xps (version 1.32.0)

fitQC: Functions for fitting probe-level models

Description

This function allows to combine different algorithms to compute background correction, normalization and fit a multichip model for summarization.

Usage

fitQC(xps.data, filename = character(), filedir = getwd(), tmpdir = "", update = FALSE, # background correction bgcorrect.method  = "rma", bgcorrect.select  = "none", bgcorrect.option  = "pmonly:epanechnikov", bgcorrect.params  = c(16384), # normalization normalize.method  = "quantile", normalize.select  = "pmonly", normalize.option  = "transcript:together:none", normalize.logbase = "0", normalize.params  = c(0.0), # quality control qualify.method    = "rlm", qualify.select    = "pmonly", qualify.qualopt   = "all", qualify.option    = "transcript", qualify.estimator = "huber", qualify.logbase   = "log2", qualify.params    = list(10, 0.01, 1.0), # reference values reference.index   = 0, reference.method  = "mean", reference.params  = list(0.0), # misc. exonlevel  = "", xps.scheme = NULL, add.data   = FALSE, bufsize    = 32000, verbose    = TRUE)
xpsQualityControl(object, ...)

Arguments

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.

update

logical. If TRUE the existing ROOT data file filename will be updated.

bgcorrect.method

background method to use.

bgcorrect.select

type of probes to select for background correction.

bgcorrect.option

type of background correction to use.

bgcorrect.params

vector of parameters for background method.

normalize.method

normalization method to use.

normalize.select

type of probes to select for normalization.

normalize.option

normalization option.

normalize.logbase

logarithm base as character, one of ‘0’, ‘log’, ‘log2’, ‘log10’.

normalize.params

vector of parameters for normalization method.

qualify.method

qualification method to use, currently rlm.

qualify.select

type of probes to select for qualification.

qualify.qualopt

option determining the data to which to apply qualification, one of ‘raw’, ‘adjusted’, ‘normalized’, ‘all’.

qualify.option

option determining the grouping of probes for qualification, one of ‘transcript’, ‘exon’, ‘probeset’; exon arrays only.

qualify.estimator

option determining the M-estimator to use, one of ‘huber’, ‘fair’, ‘cauchy’, ‘gemanmcclure’, ‘welsch’, ‘tukey’, ‘andrew’.

qualify.logbase

logarithm base as character, one of ‘0’, ‘log’, ‘log2’, ‘log10’.

qualify.params

vector of parameters for qualification method.

reference.index

index of reference tree to use, or 0.

reference.method

for refindex=0, either trimmed mean or median of trees.

reference.params

vector of parameters for reference method.

exonlevel

exon annotation level determining which probes should be used for summarization; exon/genome arrays only.

xps.scheme

optional alternative SchemeSet.

add.data

logical. If TRUE expression data will be included as slot data.

bufsize

integer which sets the buffer size of the tree branch baskets (default is 32000).

verbose

logical, if TRUE print status information.

object

object of class DataTreeSet.

...

the arguments described above.

Value

An object of type QualTreeSet.

Details

This function allows to combine different algorithms to compute background correction, normalization and fit a multichip model for summarization.

xpsQualityControl is the DataTreeSet method called by function fitQC, containing the same parameters.

See Also

fitRLM, qualify, express

Examples

## Not run: 
# ## load existing ROOT scheme file and ROOT data file
# scheme.test3 <- root.scheme(paste(path.package("xps"),"schemes/SchemeTest3.root",sep="/"))
# data.test3 <- root.data(scheme.test3, paste(path.package("xps"),"rootdata/DataTest3_cel.root",sep="/"))
# 
# ## qualification - rlm
# rlm.all <- fitQC(data.test3, "tmp_Test3RLMall", filedir=getwd(), tmpdir="",
#                  qualify.method="rlm", qualify.qualopt="all", qualify.option="transcript", add.data=FALSE)
# 
# ## get expression data.frame
# expr.rlm.all <- validData(rlm.all)
# 
# ## get borders
# brd.rlm.all <- borders(rlm.all)
# 
# ## get residuals
# res.rlm.all <- residuals(rlm.all)
# 
# ## get weights
# w.rlm.all <- weights(rlm.all)
# 
# ## plot expression levels
# if (interactive()) {
# coiplot(rlm.all)
# borderplot(rlm.all)
# nuseplot(rlm.all)
# rleplot(rlm.all)
# image(rlm.all, type="resids")
# }
# ## End(Not run)