GladModel: The GladModel class

Description

Package: aroma.core Class GladModel Object ~~| ~~+--ChromosomalModel ~~~~~~~| ~~~~~~~+--CopyNumberChromosomalModel ~~~~~~~~~~~~| ~~~~~~~~~~~~+--CopyNumberSegmentationModel ~~~~~~~~~~~~~~~~~| ~~~~~~~~~~~~~~~~~+--GladModel Directly known subclasses: public static class GladModel extends CopyNumberSegmentationModel This class represents the Gain and Loss Analysis of DNA regions (GLAD) model [1]. This class can model chip-effect estimates obtained from multiple chip types, and not all samples have to be available on all chip types.

Usage

GladModel(cesTuple=NULL, ...)

Arguments

cesTuple

A CopyNumberDataSetTuple.

...

Arguments passed to the constructor of CopyNumberSegmentationModel.

Fields and Methods

Methods: rll{ getFitFunction - writeRegions - } Methods inherited from CopyNumberSegmentationModel: fit, getAsteriskTags, getFitFunction, getFullNames, getRegions, getTags, plot, plotCopyNumberRegionLayers, writeRegions Methods inherited from CopyNumberChromosomalModel: as.character, calculateChromosomeStatistics, calculateRatios, clearCache, estimateSds, extractRawCopyNumbers, fit, getChromosomeLength, getDataFileMatrix, getNames, getOptionalArguments, getPairedNames, getReferenceSetTuple, getRefSetTuple, isPaired, newPlot, plotAxesLayers, plotChromosomesLayers, plotCytobandLayers, plotFitLayers, plotGridHorizontalLayers, plotRawCopyNumbers, plotSampleLayers Methods inherited from ChromosomalModel: as.character, clearCache, fit, getAlias, getAromaGenomeTextFile, getAsteriskTags, getChipType, getChipTypes, getChromosomes, getFullName, getFullNames, getGenome, getGenomeData, getGenomeFile, getListOfAromaUgpFiles, getName, getNames, getParentPath, getPath, getReportPath, getRootPath, getSets, getSetTuple, getTags, indexOf, nbrOfArrays, nbrOfChipTypes, setGenome 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

Benchmarking

In high-density copy numbers analysis, the most time consuming step is fitting the GLAD model. The complexity of the model grows more than linearly (squared? exponentially?) with the number of data points in the chromosome and sample being fitted. This is why it take much more than twice the time to fit two chip types together than separately.

Details

Data from multiple chip types are combined "as is". This is based on the assumption that the relative chip effect estimates are non-biased (or at the equally biased across chip types). Note that in GLAD there is no way to down weight certain data points, which is why we can control for differences in variance across chip types.

References

[1] Hupe P et al. Analysis of array CGH data: from signal ratio to gain and loss of DNA regions. Bioinformatics, 2004, 20, 3413-3422.