Package 'oligo'

Title: Preprocessing tools for oligonucleotide arrays
Description: A package to analyze oligonucleotide arrays (expression/SNP/tiling/exon) at probe-level. It currently supports Affymetrix (CEL files) and NimbleGen arrays (XYS files).
Authors: Benilton Carvalho and Rafael Irizarry
Maintainer: Benilton Carvalho <[email protected]>
License: LGPL (>= 2)
Version: 1.71.0
Built: 2024-10-30 09:11:02 UTC
Source: https://github.com/bioc/oligo

Help Index

The oligo package: a tool for low-level analysis of oligonucleotide arrays

Description

The oligo package provides tools to preprocess different oligonucleotide arrays types: expression, tiling, SNP and exon chips. The supported manufacturers are Affymetrix and NimbleGen.

It offers support to large datasets (when the bigmemory is loaded) and can execute preprocessing tasks in parallel (if, in addition to bigmemory, the snow package is also loaded).

Details

The package will read the raw intensity files (CEL for Affymetrix; XYS for NimbleGen) and allow the user to perform analyses starting at the feature-level.

Reading in the intensity files require the existence of data packages that contain the chip specific information (X/Y coordinates; feature types; sequence). These data packages packages are built using the pdInfoBuilder package.

For Affymetrix SNP arrays, users are asked to download the already built annotation packages from BioConductor. This is because these packages contain metadata that are not automatically created. The following annotation packages are available:

50K Xba - pd.mapping50kxba.240 50K Hind - pd.mapping50khind.240 250K Sty - pd.mapping250k.sty 250K Nsp - pd.mapping250k.nsp GenomeWideSnp 5 (SNP 5.0) - pd.genomewidesnp.5 GenomeWideSnp 6 (SNP 6.0) - pd.genomewidesnp.6

For users interested in genotype calls for SNP 5.0 and 6.0 arrays, we strongly recommend the use use the crlmm package, which implements a more efficient version of CRLMM.

Author(s)

Benilton Carvalho - [email protected]

References

Carvalho, B.; Bengtsson, H.; Speed, T. P. & Irizarry, R. A. Exploration, Normalization, and Genotype Calls of High Density Oligonucleotide SNP Array Data. Biostatistics, 2006.

Sequence Base Contents

Description

Function to compute the amounts of each nucleotide in a sequence.

Usage

basecontent(seq)

Arguments

seq

character vector of length n containg a valid sequence (A/T/C/G)

Value

matrix with n rows and 4 columns with the counts for each base.

Examples

sequences <- c("ATATATCCCCG", "TTTCCGAGC")
basecontent(sequences)

Simplified interface to PLM.

Description

Simplified interface to PLM.

Usage

basicPLM(pmMat, pnVec, normalize = TRUE, background = TRUE, transfo =
  log2, method = c('plm', 'plmr', 'plmrr', 'plmrc'), verbose = TRUE)

Arguments

pmMat

Matrix of intensities to be processed.
pnVec

Probeset names
normalize

Logical flag: normalize?
background

Logical flag: background adjustment?
transfo

function: function to be used for data transformation prior to summarization.
method

Name of the method to be used for normalization. 'plm' is the usual PLM model; 'plmr' is the (row and column) robust version of PLM; 'plmrr' is the row-robust version of PLM; 'plmrc' is the column-robust version of PLM.
verbose

Logical flag: verbose.

Value

A list with the following components:

Estimates

A (length(pnVec) x ncol(pmMat)) matrix with probeset summaries.
StdErrors

A (length(pnVec) x ncol(pmMat)) matrix with standard errors of 'Estimates'.
Residuals

A (nrow(pmMat) x ncol(pmMat)) matrix of residuals.

Note

Currently, only RMA-bg-correction and quantile normalization are allowed.

Author(s)

Benilton Carvalho

See Also

rcModelPLM, rcModelPLMr, rcModelPLMrr, rcModelPLMrc, basicRMA

Examples

set.seed(1)
pms <- 2^matrix(rnorm(1000), nc=20)
colnames(pms) <- paste("sample", 1:20, sep="")
pns <- rep(letters[1:10], each=5)
res <- basicPLM(pms, pns, TRUE, TRUE)
res[['Estimates']][1:4, 1:3]
res[['StdErrors']][1:4, 1:3]
res[['Residuals']][1:20, 1:3]

Simplified interface to RMA.

Description

Simple interface to RMA.

Usage

basicRMA(pmMat, pnVec, normalize = TRUE, background = TRUE, bgversion = 2, destructive = FALSE, verbose = TRUE, ...)

Arguments

pmMat

Matrix of intensities to be processed.

pnVec

Probeset names.

normalize

Logical flag: normalize?

background

Logical flag: background adjustment?

bgversion

Version of background correction.

destructive

Logical flag: use destructive methods?

verbose

Logical flag: verbose.

...

Not currently used.

Value

Matrix.

Examples

set.seed(1)
pms <- 2^matrix(rnorm(1000), nc=20)
colnames(pms) <- paste("sample", 1:20, sep="")
pns <- rep(letters[1:10], each=5)
res <- basicRMA(pms, pns, TRUE, TRUE)
res[, 1:3]

Boxplot

Description

Boxplot for observed (log-)intensities in a FeatureSet-like object (ExpressionFeatureSet, ExonFeatureSet, SnpFeatureSet, TilingFeatureSet) and ExpressionSet.

Usage

## S4 method for signature 'FeatureSet'
boxplot(x, which=c("pm", "mm", "bg", "both",
"all"), transfo=log2, nsample=10000, target = "mps1", ...)

## S4 method for signature 'ExpressionSet'
boxplot(x, which, transfo=identity, nsample=10000, ...)

Arguments

x

a FeatureSet-like object or ExpressionSet object.
which

character defining what probe types are to be used in the plot.
transfo

a function to transform the data before plotting. See 'Details'.
nsample

number of units to sample and build the plot.
...

arguments to be passed to the default boxplot method.

Details

The 'transfo' argument will set the transformation to be used. For raw data, 'transfo=log2' is a common practice. For summarized data (which are often in log2-scale), no transformation is needed (therefore 'transfo=identity').

Note

The boxplot methods for FeatureSet and Expression use a sample (via sample) of the probes/probesets to produce the plot. Therefore, the user interested in reproducibility is advised to use set.seed.

See Also

hist, image, sample, set.seed

Accessor for chromosome information

Description

Returns chromosome information.

Usage

pmChr(object)

Arguments

object

TilingFeatureSet or SnpCallSet object

Details

chromosome() returns the chromosomal information for all probes and pmChr() subsets the output to the PM probes only (if a TilingFeatureSet object).

Value

Vector with chromosome information.

Genotype Calls

Description

Performs genotype calls via CRLMM (Corrected Robust Linear Model with Maximum-likelihood based distances).

Usage

crlmm(filenames, outdir, batch_size=40000, balance=1.5,
      minLLRforCalls=c(5, 1, 5), recalibrate=TRUE,
      verbose=TRUE, pkgname, reference=TRUE)
justCRLMM(filenames, batch_size = 40000, minLLRforCalls = c(5, 1, 5),
recalibrate = TRUE, balance = 1.5, phenoData = NULL, verbose = TRUE,
pkgname = NULL, tmpdir=tempdir())

Arguments

filenames

character vector with the filenames.
outdir

directory where the output (and some tmp files) files will be saved.
batch_size

integer defining how many SNPs should be processed at a time.
recalibrate

Logical - should recalibration be performed?
balance

Control parameter to balance homozygotes and heterozygotes calls.
minLLRforCalls

Minimum thresholds for genotype calls.
verbose

Logical.
phenoData

phenoData object or NULL
pkgname

alt. pdInfo package to be used
reference

logical, defaulting to TRUE ...
tmpdir

Directory where temporary files are going to be stored at.

Value

SnpCallSetPlus object.

Create set of colors, interpolating through a set of preferred colors.

Description

Create set of colors, interpolating through a set of preferred colors.

Usage

darkColors(n)
seqColors(n)
seqColors2(n)
divColors(n)

Arguments

n

integer determining number of colors to be generated

Details

darkColors is based on the Dark2 palette in RColorBrewer, therefore useful to describe qualitative features of the data.

seqColors is based on Blues and generates a gradient of blues, therefore useful to describe quantitative features of the data. seqColors2 behaves similarly, but it is based on OrRd (white-orange-red).

divColors is based on the RdBu pallete in RColorBrewer, therefore useful to describe quantitative features ranging on two extremes.

Examples

x <- 1:10
y <- 1:10
cols1 <- darkColors(10)
cols2 <- seqColors(10)
cols3 <- divColors(10)
cols4 <- seqColors2(10)
plot(x, y, col=cols1, xlim=c(1, 13), pch=19, cex=3)
points(x+1, y, col=cols2, pch=19, cex=3)
points(x+2, y, col=cols3, pch=19, cex=3)
points(x+3, y, col=cols4, pch=19, cex=3)
abline(0, 1, lty=2)
abline(-1, 1, lty=2)
abline(-2, 1, lty=2)
abline(-3, 1, lty=2)

Tool to fit Probe Level Models.

Description

Fits robust Probe Level linear Models to all the (meta)probesets in an FeatureSet. This is carried out on a (meta)probeset by (meta)probeset basis.

Usage

fitProbeLevelModel(object, background=TRUE, normalize=TRUE, target="core", method="plm", verbose=TRUE, S4=TRUE, ...)

Arguments

object

FeatureSet object.
background

Do background correction?
normalize

Do normalization?
target

character vector describing the summarization target. Valid values are: 'probeset', 'core' (Gene/Exon), 'full' (Exon), 'extended' (Exon).
method

summarization method to be used.
verbose

verbosity flag.
S4

return final value as an S4 object (oligoPLM) if TRUE. If FALSE, final value is returned as a list.
...

subset to be passed down to getProbeInfo for subsetting. See subset for details.

Value

fitProbeLevelModel returns an oligoPLM object, if S4=TRUE; otherwise, it will return a list.

Note

This is the initial port of fitPLM to oligo. Some features found on the original work by Ben Bolstad (in the affyPLM package) may not be yet available. If you found one of this missing characteristics, please contact Benilton Carvalho.

Author(s)

This is a simplified port from Ben Bolstad's work implemented in the affyPLM package. Problems with the implementation in oligo should be reported to Benilton Carvalho.

References

Bolstad, BM (2004) Low Level Analysis of High-density Oligonucleotide Array Data: Background, Normalization and Summarization. PhD Dissertation. University of California, Berkeley.

See Also

rma, summarizationMethods, subset

Examples

if (require(oligoData)){
  data(nimbleExpressionFS)
  fit <- fitProbeLevelModel(nimbleExpressionFS)
  image(fit)
  NUSE(fit)
  RLE(fit)
}

Estimate affinity coefficients.

Description

Estimate affinity coefficients using sequence information and splines.

Usage

getAffinitySplineCoefficients(intensities, sequences)

Arguments

intensities

Intensity matrix
sequences

Probe sequences

Value

Matrix with estimated coefficients.

See Also

getBaseProfile

Compute and plot nucleotide profile.

Description

Computes and, optionally, lots nucleotide profile, describing the sequence effect on intensities.

Usage

getBaseProfile(coefs, probeLength = 25, plot = FALSE, ...)

Arguments

coefs

affinity spline coefficients.
probeLength

length of probes
plot

logical. Plots profile?
...

arguments to be passed to matplot.

Value

Invisibly returns a matrix with estimated effects.

Get container information for NimbleGen Tiling Arrays.

Description

Get container information for NimbleGen Tiling Arrays. This is useful for better identification of control probes.

Usage

getContainer(object, probeType)

Arguments

object

A TilingFeatureSet or TilingFeatureSet object.

probeType

String describing which probes to query ('pm', 'bg')

Value

'character' vector with container information.

Function to get CRLMM summaries saved to disk

Description

This will read the summaries written to disk and return them to the user as a SnpCallSetPlus or SnpCnvCallSetPlus object.

Usage

getCrlmmSummaries(tmpdir)

Arguments

tmpdir

directory where CRLMM saved the results to.

Value

If the data were from SNP 5.0 or 6.0 arrays, the function will return a SnpCnvCallSetPlus object. It will return a SnpCallSetPlus object, otherwise.

NetAffx Biological Annotations

Description

Gets NetAffx Biological Annotations saved in the annotation package (Exon and Gene ST Affymetrix arrays).

Usage

getNetAffx(object, type = "probeset")

Arguments

object

'ExpressionSet' object (eg., result of rma())
type

Either 'probeset' or 'transcript', depending on what type of summaries were obtained.

Details

This retrieves NetAffx annotation saved in the (pd) annotation package - annotation(object). It is only available for Exon ST and Gene ST arrays.

The 'type' argument should match the summarization target used to generate 'object'. The 'rma' method allows for two targets: 'probeset' (target='probeset') and 'transcript' (target='core', target='full', target='extended').

Value

'AnnotatedDataFrame' that can be used as featureData(object)

Author(s)

Benilton Carvalho

Helper function to extract color information for filenames on NimbleGen arrays.

Description

This function will (try to) extract the color information for NimbleGen arrays. This is useful when using read.xysfiles2 to parse XYS files for Tiling applications.

Usage

getNgsColorsInfo(path = ".", pattern1 = "_532", pattern2 = "_635", ...)

Arguments

path

path where to look for files
pattern1

pattern to match files supposed to go to the first channel
pattern2

pattern to match files supposed to go to the second channel
...

extra arguments for list.xysfiles

Details

Many NimbleGen samples are identified following the pattern sampleID_532.XYS / sampleID_635.XYS.

The function suggests sample names if all the filenames follow the standard above.

Value

A data.frame with, at least, two columns: 'channel1' and 'channel2'. A third column, 'sampleNames', is returned if the filenames follow the sampleID_532.XYS / sampleID_635.XYS standard.

Author(s)

Benilton Carvalho <[email protected]>

Retrieve Platform Design object

Description

Retrieve platform design object.

Usage

getPlatformDesign(object)
getPD(object)

Arguments

object

FeatureSet object

Details

Retrieve platform design object.

Value

platformDesign or PDInfo object.

Probe information selector.

Description

A tool to simplify the selection of probe information, so user does not need to use the SQL approaches.

Usage

getProbeInfo(object, field, probeType = "pm", target = "core", sortBy = c("fid", "man_fsetid", "none"), ...)

Arguments

object

FeatureSet object.
field

character string with names of field(s) of interest to be obtained from database.
probeType

character string: 'pm' or 'mm'
target

Used only for Exon or Gene ST arrays: 'core', 'full', 'extended', 'probeset'.
sortBy

Field to be used for sorting.
...

Arguments to be passed to subset

Value

A data.frame with the probe level information.

Note

The code allows for querying info on MM probes, however it has been used mostly on PM probes.

Author(s)

Benilton Carvalho

Examples

if (require(oligoData)){
   data(affyGeneFS)
   availProbeInfo(affyGeneFS)
   probeInfo <- getProbeInfo(affyGeneFS, c('fid', 'x', 'y', 'chrom'))
   head(probeInfo)
   ## Selecting antigenomic background probes
   agenGene <- getProbeInfo(affyGeneFS, field=c('fid', 'fsetid', 'type'), target='probeset', subset= type == 'control->bgp->antigenomic')
   head(agenGene)
}

Accessors for physical array coordinates.

Description

Accessors for physical array coordinates.

Usage

getX(object, type)
getY(object, type)

Arguments

object

FeatureSet object
type

'character' defining the type of the probes to be queried. Valid options are 'pm', 'mm', 'bg'

Value

A vector with the requested coordinates.

Examples

## Not run: 
x <- read.celfiles(list.celfiles())
theXpm <- getX(x, "pm")
theYpm <- getY(x, "pm")

## End(Not run)

Density estimate

Description

Plot the density estimates for each sample

Usage

## S4 method for signature 'FeatureSet'
hist(x, transfo=log2, which=c("pm", "mm", "bg", "both", "all"),
                   nsample=10000, target = "mps1", ...)

## S4 method for signature 'ExpressionSet'
hist(x, transfo=identity, nsample=10000, ...)

Arguments

x

FeatureSet or ExpressionSet object
transfo

a function to transform the data before plotting. See 'Details'.
nsample

number of units to sample and build the plot.
which

set of probes to be plotted ("pm", "mm", "bg", "both", "all").
...

arguments to be passed to matplot

Details

The 'transfo' argument will set the transformation to be used. For raw data, 'transfo=log2' is a common practice. For summarized data (which are often in log2-scale), no transformation is needed (therefore 'transfo=identity').

Note

The hist methods for FeatureSet and Expression use a sample (via sample) of the probes/probesets to produce the plot (unless nsample > nrow(x)). Therefore, the user interested in reproducibility is advised to use set.seed.

Display a pseudo-image of a microarray chip

Description

Produces a pseudo-image (graphics::image) for each sample.

Usage

## S4 method for signature 'FeatureSet'
image(x, which, transfo=log2, ...)

## S4 method for signature 'PLMset'
image(x, which=0,
                   type=c("weights","resids", "pos.resids","neg.resids","sign.resids"),
                   use.log=TRUE, add.legend=FALSE, standardize=FALSE,
                   col=NULL, main, ...)

Arguments

x

FeatureSet object
which

integer indices of samples to be plotted (optional).
transfo

function to be applied to the data prior to plotting.
type

Type of statistics to be used.
use.log

Use log.
add.legend

Add legend.
standardize

Standardize residuals.
col

Colors to be used.
main

Main title.
...

parameters to be passed to image

Examples

if(require(oligoData) & require(pd.hg18.60mer.expr)){
  data(nimbleExpressionFS)
  par(mfrow=c(1, 2))
  image(nimbleExpressionFS, which=4)
##  fit <- fitPLM(nimbleExpressionFS)
##  image(fit, which=4)
  plot(1) ## while fixing fitPLM TODO
}

Summarization of SNP data

Description

This function implements the SNPRMA method for summarization of SNP data. It works directly with the CEL files, saving memory.

Usage

justSNPRMA(filenames, verbose = TRUE, phenoData = NULL, normalizeToHapmap = TRUE)

Arguments

filenames

character vector with the filenames.
verbose

logical flag for verbosity.
phenoData

a phenoData object or NULL
normalizeToHapmap

Normalize to Hapmap? Should always be TRUE, but it's kept here for future use.

Value

SnpQSet or a SnpCnvQSet, depending on the array type.

Examples

## snprmaResults <- justSNPRMA(list.celfiles())

List XYS files

Description

Lists the XYS files.

Usage

list.xysfiles(...)

Arguments

...

parameters to be passed to list.files

Details

The functions interface list.files and the user is asked to check that function for further details.

Value

Character vector with the filenames.

See Also

list.files

Examples

list.xysfiles()

MA plots

Description

Create MA plots using a reference array (if one channel) or using channel2 as reference (if two channel).

Usage

MAplot(object, ...)

## S4 method for signature 'FeatureSet'
MAplot(object, what=pm, transfo=log2, groups,
       refSamples, which, pch=".", summaryFun=rowMedians,
       plotFun=smoothScatter, main="vs pseudo-median reference chip",
       pairs=FALSE, ...)

## S4 method for signature 'TilingFeatureSet'
MAplot(object, what=pm, transfo=log2, groups,
       refSamples, which, pch=".", summaryFun=rowMedians,
       plotFun=smoothScatter, main="vs pseudo-median reference chip",
       pairs=FALSE, ...)

## S4 method for signature 'PLMset'
MAplot(object, what=coefs, transfo=identity, groups,
       refSamples, which, pch=".", summaryFun=rowMedians,
       plotFun=smoothScatter, main="vs pseudo-median reference chip",
       pairs=FALSE, ...)

## S4 method for signature 'matrix'
MAplot(object, what=identity, transfo=identity,
       groups, refSamples, which, pch=".", summaryFun=rowMedians,
       plotFun=smoothScatter, main="vs pseudo-median reference chip",
       pairs=FALSE, ...)

## S4 method for signature 'ExpressionSet'
MAplot(object, what=exprs, transfo=identity,
       groups, refSamples, which, pch=".", summaryFun=rowMedians,
       plotFun=smoothScatter, main="vs pseudo-median reference chip",
       pairs=FALSE, ...)

Arguments

object

FeatureSet, PLMset or ExpressionSet object.
what

function to be applied on object that will extract the statistics of interest, from which log-ratios and average log-intensities will be computed.
transfo

function to transform the data prior to plotting.
groups

factor describing groups of samples that will be combined prior to plotting. If missing, MvA plots are done per sample.
refSamples

integers (indexing samples) to define which subjects will be used to compute the reference set. If missing, a pseudo-reference chip is estimated using summaryFun.
which

integer (indexing samples) describing which samples are to be plotted.
pch

same as pch in plot
summaryFun

function that operates on a matrix and returns a vector that will be used to summarize data belonging to the same group (or reference) on the computation of grouped-stats.
plotFun

function to be used for plotting. Usually smoothScatter, plot or points.
main

string to be used in title.
pairs

logical flag to determine if a matrix of MvA plots is to be generated
...

Other arguments to be passed downstream, like plot arguments.

Details

MAplot will take the following extra arguments:

  1. subset: indices of elements to be plotted to reduce impact of plotting 100's thousands points (if pairs=FALSE only);

  2. span: see loess;

  3. family.loess: see loess;

  4. addLoess: logical flag (default TRUE) to add a loess estimate;

  5. parParams: list of params to be passed to par() (if pairs=TRUE only);

Value

Plot

Author(s)

Benilton Carvalho - based on Ben Bolstad's original MAplot function.

See Also

plot, smoothScatter

Examples

if(require(oligoData) & require(pd.hg18.60mer.expr)){
  data(nimbleExpressionFS)
  nimbleExpressionFS
  groups <- factor(rep(c('brain', 'UnivRef'), each=3))
  data.frame(sampleNames(nimbleExpressionFS), groups)
  MAplot(nimbleExpressionFS, pairs=TRUE, ylim=c(-.5, .5), groups=groups)
}

Accessors and replacement methods for the intensity/PM/MM/BG matrices.

Description

Accessors and replacement methods for the PM/MM/BG matrices.

Usage

intensity(object)
mm(object, subset = NULL, target='core')
pm(object, subset = NULL, target='core')
bg(object, subset = NULL)
mm(object, subset = NULL, target='core')<-value
pm(object, subset = NULL, target='core')<-value
bg(object)<-value

Arguments

object

FeatureSet object.
subset

Not implemented yet.
value

matrix object.
target

One of 'probeset', 'core', 'full', 'extended'. This is ignored if the array design is something other than Gene ST or Exon ST.

Details

For all objects but TilingFeatureSet, these methods will return matrices. In case of TilingFeatureSet objects, the value is a 3-dimensional array (probes x samples x channels).

intensity will return the whole intensity matrix associated to the object. pm, mm, bg will return the respective PM/MM/BG matrix.

When applied to ExonFeatureSet or GeneFeatureSet objects, pm will return the PM matrix at the transcript level ('core' probes) by default. The user should set the target argument accordingly if something else is desired. The valid values are: 'probeset' (Exon and Gene arrays), 'core' (Exon and Gene arrays), 'full' (Exon arrays) and 'extended' (Exon arrays).

The target argument has no effects when used on designs other than Gene and Exon ST.

Examples

if (require(maqcExpression4plex) & require(pd.hg18.60mer.expr)){
xysPath <- system.file("extdata", package="maqcExpression4plex")
xysFiles <- list.xysfiles(xysPath, full.name=TRUE)
ngsExpressionFeatureSet <- read.xysfiles(xysFiles)
pm(ngsExpressionFeatureSet)[1:10,]
}

Accessors for PM, MM or background probes indices.

Description

Extracts the indexes for PM, MM or background probes.

Usage

mmindex(object, ...)
pmindex(object, ...)
bgindex(object, ...)

Arguments

object

FeatureSet or DBPDInfo object
...

Extra arguments, not yet implemented

Details

The indices are ordered by 'fid', i.e. they follow the order that the probes appear in the CEL/XYS files.

Value

A vector of integers representing the rows of the intensity matrix that correspond to PM, MM or background probes.

Examples

## How pm() works
## Not run: 
x <- read.celfiles(list.celfiles())
pms0 <- pm(x)
pmi <- pmindex(x)
pms1 <- exprs(x)[pmi,]
identical(pms0, pms1)

## End(Not run)

Probe Sequeces

Description

Accessor to the (PM/MM/background) probe sequences.

Usage

mmSequence(object)
pmSequence(object, ...)
bgSequence(object, ...)

Arguments

object

FeatureSet, AffySNPPDInfo or DBPDInfo object
...

additional arguments

Value

A DNAStringSet containing the PM/MM/background probe sequence associated to the array.

Defunct Functions in Package 'oligo'

Description

The functions or variables listed here are no longer part of 'oligo'

Usage

fitPLM(...)
coefs(...)
resids(...)

Arguments

...

Arguments.

Details

fitPLM was replaced by fitProbeLevelModel, allowing faster execution and providing more specific models. fitPLM was based in the code written by Ben Bolstad in the affyPLM package. However, all the model-fitting functions are now in the package preprocessCore, on which fitProbeLevelModel depends.

coefs and resids, like fitPLM, were inherited from the affyPLM package. They were replaced respectively by coef and residuals, because this is how these statistics are called everywhere else in R.

Class "oligoPLM"

Description

A class to represent Probe Level Models.

Objects from the Class

Objects can be created by calls of the form fitProbeLevelModel(FeatureSetObject), where FeatureSetObject is an object obtained through read.celfiles or read.xysfiles, representing intensities observed for different probes (which are grouped in probesets or meta-probesets) across distinct samples.

Slots

chip.coefs:

"matrix" with chip/sample effects - probeset-level

description:

"MIAME" compliant description information.

phenoData:

"AnnotatedDataFrame" with phenotypic data.

protocolData:

"AnnotatedDataFrame" with protocol data.

probe.coefs:

"numeric" vector with probe effects

weights:

"matrix" with weights - probe-level

residuals:

"matrix" with residuals - probe-level

se.chip.coefs:

"matrix" with standard errors for chip/sample coefficients

se.probe.coefs:

"numeric" vector with standard errors for probe effects

residualSE:

scale - residual standard error

geometry:

array geometry used for plots

method:

"character" string describing method used for PLM

manufacturer:

"character" string with manufacturer name

annotation:

"character" string with the name of the annotation package

narrays:

"integer" describing the number of arrays

nprobes:

"integer" describing the number of probes before summarization

nprobesets:

"integer" describing the number of probesets after summarization

Methods

annotation

signature(object = "oligoPLM"): accessor/replacement method to annotation slot

boxplot

signature(x = "oligoPLM"): boxplot method

coef

signature(object = "oligoPLM"): accessor/replacement method to coef slot

coefs.probe

signature(object = "oligoPLM"): accessor/replacement method to coefs.probe slot

geometry

signature(object = "oligoPLM"): accessor/replacement method to geometry slot

image

signature(x = "oligoPLM"): image method

manufacturer

signature(object = "oligoPLM"): accessor/replacement method to manufacturer slot

method

signature(object = "oligoPLM"): accessor/replacement method to method slot

ncol

signature(x = "oligoPLM"): accessor/replacement method to ncol slot

nprobes

signature(object = "oligoPLM"): accessor/replacement method to nprobes slot

nprobesets

signature(object = "oligoPLM"): accessor/replacement method to nprobesets slot

residuals

signature(object = "oligoPLM"): accessor/replacement method to residuals slot

residualSE

signature(object = "oligoPLM"): accessor/replacement method to residualSE slot

se

signature(object = "oligoPLM"): accessor/replacement method to se slot

se.probe

signature(object = "oligoPLM"): accessor/replacement method to se.probe slot

show

signature(object = "oligoPLM"): show method

weights

signature(object = "oligoPLM"): accessor/replacement method to weights slot

NUSE

signature(x = "oligoPLM") : Boxplot of Normalized Unscaled Standard Errors (NUSE) or NUSE values.

RLE

signature(x = "oligoPLM") : Relative Log Expression boxplot or values.

opset2eset

signature(x = "oligoPLM") : Convert to ExpressionSet.

Author(s)

This is a port from Ben Bolstad's work implemented in the affyPLM package. Problems with the implementation in oligo should be reported to the package's maintainer.

References

Bolstad, BM (2004) Low Level Analysis of High-density Oligonucleotide Array Data: Background, Normalization and Summarization. PhD Dissertation. University of California, Berkeley.

See Also

rma, summarize

Examples

## TODO: review code and fix broken
## Not run: 
if (require(oligoData)){
  data(nimbleExpressionFS)
  fit <- fitProbeLevelModel(nimbleExpressionFS)
  image(fit)
  NUSE(fit)
  RLE(fit)
}

## End(Not run)

Methods for P/A Calls

Description

Methods for Present/Absent Calls are meant to provide means of assessing whether or not each of the (PM) intensities are compatible with observations generated by background probes.

Usage

paCalls(object, method, ..., verbose=TRUE)
## S4 method for signature 'ExonFeatureSet'
paCalls(object, method, verbose = TRUE)
## S4 method for signature 'GeneFeatureSet'
paCalls(object, method, verbose = TRUE)
## S4 method for signature 'ExpressionFeatureSet'
paCalls(object, method, ..., verbose = TRUE)

Arguments

object

Exon/Gene/Expression-FeatureSet object.
method

String defining what method to use. See 'Details'.
...

Additional arguments passed to MAS5. See 'Details'
verbose

Logical flag for verbosity.

Details

For Whole Transcript arrays (Exon/Gene) the valid options for method are 'DABG' (p-values for each probe) and 'PSDABG' (p-values for each probeset). For Expression arrays, the only option currently available for method is 'MAS5'.

ABOUT MAS5 CALLS:

The additional arguments that can be passed to MAS5 are:

  1. alpha1: a significance threshold in (0, alpha2);

  2. alpha2: a significance threshold in (alpha1, 0.5);

  3. tau: a small positive constant;

  4. ignore.saturated: if TRUE, do the saturation correction described in the paper, with a saturation level of 46000;

This function performs the hypothesis test:

H0: median(Ri) = tau, corresponding to absence of transcript H1: median(Ri) > tau, corresponding to presence of transcript

where Ri = (PMi - MMi) / (PMi + MMi) for each i a probe-pair in the probe-set represented by data.

The p-value that is returned estimates the usual quantity:

Pr(observing a more "present looking" probe-set than data | data is absent)

So that small p-values imply presence while large ones imply absence of transcript. The detection call is computed by thresholding the p-value as in:

call "P" if p-value < alpha1 call "M" if alpha1 <= p-value < alpha2 call "A" if alpha2 <= p-value

Value

A matrix (of dimension dim(PM) if method="DABG" or "MAS5"; of dimension length(unique(probeNames(object))) x ncol(object) if method="PSDABG") with p-values for P/A Calls.

Author(s)

Benilton Carvalho

References

Clark et al. Discovery of tissue-specific exons using comprehensive human exon microarrays. Genome Biol (2007) vol. 8 (4) pp. R64

Liu, W. M. and Mei, R. and Di, X. and Ryder, T. B. and Hubbell, E. and Dee, S. and Webster, T. A. and Harrington, C. A. and Ho, M. H. and Baid, J. and Smeekens, S. P. (2002) Analysis of high density expression microarrays with signed-rank call algorithms, Bioinformatics, 18(12), pp. 1593–1599.

Liu, W. and Mei, R. and Bartell, D. M. and Di, X. and Webster, T. A. and Ryder, T. (2001) Rank-based algorithms for analysis of microarrays, Proceedings of SPIE, Microarrays: Optical Technologies and Informatics, 4266.

Affymetrix (2002) Statistical Algorithms Description Document, Affymetrix Inc., Santa Clara, CA, whitepaper. http://www.affymetrix.com/support/technical/whitepapers/sadd_whitepaper.pdf

Examples

## Not run: 
if (require(oligoData) & require(pd.huex.1.0.st.v2)){
  data(affyExonFS)
  ## Get only 2 samples for example
  dabgP = paCalls(affyExonFS[, 1:2])
  dabgPS = paCalls(affyExonFS[, 1:2], "PSDABG")
  head(dabgP) ## for probe
  head(dabgPS) ## for probeset
}

## End(Not run)

Methods for Log-Ratio plotting

Description

The plotM methods are meant to plot log-ratios for different classes of data.

Methods

object = "SnpQSet", i = "character"

Plot log-ratio for SNP data for sample i.

object = "SnpQSet", i = "integer"

Plot log-ratio for SNP data for sample i.

object = "SnpQSet", i = "numeric"

Plot log-ratio for SNP data for sample i.

object = "TilingQSet", i = "missing"

Plot log-ratio for Tiling data for sample i.

Access the allele information for PM probes.

Description

Accessor to the allelic information for PM probes.

Usage

pmAllele(object)

Arguments

object

SnpFeatureSet or PDInfo object.

Access the fragment length for PM probes.

Description

Accessor to the fragment length for PM probes.

Usage

pmFragmentLength(object, enzyme, type=c('snp', 'cn'))

Arguments

object

PDInfo or SnpFeatureSet object.
enzyme

Enzyme to be used for query. If missing, all enzymes are used.
type

Type of probes to be used: 'snp' for SNP probes; 'cn' for Copy Number probes.

Value

A list of length equal to the number of enzymes used for digestion. Each element of the list is a data.frame containing:

  • row: the row used to link to the PM matrix;

  • length: expected fragment length.

Note

There is not a 1:1 relationship between probes and expected fragment length. For one enzyme, a given probe may be associated to multiple fragment lengths. Therefore, the number of rows in the data.frame may not match the number of PM probes and the row column should be used to match the fragment length with the PM matrix.

Accessor to position information

Description

pmPosition will return the genomic position for the (PM) probes.

Usage

pmPosition(object)
pmOffset(object)

Arguments

object

AffySNPPDInfo, TilingFeatureSet or SnpCallSet object

Details

pmPosition will return genomic position for PM probes on a tiling array.

pmOffset will return the offset information for PM probes on SNP arrays.

Accessor to the strand information

Description

Returns the strand information for PM probes (0 - sense / 1 - antisense).

Usage

pmStrand(object)

Arguments

object

AffySNPPDInfo or TilingFeatureSet object

Accessor to feature names

Description

Accessors to featureset names.

Usage

probeNames(object, subset = NULL, ...)
probesetNames(object, ...)

Arguments

object

FeatureSet or DBPDInfo
subset

not implemented yet.
...

Arguments (like 'target') passed to downstream methods.

Value

probeNames returns a string with the probeset names for *each probe* on the array. probesetNames, on the other hand, returns the *unique probeset names*.

Parser to CEL files

Description

Reads CEL files.

Usage

read.celfiles(..., filenames, pkgname, phenoData, featureData,
experimentData, protocolData, notes, verbose=TRUE, sampleNames,
rm.mask=FALSE, rm.outliers=FALSE, rm.extra=FALSE, checkType=TRUE)

read.celfiles2(channel1, channel2, pkgname, phenoData, featureData,
experimentData, protocolData, notes, verbose=TRUE, sampleNames,
rm.mask=FALSE, rm.outliers=FALSE, rm.extra=FALSE, checkType=TRUE)

Arguments

...

names of files to be read.
filenames

a character vector with the CEL filenames.
channel1

a character vector with the CEL filenames for the first 'channel' on a Tiling application
channel2

a character vector with the CEL filenames for the second 'channel' on a Tiling application
pkgname

alternative data package to be loaded.
phenoData

phenoData
featureData

featureData
experimentData

experimentData
protocolData

protocolData
notes

notes
verbose

logical
sampleNames

character vector with sample names (usually better descriptors than the filenames)
rm.mask

logical. Read masked?
rm.outliers

logical. Remove outliers?
rm.extra

logical. Remove extra?
checkType

logical. Check type of each file? This can be time consuming.

Details

When using 'affyio' to read in CEL files, the user can read compressed CEL files (CEL.gz). Additionally, 'affyio' is much faster than 'affxparser'.

The function guesses which annotation package to use from the header of the CEL file. The user can also provide the name of the annotaion package to be used (via the pkgname argument). If the annotation package cannot be loaded, the function returns an error. If the annotation package is not available from BioConductor, one can use the pdInfoBuilder package to build one.

Value

ExpressionFeatureSet

if Expresssion arrays
ExonFeatureSet

if Exon arrays
SnpFeatureSet

if SNP arrays
TilingFeatureSet

if Tiling arrays

See Also

list.celfiles, read.xysfiles

Examples

if(require(pd.mapping50k.xba240) & require(hapmap100kxba)){
celPath <- system.file("celFiles", package="hapmap100kxba")
celFiles <- list.celfiles(celPath, full.name=TRUE)
affySnpFeatureSet <- read.celfiles(celFiles)
}

Parser to XYS files

Description

NimbleGen provides XYS files which are read by this function.

Usage

read.xysfiles(..., filenames, pkgname, phenoData, featureData,
experimentData, protocolData, notes, verbose=TRUE, sampleNames,
checkType=TRUE)

read.xysfiles2(channel1, channel2, pkgname, phenoData, featureData,
experimentData, protocolData, notes, verbose=TRUE, sampleNames,
checkType=TRUE)

Arguments

...

file names
filenames

character vector with filenames.
channel1

a character vector with the XYS filenames for the first 'channel' on a Tiling application
channel2

a character vector with the XYS filenames for the second 'channel' on a Tiling application
pkgname

character vector with alternative PD Info package name
phenoData

phenoData
featureData

featureData
experimentData

experimentData
protocolData

protocolData
notes

notes
verbose

verbose
sampleNames

character vector with sample names (usually better descriptors than the filenames)
checkType

logical. Check type of each file? This can be time consuming.

Details

The function will read the XYS files provided by NimbleGen Systems and return an object of class FeatureSet.

The function guesses which annotation package to use from the header of the XYS file. The user can also provide the name of the annotaion package to be used (via the pkgname argument). If the annotation package cannot be loaded, the function returns an error. If the annotation package is not available from BioConductor, one can use the pdInfoBuilder package to build one.

Value

ExpressionFeatureSet

if Expresssion arrays
TilingFeatureSet

if Tiling arrays

See Also

list.xysfiles, read.celfiles

Examples

if (require(maqcExpression4plex) & require(pd.hg18.60mer.expr)){
xysPath <- system.file("extdata", package="maqcExpression4plex")
xysFiles <- list.xysfiles(xysPath, full.name=TRUE)
ngsExpressionFeatureSet <- read.xysfiles(xysFiles)
}

Read summaries generated by crlmm

Description

This function read the different summaries generated by crlmm.

Usage

readSummaries(type, tmpdir)

Arguments

type

type of summary of character class: 'alleleA', 'alleleB', 'alleleA-sense', 'alleleA-antisense', 'alleleB-sense', 'alleleB-antisense', 'calls', 'llr', 'conf'.
tmpdir

directory containing the output saved by crlmm

Details

On the 50K and 250K arrays, given a SNP, there are probes on both strands (sense and antisense). For this reason, the options 'alleleA-sense', 'alleleA-antisense', 'alleleB-sense' and 'alleleB-antisense' should be used **only** with such arrays (XBA, HIND, NSP or STY).

On the SNP 5.0 and SNP 6.0 platforms, this distinction does not exist in terms of algorithm (note that the actual strand could be queried from the annotation package). For these arrays, options 'alleleA', 'alleleB' are the ones to be used.

The options calls, llr and conf will return, respectivelly, the CRLMM calls, log-likelihood ratios (for devel purpose **only**) and CRLMM confidence calls matrices.

Value

Matrix with values of summaries.

RMA - Robust Multichip Average algorithm

Description

Robust Multichip Average preprocessing methodology. This strategy allows background subtraction, quantile normalization and summarization (via median-polish).

Usage

## S4 method for signature 'ExonFeatureSet'
rma(object, background=TRUE, normalize=TRUE, subset=NULL, target="core")
  ## S4 method for signature 'HTAFeatureSet'
rma(object, background=TRUE, normalize=TRUE, subset=NULL, target="core")
  ## S4 method for signature 'ExpressionFeatureSet'
rma(object, background=TRUE, normalize=TRUE, subset=NULL)
  ## S4 method for signature 'GeneFeatureSet'
rma(object, background=TRUE, normalize=TRUE, subset=NULL, target="core")
  ## S4 method for signature 'SnpCnvFeatureSet'
rma(object, background=TRUE, normalize=TRUE, subset=NULL)

Arguments

object

Exon/HTA/Expression/Gene/SnpCnv-FeatureSet object.
background

Logical - perform RMA background correction?
normalize

Logical - perform quantile normalization?
subset

To be implemented.
target

Level of summarization (only for Exon/Gene arrays)

Methods

signature(object = "ExonFeatureSet")

When applied to an ExonFeatureSet object, rma can produce summaries at different levels: probeset (as defined in the PGF), core genes (as defined in the core.mps file), full genes (as defined in the full.mps file) or extended genes (as defined in the extended.mps file). To determine the level for summarization, use the target argument.

signature(object = "ExpressionFeatureSet")

When used on an ExpressionFeatureSet object, rma produces summaries at the probeset level (as defined in the CDF or NDF files, depending on the manufacturer).

signature(object = "GeneFeatureSet")

When applied to a GeneFeatureSet object, rma can produce summaries at different levels: probeset (as defined in the PGF) and 'core genes' (as defined in the core.mps file). To determine the level for summarization, use the target argument.

signature(object = "HTAFeatureSet")

When applied to a HTAFeatureSet object, rma can produce summaries at different levels: probeset (as defined in the PGF) and 'core genes' (as defined in the core.mps file). To determine the level for summarization, use the target argument.

signature(object = "SnpCnvFeatureSet")

If used on a SnpCnvFeatureSet object (ie., SNP 5.0 or SNP 6.0 arrays), rma will produce summaries for the CNV probes. Note that this is an experimental feature for internal (and quick) assessment of CNV probes. We recommend the use of the 'crlmm' package, which contains a Copy Number tool specifically designed for these data.

References

Rafael. A. Irizarry, Benjamin M. Bolstad, Francois Collin, Leslie M. Cope, Bridget Hobbs and Terence P. Speed (2003), Summaries of Affymetrix GeneChip probe level data Nucleic Acids Research 31(4):e15

Bolstad, B.M., Irizarry R. A., Astrand M., and Speed, T.P. (2003), A Comparison of Normalization Methods for High Density O ligonucleotide Array Data Based on Bias and Variance. Bioinformatics 19(2):185-193

Irizarry, RA, Hobbs, B, Collin, F, Beazer-Barclay, YD, Antonellis, KJ, Scherf, U, Speed, TP (2003) Exploration, Normalizati on, and Summaries of High Density Oligonucleotide Array Probe Level Data. Biostatistics. Vol. 4, Number 2: 249-264

See Also

snprma

Examples

if (require(maqcExpression4plex) & require(pd.hg18.60mer.expr)){
xysPath <- system.file("extdata", package="maqcExpression4plex")
xysFiles <- list.xysfiles(xysPath, full.name=TRUE)
ngsExpressionFeatureSet <- read.xysfiles(xysFiles)
summarized <- rma(ngsExpressionFeatureSet)
show(summarized)
}

Date of scan

Description

Retrieves date information in CEL/XYS files.

Usage

runDate(object)

Arguments

object

'FeatureSet' object.

Create design matrix for sequences

Description

Creates design matrix for sequences.

Usage

sequenceDesignMatrix(seqs)

Arguments

seqs

character vector of 25-mers.

Details

This assumes all sequences are 25bp long.

The design matrix is often used when the objecive is to adjust intensities by sequence.

Value

Matrix with length(seqs) rows and 75 columns.

Examples

genSequence <- function(x)
    paste(sample(c("A", "T", "C", "G"), 25, rep=TRUE), collapse="", sep="")
seqs <- sapply(1:10, genSequence)
X <- sequenceDesignMatrix(seqs)
Y <- rnorm(10, mean=12, sd=2)
Ydemean <- Y-mean(Y)
X[1:10, 1:3]
fit <- lm(Ydemean~X)
coef(fit)

Preprocessing SNP Arrays

Description

This function preprocess SNP arrays.

Usage

snprma(object, verbose = TRUE, normalizeToHapmap = TRUE)

Arguments

object

SnpFeatureSet object
verbose

Verbosity flag. logical
normalizeToHapmap

internal

Value

A SnpQSet object.

Tools for microarray preprocessing.

Description

These are tools to preprocess microarray data. They include background correction, normalization and summarization methods.

Usage

backgroundCorrectionMethods()
normalizationMethods()
summarizationMethods()
backgroundCorrect(object, method=backgroundCorrectionMethods(), copy=TRUE, extra, subset=NULL, target='core', verbose=TRUE)
summarize(object, probes=rownames(object), method="medianpolish", verbose=TRUE, ...)
## S4 method for signature 'FeatureSet'
normalize(object, method=normalizationMethods(), copy=TRUE, subset=NULL,target='core', verbose=TRUE, ...)
## S4 method for signature 'matrix'
normalize(object, method=normalizationMethods(), copy=TRUE, verbose=TRUE, ...)
## S4 method for signature 'ff_matrix'
normalize(object, method=normalizationMethods(), copy=TRUE, verbose=TRUE, ...)
normalizeToTarget(object, targetDist, method="quantile", copy=TRUE, verbose=TRUE)

Arguments

object

Object containing probe intensities to be preprocessed.
method

String determining which method to use at that preprocessing step.
targetDist

Vector with the target distribution
probes

Character vector that identifies the name of the probes represented by the rows of object.
copy

Logical flag determining if data must be copied before processing (TRUE), or if data can be overwritten (FALSE).
subset

Not yet implemented.
target

One of the following values: 'core', 'full', 'extended', 'probeset'. Used only with Gene ST and Exon ST designs.
extra

Extra arguments to be passed to other methods.
verbose

Logical flag for verbosity.
...

Arguments to be passed to methods.

Details

Number of rows of object must match the length of probes.

Value

backgroundCorrectionMethods and normalizationMethods will return a character vector with the methods implemented currently.

backgroundCorrect, normalize and normalizeToTarget will return a matrix with same dimensions as the input matrix. If they are applied to a FeatureSet object, the PM matrix will be used as input.

The summarize method will return a matrix with length(unique(probes)) rows and ncol(object) columns.

Examples

ns <- 100
nps <- 1000
np <- 10
intensities <- matrix(rnorm(ns*nps*np, 8000, 400), nc=ns)
ids <- rep(as.character(1:nps), each=np)
bgCorrected <- backgroundCorrect(intensities)
normalized <- normalize(bgCorrected)
summarizationMethods()
expression <- summarize(normalized, probes=ids)
intensities[1:20, 1:3]
expression[1:20, 1:3]
target <- rnorm(np*nps)
normalizedToTarget <- normalizeToTarget(intensities, target)

if (require(oligoData) & require(pd.hg18.60mer.expr)){
  ## Example of normalization with real data
  data(nimbleExpressionFS)
  boxplot(nimbleExpressionFS, main='Original')
  for (mtd in normalizationMethods()){
    message('Normalizing with ', mtd)
    res <- normalize(nimbleExpressionFS, method=mtd, verbose=FALSE)
    boxplot(res, main=mtd)
  }
}