  params = list(),
  varPosterior = NA_real_,
  dfPosterior = NA_real_
)

Arguments

object	StatModel object
type	default set to fit-error, can be a "lm", "rlm" (robust lm with M estimation), "lmer" (when mixed models or ridge regression is adopted), "quasibinomial" (when peptide counts are fitted)
params	A list containing the parameters of the fitted model
varPosterior	Numeric, posterior variance, default is NA
dfPosterior	Numeric, posterior degrees of freedom, default is NA

Value

A StatModel object

Slots

type

character(1) defining type of the used model. Default is "fitError", i.e. a error model. Other include "lm", "rlm", ...

params

A list() containing information of the used model.

varPosterior

numeric() of posterior variance.

dfPosterior

numeric() of posterior degrees of freedom.

Author(s)

Oliver M. Crook, Laurent Gatto, Lieven Clement

Examples

## A fully specified dummy model
myModel <- StatModel(
    type = "rlm",
    params = list(x = 3, y = 7, b = 4),
    varPosterior = c(0.1, 0.2, 0.3),
    dfPosterior = c(6, 7, 8)
)
myModel



## A collection of models stored as a variable in a DataFrame
mod1 <- StatModel(type = "rlm")
mod2 <- StatModel(type = "lm")
df <- DataFrame(x = 1:2)
df$mods <- c(mod1, mod2)
df
# TODO

statOmics/msqrob2 documentation built on April 23, 2024, 8:52 a.m.