firma: Finding Isoforms using Robust Multichip Analysis
In xps: Processing and Analysis of Affymetrix Oligonucleotide Arrays including Exon Arrays, Whole Genome Arrays and Plate Arrays

Description Usage Arguments Details Value Note Author(s) References Examples

This function converts a DataTreeSet for exon arrays into an ExprTreeSet using the Finding Isoforms using Robust Multichip Analysis (FIRMA).

firma(xps.data,
      filename   = character(0),
      filedir    = getwd(),
      tmpdir     = "",
      background = "antigenomic",
      normalize  = TRUE,
      option     = "probeset",
      exonlevel  = "metacore",
      method     = "mdp",
      params     = list(16384, 0.0, 1.0, 10, 0.01, 1.0),
      xps.scheme = NULL,
      add.data   = TRUE,
      verbose    = TRUE)

xpsFIRMA(object, ...)

`xps.data`	object of class `DataTreeSet`.
`filename`	file name of ROOT data file.
`filedir`	system directory where ROOT data file should be stored.
`tmpdir`	optional temporary directory where temporary ROOT files should be stored.
`background`	probes used to compute background, one of ‘genomic’, ‘antigenomic’
`normalize`	logical. If `TRUE` normalize data using quantile normalization.
`option`	option determining the grouping of probes for summarization, one of ‘exon’, ‘probeset’.
`exonlevel`	exon annotation level determining which probes should be used for summarization.
`method`	method to be used for summarization, currently ‘mdp’.
`params`	list of (default) parameters for `rma`.
`xps.scheme`	optional alternative `SchemeTreeSet`.
`add.data`	logical. If `TRUE` expression data will be included as slot `data`.
`verbose`	logical, if `TRUE` print status information.
`object`	object of class `DataTreeSet`.
`...`	the arguments described above.

This function computes FIRMA (Finding Isoforms using Robust Multichip Analysis) for detecting differential alternative splicing for exon arrays, as described in Purdom et al.

Following options are valid for exon arrays:

`probeset`:	expression levels are computed for individual probe sets, i.e. for each 'probeset_id'.
`exon`:	expression levels are computed for exon clusters, i.e. probe sets containing the same 'exon_id', where each exon cluster consists of one or more `probeset`s.

Following exonlevel annotations are valid for exon arrays:

	`core`:	probesets supported by RefSeq and full-length GenBank transcripts.
	`metacore`:	core meta-probesets.
	`extended`:	probesets with other cDNA support.
	`metaextended`:	extended meta-probesets.
	`full`:	probesets supported by gene predictions only.
	`metafull`:	full meta-probesets.
	`ambiguous`:	ambiguous probesets only.
	`affx`:	standard AFFX controls.
	`all`:	combination of above (including affx).

Exon levels can also be combined, with following combinations being most useful:

`exonlevel="metacore+affx"`:	core meta-probesets plus AFFX controls
`exonlevel="core+extended"`:	probesets with cDNA support
`exonlevel="core+extended+full"`:	supported plus predicted probesets

Exon level annotations are described in the Affymetrix whitepaper exon_probeset_trans_clust_whitepaper.pdf:
“Exon Probeset Annotations and Transcript Cluster Groupings”.

Method xpsFIRMA is the DataTreeSet method called by function firma, containing the same parameters.

An ExprTreeSet

In contrary to other implementations of (FI)RMA the expression measure of FIRMA is given in linear scale, analogously to the expression measures computed with mas5 and mas4.

Please note that the current implementation of FIRMA is based on median-polish only, see: http://www.aroma-project.org/node/81

Please note that the default settings of params gives results which are identical to the results obtained with APT (Affymetrix Power Tools) and with package affy_1.14.2 or earlier. If you want to obtain results which are identical to the results obtained with affy_1.16.0 or later then you need to set params = list(16384, 0.0, 0.4, 10, 0.01, 1.0).

By setting parameter background="none" it is possible to skip background correction .

For the analysis of many exon arrays it may be better to define a tmpdir, since this will store only the results in the main file and not e.g. background and normalized intensities, and thus will reduce the file size of the main file. For quantile normalization memory should not be an issue, however medianpolish depends on RAM unless you are using a temporary file.

Parameter exonlevel determines not only which probes are used for medianpolish, but also the probes used for background calculation and for quantile normalization. If you want to use seperate probes for background calculation, quantile normalization and medianpolish summarization, you can pass a numeric vector containing three integer values corresponding to the respective exonlevel, e.g. you can use exonlevel=c(16316,8252,8252), see function exonLevel for more details.

Christian Stratowa

Purdom, E., Simpson K.M., Robinson M.D., Conboy J.G., Lapuk A.V. and Speed, T.P. (2008), FIRMA: a method for detection of alternative splicing from exon array data. Bioinformatics 24(15):1707-1714

## Not run: 
## load ROOT scheme file
scmdir <- "/Volumes/GigaDrive/CRAN/Workspaces/Schemes"
scheme.exon <- root.scheme(paste(scmdir,"Scheme_HuEx10stv2r2_na27.root",sep="/"))

## load subset of ROOT data file
datdir <- "/Volumes/GigaDrive/CRAN/Workspaces/ROOTData"
subnames <- c("HeartA","HeartB","HeartC", "MuscleA","MuscleB","MuscleC")
sub.exon <- root.data(scheme.exon, rootFile(data.exon), celnames=subnames)

## firma
outdir <- getwd()
sub.firma.ps <- firma(sub.exon,"HeartMuscleFIRMAcorePS",filedir=outdir,tmpdir="",background="antigenomic",
                      normalize=TRUE,option="probeset",exonlevel="core")

## get transcript expression levels for all transcripts or transcript=2429277
expr.firma <- firma.expr(sub.firma.ps, probeset=NULL, option="transcript")
expr.firma <- firma.expr(sub.firma.ps, probeset=2429277, option="transcript")

## get probeset expression levels for all probeset or probeset=2429278 or transcript=2429277
expr.firma <- firma.expr(sub.firma.ps, probeset=NULL, option="probeset")
expr.firma <- firma.expr(sub.firma.ps, probeset=2429278, option="probeset")
expr.firma <- firma.expr(sub.firma.ps, probeset=2429277, option="probeset")

## get probeset splice scores for all probeset or probeset=2429278 or transcript=2429277
score.firma <- firma.score(sub.firma.ps, probeset=NULL, option="probeset")
score.firma <- firma.score(sub.firma.ps, probeset=2429278, option="probeset")
score.firma <- firma.score(sub.firma.ps, probeset=2429277, option="probeset")

## different plots
boxplot(sub.firma.ps, which="UnitName:LEVEL_PS")
boxplot(sub.firma.ps, which="UnitName:LEVEL_TS")

hist(sub.firma.ps, which="UnitName:LEVEL_PS")
hist(sub.firma.ps, which="UnitName:LEVEL_TS")

rleplot(sub.firma.ps, which="UnitName:LEVEL_PS")
rleplot(sub.firma.ps, which="UnitName:LEVEL_TS")

nuseplot(sub.firma.ps, which="UnitName:STDEV_PS")
nuseplot(sub.firma.ps, which="UnitName:STDEV_TS")

## End(Not run)