Package 'MiChip'

Create Boxplot of data

Description

Creates a boxplot of expression data contained in a matrix and writes this to a file.

Usage

boxplotData(dmat, exptname, dlevel)

Arguments

dmat	matrix containing expression data to be boxplotted
exptname	Name of the experiment, used to build filename
dlevel	Stage of the experiment e.g. raw, summarized, normalized

Examples

#Create a file of a boxplot containing normalized expression data for myexpt
## Not run: 
boxplotData(dmat, "MyExpt", "mednormed")

## End(Not run)

---

Create Boxplot of data

Description

Creates a boxplot of expression data contained in a matrix.

Usage

boxplotDataNoFile(dmat, exptname, dlevel)

Arguments

dmat	matrix containing expression data to be boxplotted
exptname	Name of the experiment, used to build plot title
dlevel	Stage of the experiment e.g. raw, summarized, normalized

Examples

#Create a boxplot containing normalized expression data for myexpt
## Not run: boxplotDataNoFile(dmat, "MyExpt", "mednormed")

---

Corrects for spots flagged as not present

Description

Spots flagged with a -ve quality flag value by the scanner may be regarded as not present. This method sets their intensity to NA.

Usage

correctForFlags(eset, intensityCutoff=0)

Arguments

eset	ExpressionSet containing intensity values and flags to be filtered
intensityCutoff	value of lowest acceptable intensity value in the experiment

Examples

#Correct ExpressionSet for flags of spots marked as unreadable
## Not run: 
myCorrectedEset <-correctForFlags(eset, intensityCutoff=0)

## End(Not run)

---

Introduction to the MiChip Package

Description

Alibrary for processing MiChip hybridizations

Author(s)

Jonathon Blake

---

Produce Median from Probe Intensity values

Description

Creates a median to summarize the intensities for individual probes, giving that not all probes will have a valid intensity reading

Usage

myForgivingMedian(mat, minSumlength=0)

Arguments

mat	matrix of data to calculate the median from
minSumlength	The lowest acceptable length of the matrix to calculate a median

Examples

#Calculate the median of a matrix omiting NAs
## Not run: 
myForgivingMedian(mat, minSumlength=0)

## End(Not run)

---

Calculates the median of an array excluding NAs

Description

Calculates the median of an array excluding NAs

Usage

naOmitMedian(mat, madAdjust=FALSE)

Arguments

mat	A single dimensional matrix
madAdjust	if TRUE then summarized data will be filtered according to the MAD median absolute deviation and set to NA if the median is less than MAD

Examples

#Calculate the median of matrix mat omiting NAs
## Not run: 
	myMedian <-naOmitMedian(mat, madAdjust=TRUE)

## End(Not run)

---

Normalize to median intensity

Description

Normalizes intensity values to the median of each chip

Usage

normalizePerChipMedian(eset)

Arguments

eset	ExpressionSet containing chip intensity values to be normalized

Examples

#Normalize expression data in an Eset to the median
## Not run: 
normedDataEset <- normalizePerChipMedian(eset)

## End(Not run)

---

Outputs a tab delimited file from an ExpressionSet

Description

Takes an ExpressionSet and outputs a tab delimited file containing feature annotation to the left and hyb specific expression/flag data to the right

Usage

outputAnnotatedDataMatrix(eset, exptname, stage, dataElement)

Arguments

eset	ExpressionSet containing the matrix and annotation to output
exptname	a string containing the name of the experiment. Used to build file name
stage	a string containing the stage of the data in the matrix e.g. normalized
dataElement	a string containing the name of the data element in the ExpressionSet to be output

Examples

#Write out an annotated tab delimited file for the normalized data
## Not run: 
outputAnnotatedDataMatrix(normedEset, "MyMicroArrayExpt","Median_Normalized", "exprs")
## End(Not run)

---

Add Pearson Correlation value to plots

Description

Adds a pearson correlation value to the scatter plots

Usage

panelCor(x,y, digits=2, prefix="r=")

Arguments

x	matrix of x values
y	matrix of y values to correlate with x
digits	number of digits to display
prefix	The string prefix that should be display on the scatterplot panel

Examples

#Calculate the median of a matrix omiting NAs
## Not run: 
panelCor(x,y, digits=2, prefix="r=")

## End(Not run)

---

Parse raw data files to create an ExpressionSet

Description

Loads all the gpr scanner output files in a particular directory and returns an ExpressionSet of the hybridizations in a MiChip experiment

Usage

parseRawData(datadir=".", pat="gpr")

Arguments

datadir	a directory containing one or my files of scanner output from MiChip hybridizations
pat	a string containing the three letter extension of the scanner output files

Examples

## Not run: 
## Load all *.gpr files in current directory
parseRawData(datadir=".", pat="gpr")


## Load all *.gpr files in a specified directory , windows
parseRawData(datadir="c:\mydata\grpdata\expt1\", pat="gpr")


## Load all *.gpr files in a specified directory, linux
parseRawData(datadir="/home/myuser/gprdata/extp1/", pat="gpr")

## End(Not run)

---

Plot pairwise intensity scatter

Description

Creates a pairwise set of scatter plots from a data matrix and writes it out to file

Usage

plotIntensitiesScatter(dmat, controls=NULL, exptname, maintitle)

Arguments

dmat	matrix containing data from an experiment to be plotted
controls	matrix of row numbers containing control data to be plotted in a different colour
exptname	Name of the experiment, used for build the filename
maintitle	String used to build the maintitle of the graph

Examples

#Plot the pairwise intensities from myexpt
## Not run: 
plotIntensitiesScatter(dmat, NULL, "MyExpt", "Median_Normalized")

## End(Not run)

---

Removes unwanted rows from data matrix

Description

Due to the requirements of spotting the chips, some of the spots are empty. Others contain controls or features from another species that may not be wanted in the analysis. This method removes them

Usage

removeUnwantedRows(rawData, filters)

Arguments

rawData	ExpressionSet containing matrix of data to be filtered
filters	list of strings to be to be filtered from annotation gene name column

Examples

#Removes empty and control spots from data matrix
## Not run: 
filters=c("empty", "control")
filteredData <- removeUnwantedRows(rawData, filters)
## End(Not run)

---

returns and annotated data matrix from an ExpressionSet

Description

Takes an ExpressionSet and returns a data martix of feature annotation to the left and hyb specific expression/flag data to the right

Usage

returnAnnotatedDataMatrix(eset, dataElement)

Arguments

eset	ExpressionSet containing the matrix and annotation to output
dataElement	a string containing the name of the data element in the ExpressionSet to be output

Examples

#Write out an annotated tab delimited file for the normalized data
## Not run: 
returnAnnotatedDataMatrix(normedEset,"exprs")
## End(Not run)

---

Sets a cutoff for the lowest intensity value

Description

Any value less than the cutoff value will be set to NA. This allows near background intensity values to be excluded

Usage

setIntensityCutoff(dmat, intensityCutoff)

Arguments

dmat	matrix of intensity values to which the cutoff value is applied
intensityCutoff	value of lowest acceptable intensity value in the experiment

Examples

#Set all the values under 50 in a matrix to NA
## Not run: 
dmatOver50 <- setIntensityCutoff(dmat, 50)

## End(Not run)

---

Removes a standard list of features for MiChip processing

Description

Removes all empty spots, control spots, U6 RNA, non human spots from an ExpressionSet in the standard fashion. A wrapper for removeUnwantedRows

Usage

standardRemoveRows(rawData)

Arguments

rawData

ExpressionSet containing the matrix to be filtered

Examples

#Filter standard rows from an ExpressionSet
## Not run: 
myfilterdEset <-standardRemoveRows(rawData)

## End(Not run)

---

Summarizes the probe intensity as median of replicates spotted

Description

As the probes are spotted onto the in quaduplet or duplicate the values have to be combined in some way. This function takes the median of the intensities for the spots. Effectively the mean for duplicates. If less than half of the spots are present an NA is added

Usage

summarizeIntensitiesAsMedian(eset,minSumlength=0, madAdjust=FALSE)

Arguments

eset	ExpressionSet containing probe intensity data to be summarized
minSumlength	The lowest acceptable length of the matrix to calculate a median
madAdjust	if TRUE then summarized data will be filtered according to the MAD median absolute deviation and set to NA if the median is less than MAD

Examples

#Calculate the median of a matrix omiting NAs
## Not run: 
summarizeIntensitiesAsMedian(eset,minSumlength=0,madAdjust=TRUE)

## End(Not run)

---

Worked Example of MiChip Processing

Description

Loads a set of hybridizations into a matrix and them proceeds to filter, summarize and median normalize them

Usage

workedExampleMedianNormalize(exptname, intensityCutoff=0, datadir=".", minSumlength, madAdjust = FALSE)

Arguments

exptname	string indicating the name of the experiment
intensityCutoff	The intensity value for accepting the spots intensity value in the experiment
datadir	The directory where hybridization files are found.
minSumlength	Minimum exceptable number of values to summarize intensity value.
madAdjust	if TRUE then summarized data will be filtered according to the MAD median absolute deviation and set to NA if the median is less than MAD

Examples

#Normalize data in the current directory to the median per chip
datadir <- system.file("extdata", package="MiChip")
myNormedEset <-workedExampleMedianNormalize("MyExpt", intensityCutoff=0, datadir, minSumlength=0, madAdjust=TRUE)

---

Worked Example of MiChip Processing

Description

Loads a set of hybridizations into a matrix and them proceeds to filter and summarize these data

Usage

workedExampleNotNormalizedData(exptname, intensityCutoff=0, datadir=".", minSumlength, madAdjust = FALSE )

Arguments

exptname	string indicating the name of the experiment
intensityCutoff	The intensity value for accepting the spots intensity value in the experment
datadir	The directory contain data from the experiment
minSumlength	Minimum exceptable number of values to summarize intensity value.
madAdjust	if TRUE then summarized data will be filtered according to the MAD median absolute deviation and set to NA if the median is less than MAD

Examples

#Summarizes the data in the current directory 
## Not run: 
mySummarizedEset <-workedExampleNotNormalizedData("MyExpt", intensityCutoff=0, datadir=".", minSumlength=0, madAdjust=TRUE)

## End(Not run)

---