ProbeLevelModel: The ProbeLevelModel class

Description

Package: aroma.affymetrix
Class ProbeLevelModel

Directly known subclasses:
AffineCnPlm, AffinePlm, AffineSnpPlm, AvgCnPlm, AvgPlm, AvgSnpPlm, ExonRmaPlm, HetLogAddCnPlm, HetLogAddPlm, HetLogAddSnpPlm, MbeiCnPlm, MbeiPlm, MbeiSnpPlm, RmaCnPlm, RmaPlm, RmaSnpPlm

public abstract static class ProbeLevelModel
extends MultiArrayUnitModel

This abstract class represents a probe-level model (PLM) as defined by the affyPLM package: "A [...] PLM is a model that is fit to probe-intensity data. More specifically, it is where we fit a model with probe level and chip level parameters on a probeset by probeset basis", where the more general case for a probeset is a unit group in Affymetrix CDF terms.

Usage

ProbeLevelModel(..., standardize=TRUE)

Arguments

...: Arguments passed to MultiArrayUnitModel.
standardize: If TRUE, chip-effect and probe-affinity estimates are rescaled such that the product of the probe affinities is one.

Fields and Methods

Methods:

	`fit`	-
	`getChipEffectSet`	-
	`getProbeAffinityFile`	-
	`getResidualSet`	-
	`getWeightsSet`	-

Methods inherited from MultiArrayUnitModel:
getListOfPriors, setListOfPriors, validate

Methods inherited from UnitModel:
findUnitsTodo, getAsteriskTags, getFitSingleCellUnitFunction, getParameters

Methods inherited from Model:
as.character, fit, getAlias, getAsteriskTags, getDataSet, getFullName, getName, getPath, getRootPath, getTags, setAlias, setTags

Methods inherited from ParametersInterface:
getParameterSets, getParameters, getParametersAsString

Methods inherited from Object:
$, $<-, [[, [[<-, as.character, attach, attachLocally, clearCache, clearLookupCache, clone, detach, equals, extend, finalize, getEnvironment, getFieldModifier, getFieldModifiers, getFields, getInstantiationTime, getStaticInstance, hasField, hashCode, ll, load, names, objectSize, print, save, asThis

Author

Henrik Bengtsson

Details

In order to minimize the risk for mistakes, but also to be able compare results from different PLMs, all PLM subclasses must meet the following criteria:

All parameter estimates must be (stored and returned) on the intensity scale, e.g. log-additive models such as RmaPlm have to transform the parameters on the log-scale to the intensity scale.
The probe-affinity estimates $\phi_k$ for a unit group must be constrained such that $\prod_k \phi_k = 1$, or equivalently if $\phi_k > 0$,$\sum_k \log(\phi_k) = 0$.