Package 'affyPLM'

LESN - Low End Signal is Noise Background corrections

Description

This function background corrects PM probe data using LESN - Low End Signal is Noise concepts.

Usage

bg.correct.LESN(object, method=2, baseline=0.25, theta=4)

Arguments

object	an AffyBatch
method	an integer code specifying which method to use
baseline	A baseline value to use
theta	A parameter used in the background correction process

Details

This method will be more formally documented at a later date.

The basic concept is to consider that the lowest end of intensites is most likely just noise (and should be heavily corrected) and the highest end signals are most likely signal and should have little adjustment. Low end signals are made much smaller while high end signals get less adjustment relative adjustment.

Value

An AffyBatch

Author(s)

Ben Bolstad [email protected]

References

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

Examples

if (require(affydata)) {
  data(Dilution)
  Dilution.example.bgcorrect <- bg.correct.LESN(Dilution)
}

---

Fit a Probe Level Model to Affymetrix Genechip Data.

Description

This function converts an AffyBatch into an PLMset by fitting a specified robust linear model to the probe level data.

Usage

fitPLM(object,model=PM ~ -1 + probes +samples,
       variable.type=c(default="factor"),
       constraint.type=c(default="contr.treatment"),
       subset=NULL,
       background=TRUE, normalize=TRUE, background.method="RMA.2",
       normalize.method="quantile", background.param=list(),
       normalize.param=list(), output.param=verify.output.param(),
       model.param=verify.model.param(object, model),
       verbosity.level=0)

Arguments

object	an AffyBatch
model	A formula describing the model to fit. This is slightly different from the standard method of specifying formulae in R. Read the description below
variable.type	a way to specify whether variables in the model are factors or standard variables
constraint.type	should factor variables sum to zero or have first variable set to zero (endpoint constraint)
subset	a vector with the names of probesets to be used. If NULL then all probesets are used.
normalize	logical value. If TRUE normalize data using quantile normalization
background	logical value. If TRUE background correct using RMA background correction
background.method	name of background method to use.
normalize.method	name of normalization method to use.
background.param	A list of parameters for background routines
normalize.param	A list of parameters for normalization routines
output.param	A list of parameters controlling optional output from the routine.
model.param	A list of parameters controlling model procedure
verbosity.level	An integer specifying how much to print out. Higher values indicate more verbose. A value of 0 will print nothing

Details

This function fits robust Probe Level linear Models to all the probesets in an AffyBatch. This is carried out on a probeset by probeset basis. The user has quite a lot of control over which model is used and what outputs are stored. For more details please read the vignette.

Value

An PLMset

Author(s)

Ben Bolstad [email protected]

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

expresso, rma, threestep

Examples

if (require(affydata)) {
  data(Dilution)
  Pset <- fitPLM(Dilution, model=PM ~ -1 + probes + samples)
  se(Pset)[1:5,]

  image(Pset)
  NUSE(Pset) 

  #now lets try a wider class of models
  ## Not run: Pset <- fitPLM(Dilution,model=PM ~ -1 + probes +liver,
  normalize=FALSE,background=FALSE)
## End(Not run) 
  ## Not run: coefs(Pset)[1:10,]

  ## Not run: Pset <- fitPLM(Dilution,model=PM ~ -1 + probes + liver +
  scanner, normalize=FALSE,background=FALSE)
## End(Not run)
  coefs(Pset)[1:10,]

  #try liver as a covariate
  logliver <- log2(c(20,20,10,10))
  ## Not run: Pset <- fitPLM(Dilution, model=PM~-1+probes+logliver+scanner,
  normalize=FALSE, background=FALSE, variable.type=c(logliver="covariate"))
## End(Not run) 
  coefs(Pset)[1:10,]

  #try a different se.type
  ## Not run: Pset <- fitPLM(Dilution, model=PM~-1+probes+scanner,
  normalize=FALSE,background=FALSE,m odel.param=list(se.type=2))
## End(Not run) 
  se(Pset)[1:10,]
}

---

Relative M vs. A plots

Description

Create boxplots of M or M vs A plots. Where M is determined relative to a specified chip or to a pseudo-median reference chip.

Arguments

...	Additional parameters for the routine
A	A vector to plot along the horizonal axis
M	A vector to plot along vertical axis
subset	A set of indices to use when drawing the loess curve
show.statistics	If true some summary statistics of the M values are drawn
span	span to be used for loess fit.
family.loess	"guassian" or "symmetric" as in loess.
cex	Size of text when writing summary statistics on plot

See Also

mva.pairs

---

Normalization applied to ExpressionSets

Description

Allows the user to apply normalization routines to ExpressionSets.

Usage

normalize.ExpressionSet.quantiles(eset, transfn=c("none","log","antilog"))
  normalize.ExpressionSet.loess(eset, transfn=c("none","log","antilog"),...)
  normalize.ExpressionSet.contrasts(eset, span = 2/3,
      choose.subset=TRUE, subset.size=5000, verbose=TRUE, family="symmetric",
      transfn=c("none","log","antilog")) 
  normalize.ExpressionSet.qspline(eset, transfn=c("none","log","antilog"),...)
  normalize.ExpressionSet.invariantset(eset,prd.td=c(0.003, 0.007),
      verbose=FALSE, transfn=c("none","log","antilog"),
      baseline.type=c("mean","median","pseudo-mean","pseudo-median")) 
  normalize.ExpressionSet.scaling(eset, trim=0.02, baseline=-1,
      transfn=c("none","log","antilog"))

Arguments

eset	An ExpressionSet
span	parameter to be passed to the function loess.
choose.subset	use a subset of values to establish the normalization relationship
subset.size	number to use for subset
verbose	verbosity flag
family	parameter to be passed to the function loess.
prd.td	cutoff parameter (details in the bibliographic reference)
trim	How much to trim from the top and bottom before computing the mean when using the scaling normalization
baseline	Index of array to use as baseline, negative values (-1,-2,-3,-4) control different baseline selection methods
transfn	Transform the ExpressionSet before normalizing. Useful when dealing with expression values that are log-scale
baseline.type	A method of selecting the baseline array
...	Additional parameters that may be passed to the normalization routine

Details

This function carries out normalization of expression values. In general you should either normalize at the probe level or at the expression value level, not both.

Typing normalize.ExpressionSet.methods should give you a list of methods that you may use. note that you can also use the normalize function on ExpressionSets. Use method to select the normalization method.

Value

A normalized ExpressionSet.

Author(s)

Ben Bolstad, [email protected]

References

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

Examples

if (require(affydata)) {
  data(Dilution)
  eset <- rma(Dilution, normalize=FALSE, background=FALSE)
  normalize(eset)
}

---

Quantile Normalization applied to probesets

Description

Using a normalization based upon quantiles, this function normalizes a matrix of probe level intensities.

Usage

normalize.AffyBatch.quantiles.probeset(abatch,type=c("separate","pmonly","mmonly","together"),use.median=FALSE,use.log=TRUE)

Arguments

abatch	An AffyBatch
type	how should MM and PM values be handled
use.median	use median rather than mean
use.log	take logarithms, then normalize

Details

This function applies the quantile method in a probeset specific manner.

In particular a probeset summary is normalized using the quantile method and then the probes adjusted accordingly.

Value

A normalized AffyBatch.

Author(s)

Ben Bolstad, [email protected]

References

Bolstad, B (2001) Probe Level Quantile Normalization of High Density Oligonucleotide Array Data. Unpublished manuscript http://oz.berkeley.edu/~bolstad/stuff/qnorm.pdf

Bolstad, B. M., Irizarry R. A., Astrand, M, and Speed, T. P. (2003) A Comparison of Normalization Methods for High Density Oligonucleotide Array Data Based on Bias and Variance. Bioinformatics 19(2) ,pp 185-193. http://www.stat.berkeley.edu/~bolstad/normalize/normalize.html

See Also

normalize.quantiles

---

Scaling normalization

Description

Allows the user to apply scaling normalization.

Usage

normalize.scaling(X,trim=0.02, baseline=-1, log.scalefactors=FALSE)
normalize.AffyBatch.scaling(abatch,
    type=c("together","pmonly","mmonly","separate"),
    trim=0.02, baseline=-1, log.scalefactors=FALSE)

Arguments

X	A matrix. The columns of which are to be normalized.
abatch	An AffyBatch
type	A parameter controlling how normalization is applied to the Affybatch.
trim	How much to trim from the top and bottom before computing the mean when using the scaling normalization.
baseline	Index of array to use as baseline, negative values (-1,-2,-3,-4) control different baseline selection methods.
log.scalefactors	Compute the scale factors based on log2 transformed data.

Details

These function carries out scaling normalization of expression values.

Value

A normalized ExpressionSet.

Author(s)

Ben Bolstad, [email protected]

Examples

if (require(affydata)) {
  data(Dilution)
  normalize.AffyBatch.scaling(Dilution)
}

---

Class PLMset

Description

This is a class representation for Probe level Linear Models fitted to Affymetrix GeneChip probe level data.

Objects from the Class

Objects can be created using the function fitPLM

Slots

probe.coefs:

Object of class "matrix". Contains model coefficients related to probe effects.

se.probe.coefs:

Object of class "matrix". Contains standard error estimates for the probe coefficients.

chip.coefs:

Object of class "matrix". Contains model coefficients related to chip (or chip level) effects for each fit.

se.chip.coefs:

Object of class "matrix". Contains standard error estimates for the chip coefficients.

const.coefs:

Object of class "matrix". Contains model coefficients related to intercept effects for each fit.

se.const.coefs:

Object of class "matrix". Contains standard error estimates for the intercept estimates

model.description:

Object of class "character". This string describes the probe level model fitted.

weights:

List of objects of class "matrix". Contains probe weights for each fit. The matrix has columns for chips and rows are probes.

phenoData:

Object of class "phenoData" This is an instance of class phenoData containing the patient (or case) level data. The columns of the pData slot of this entity represent variables and the rows represent patients or cases.

annotation

A character string identifying the annotation that may be used for the ExpressionSet instance.

experimentData:

Object of class "MIAME". For compatibility with previous version of this class description can also be a "character". The class characterOrMIAME has been defined just for this.

cdfName:

A character string giving the name of the cdfFile.

nrow:

Object of class "numeric". Number of rows in chip.

ncol:

Object of class "numeric". Number of cols in chip.

narrays:

Object of class "numeric". Number of arrays used in model fit.

normVec:

Object of class "matrix". For storing normalization vector(s). Not currentl used

varcov:

Object of class "list". A list of variance/covariance matrices.

residualSE:

Object of class "matrix". Contains residual standard error and df.

residuals:

List of objects of class "matrix". Contains residuals from model fit (if stored).

model.call:

Object of class "call"

Methods

weights<-

signature(object = "PLMset"): replaces the weights.

weights

signature(object = "PLMset"): extracts the model fit weights.

coefs<-

signature(object = "PLMset"): replaces the chip coefs.

coefs

signature(object = "PLMset"): extracts the chip coefs.

se

signature(object = "PLMset"): extracts the standard error estimates of the chip coefs.

se<-

signature(object = "PLMset"): replaces the standard error estimates of the chip coefs.

coefs.probe

signature(object = "PLMset"): extracts the probe coefs.

se.probe

signature(object = "PLMset"): extracts the standard error estimates of the probe coefs.

coefs.const

signature(object = "PLMset"): extracts the intercept coefs.

se.const

signature(object = "PLMset"): extracts the standard error estimates of the intercept coefs.

getCdfInfo

signature(object = "PLMset"): retrieve the environment that defines the location of probes by probe set.

image

signature(x = "PLMset"): creates an image of the robust linear model fit weights for each sample.

indexProbes

signature(object = "PLMset", which = "character"): returns a list with locations of the probes in each probe set. The list names defines the probe set names. which can be "pm", "mm", or "both". If "both" then perfect match locations are given followed by mismatch locations.

Mbox

signature(object = "PLMset"): gives a boxplot of M's for each chip. The M's are computed relative to a "median" chip.

normvec

signature(x = "PLMset"): will return the normalization vector (if it has been stored).

residSE

signature(x = "PLMset"): will return the residual SE (if it has been stored).

boxplot

signature(x = "PLMset"): Boxplot of Normalized Unscaled Standard Errors (NUSE).

NUSE

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

RLE|

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

Note

This class is better described in the vignette.

Author(s)

B. M. Bolstad [email protected]

References

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

---

Convert a PLMset to an ExpressionSet

Description

This function converts a PLMset to an ExpressionSet. This is often useful since many Bioconductor functions operate on ExpressionSet objects.

Usage

PLMset2exprSet(pset)
pset2eset(pset)

Arguments

pset	The PLMset to convert to ExpressionSet.

Details

These functions convert PLMset objects to ExpressionSet objects. This is often useful since many Bioconductor functions operate on ExpressionSet objects. Note that the function pset2eset is a wrapper for PLMset2exprSet.

Value

returns a ExpressionSet

Author(s)

Ben Bolstad [email protected]

See Also

ExpressionSet

Examples

if (require(affydata)) {
  data(Dilution)
  Pset <- fitPLM(Dilution)
  eset <- pset2eset(Pset)
}

---

Background correct and Normalize

Description

This function pre-processes an AffyBatch.

Usage

preprocess(object, subset=NULL, normalize=TRUE, background=TRUE,
           background.method="RMA.2", normalize.method="quantile",
           background.param=list(), normalize.param=list(),
           verbosity.level=0)

Arguments

object	an AffyBatch
subset	a vector with the names of probesets to be used. If NULL then all probesets are used.
normalize	logical value. If TRUE normalize data using quantile normalization
background	logical value. If TRUE background correct using RMA background correction
background.method	name of background method to use.
normalize.method	name of normalization method to use.
background.param	list of parameters for background correction methods
normalize.param	list of parameters for normalization methods
verbosity.level	An integer specifying how much to print out. Higher values indicate more verbose. A value of 0 will print nothing

Details

This function carries out background correction and normalization pre-processing steps. It does not summarize to produce gene expression measures. All the same pre-processing methods supplied by threestep are supported by this function.

Value

An AffyBatch

Author(s)

Ben Bolstad [email protected]

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

expresso, rma

Examples

if (require(affydata)) {
  data(Dilution)

  # should be equivalent to the bg and norm of rma()
  abatch.preprocessed <- preprocess(Dilution)
}

---

Coloring pseudo chip images

Description

These are routines used for coloring pseudo chip images.

Usage

pseudoPalette(low = "white", high = c("green", "red"), mid = NULL,k =50) 
  pseudoColorBar(x, horizontal = TRUE, col = heat.colors(50), scale = 1:length(x),k = 11, log.ticks=FALSE,...)

Arguments

low	color at low end of scale
high	color at high end of scale
mid	color at exact middle of scale
k	number of colors to have
x	A data series
horizontal	If TRUE then color bar is to be draw horizontally
col	colors for color bar
scale	tickmarks for x if x is not numeric
log.ticks	use a log type transformation to assign the colors
...	additional parameters to plotting routine

Details

Adapted from similar tools in maPlots pacakge.

Author(s)

Ben Bolstad [email protected]

See Also

AffyBatch, read.affybatch

---

Read RMAExpress computed expression values

Description

Read RMAExpress computed binary output files into a matrix or ExpressionSet

Usage

ReadRMAExpress(filename, return.value=c("ExpressionSet","matrix"))

Arguments

filename	The name of the file containing RMAExpress output to be read in
return.value	should a matrix or an ExpressionSet be returned

Value

returns an ExpressionSet

Author(s)

Ben Bolstad [email protected]

References

http://rmaexpress.bmbolstad.com

---

Fit a RMA to Affymetrix Genechip Data as a PLMset

Description

This function converts an AffyBatch into an PLMset by fitting a multichip model. In particular we concentrate on the RMA model.

Usage

rmaPLM(object, subset=NULL, normalize=TRUE, background=TRUE,
       background.method="RMA.2", normalize.method="quantile",
       background.param=list(), normalize.param=list(), output.param=list(),
       model.param=list(), verbosity.level=0)

Arguments

object	an AffyBatch
subset	a vector with the names of probesets to be used. If NULL then all probesets are used.
normalize	logical value. If TRUE normalize data using quantile normalization
background	logical value. If TRUE background correct using RMA background correction
background.method	name of background method to use.
normalize.method	name of normalization method to use.
background.param	A list of parameters for background routines
normalize.param	A list of parameters for normalization routines
output.param	A list of parameters controlling optional output from the routine.
model.param	A list of parameters controlling model procedure
verbosity.level	An integer specifying how much to print out. Higher values indicate more verbose. A value of 0 will print nothing

Details

This function fits the RMA as a Probe Level Linear models to all the probesets in an AffyBatch.

Value

An PLMset

Author(s)

Ben Bolstad [email protected]

References

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

Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B and Speed TP (2003) Summaries of Affymetrix GeneChip probe level data Nucleic Acids Research 31(4):e15

Bolstad, BM, Irizarry RA, Astrand, M, and Speed, TP (2003) A Comparison of Normalization Methods for High Density Oligonucleotide Array Data Based on Bias and Variance. Bioinformatics 19(2):185-193

See Also

expresso, rma, threestep,fitPLM, threestepPLM

Examples

if (require(affydata)) {
  # A larger example testing weight image function
  data(Dilution)
  ## Not run: Pset <- rmaPLM(Dilution,output.param=list(weights=TRUE))
  ## Not run: image(Pset)
}

---

Three Step expression measures

Description

This function converts an AffyBatch into an ExpressionSet using a three step expression measure.

Usage

threestep(object, subset=NULL, normalize=TRUE, background=TRUE,
          background.method="RMA.2", normalize.method="quantile",
          summary.method="median.polish", background.param=list(),
          normalize.param=list(), summary.param=list(), verbosity.level=0)

Arguments

object	an AffyBatch.
subset	a vector with the names of probesets to be used. If NULL, then all probesets are used.
normalize	logical value. If TRUE normalize data using quantile normalization
background	logical value. If TRUE background correct using RMA background correction
background.method	name of background method to use.
normalize.method	name of normalization method to use.
summary.method	name of summary method to use.
background.param	list of parameters for background correction methods.
normalize.param	list of parameters for normalization methods.
summary.param	list of parameters for summary methods.
verbosity.level	An integer specifying how much to print out. Higher values indicate more verbose. A value of 0 will print nothing.

Details

This function computes the expression measure using threestep methods. Greater details can be found in a vignette.

Value

An ExpressionSet

Author(s)

Ben Bolstad [email protected]

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

expresso, rma

Examples

if (require(affydata)) {
  data(Dilution)

  # should be equivalent to rma()
  eset <- threestep(Dilution)

  # Using Tukey Biweight summarization
  eset <- threestep(Dilution, summary.method="tukey.biweight")

  # Using Average Log2 summarization
  eset <- threestep(Dilution, summary.method="average.log")

  # Using IdealMismatch background and Tukey Biweight and no normalization.
  eset <- threestep(Dilution, normalize=FALSE,background.method="IdealMM",
                    summary.method="tukey.biweight")

  # Using average.log summarization and no background or normalization.
  eset <- threestep(Dilution, background=FALSE, normalize=FALSE,
                    background.method="IdealMM",summary.method="tukey.biweight")

  # Use threestep methodology with the rlm model fit
  eset <- threestep(Dilution, summary.method="rlm")

  # Use threestep methodology with the log of the average
  # eset <- threestep(Dilution, summary.method="log.average")

  # Use threestep methodology with log 2nd largest method
  eset <- threestep(Dilution, summary.method="log.2nd.largest")

  eset <- threestep(Dilution, background.method="LESN2")
}

---

Three Step expression measures returned as a PLMset

Description

This function converts an AffyBatch into an PLMset using a three step expression measure.

Usage

threestepPLM(object,subset=NULL, normalize=TRUE, background=TRUE,
             background.method="RMA.2", normalize.method="quantile",
             summary.method="median.polish", background.param = list(),
             normalize.param=list(), output.param=list(),
             model.param=list(), verbosity.level=0)

Arguments

object	an AffyBatch
subset	a vector with the names of probesets to be used. If NULL then all probesets are used.
normalize	logical value. If TRUE normalize data using quantile normalization
background	logical value. If TRUE background correct using RMA background correction
background.method	name of background method to use.
normalize.method	name of normalization method to use.
summary.method	name of summary method to use.
background.param	list of parameters for background correction methods
normalize.param	list of parameters for normalization methods
output.param	list of parameters for output methods
model.param	list of parameters for model methods
verbosity.level	An integer specifying how much to print out. Higher values indicate more verbose. A value of 0 will print nothing

Details

This function computes the expression measure using threestep methods. It returns a PLMset. The most important difference is that the PLMset allows you to access the residuals which the threestep function does not do.

Value

An PLMset

Author(s)

Ben Bolstad [email protected]

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

expresso, rma, threestep, rmaPLM, fitPLM

Examples

if (require(affydata)) {
  data(Dilution)

  # should be equivalent to rma()
  ## Not run: eset <- threestepPLM(Dilution)
}

---