AvgPlm: The AvgPlm class

Description

Package: aroma.affymetrix Class AvgPlm Object ~~| ~~+--Model ~~~~~~~| ~~~~~~~+--UnitModel ~~~~~~~~~~~~| ~~~~~~~~~~~~+--MultiArrayUnitModel ~~~~~~~~~~~~~~~~~| ~~~~~~~~~~~~~~~~~+--ProbeLevelModel ~~~~~~~~~~~~~~~~~~~~~~| ~~~~~~~~~~~~~~~~~~~~~~+--AvgPlm Directly known subclasses: AvgCnPlm, AvgSnpPlm public static class AvgPlm extends ProbeLevelModel This class represents a PLM where the probe intensities are averaged assuming identical probe affinities. For instance, one may assume that replicated probes with identical sequences have the same probe affinities, cf. the GenomeWideSNP_6 chip type.

Usage

AvgPlm(..., flavor=c("median", "mean"))

Arguments

...

Arguments passed to ProbeLevelModel.

flavor

A character string specifying what model fitting algorithm to be used. This makes it possible to get identical estimates as other packages.

Fields and Methods

Methods: rll{ getAsteriskTags - } Methods inherited from ProbeLevelModel: calculateResidualSet, calculateWeights, fit, getAsteriskTags, getCalculateResidualsFunction, getChipEffectSet, getProbeAffinityFile, getResidualSet, getWeightsSet Methods inherited from MultiArrayUnitModel: getListOfPriors, setListOfPriors, validate Methods inherited from UnitModel: findUnitsTodo, getAsteriskTags, getFitSingleCellUnitFunction Methods inherited from Model: fit, getAlias, getAsteriskTags, getDataSet, getFullName, getName, getPath, getRootPath, getTags, setAlias, setTags Methods inherited from Object: asThis, getChecksum, $, $<-, [[, [[<-, as.character, attach, attachLocally, clearCache, clearLookupCache, clone, detach, equals, extend, finalize, gc, getEnvironment, getFieldModifier, getFieldModifiers, getFields, getInstantiationTime, getStaticInstance, hasField, hashCode, ll, load, objectSize, print, registerFinalizer, save

Model

For a single unit group, the averaging PLM of K probes is:

y_{i k} = θ_{i} + ε_{i k}

where $\theta_i$ are the chip effects for arrays $i=1,...,I$. The $\varepsilon_{ik}$ are zero-mean noise with equal variance.

Different flavors of model fitting

The above model can be fitted in two ways, either robustly or non-robustly. Use argument flavor="mean" to fit the model non-robustly, i.e.

{\hat{θ}}_{i} = 1 / K \sum_{k} y_{i k}

. Use argument flavor="median" to fit the model robustly, i.e.

{\hat{θ}}_{i} = m e d i a n_{k} y_{i k}