| Title: | Processing and Differential Expression Analysis of Agilent microRNA chips |
|---|---|
| Description: | Processing and Analysis of Agilent microRNA data |
| Authors: | Pedro Lopez-Romero <[email protected]> |
| Maintainer: | Pedro Lopez-Romero <[email protected]> |
| License: | GPL-3 |
| Version: | 2.55.0 |
| Built: | 2024-07-03 05:36:25 UTC |
| Source: | https://github.com/bioc/AgiMicroRna |
Differential expression analysis using the linear model features implemented in the limma package. A linear model is fitted to each miRNA gene so that the fold change between different experimental conditions and their standard errors can be estimated. Empirical Bayes methods are applied to obtain moderated statistics
basicLimma(eset, design, CM,verbose = FALSE)basicLimma(eset, design, CM,verbose = FALSE)
eset |
ExpressionSet containing the processed log-expression values |
design |
design matrix |
CM |
contrast matrix |
verbose |
logical, if |
In our data example (see the target file in Table 1 in vignette), we have used a paired design (by subject) to assess the differential expression between two treatments B and C vs a control treatment A. That is, we want to obtain the microRNAS that are differentially expressed between conditions A vs B and A vs C. The linear model that we are going to fit to every miRNA is defined by equation: y = Treatment + Subject + error term. This model is going to estimate the treatment effect and then, the comparison between the different treatments are done in terms of contrasts between the estimates of the treatment effects. To fit the model, we need first to define a design matrix. The design matrix is an incidence matrix that relates each array/sample/file to its given experimental conditions, in our case, relates each file to one of the three treatments and with its particular subject. If treatment is a factor variable, we can define de desing matrix using model.matrix(~ -1 + treatment + subject). Then the linear model can be fitted using fit=lmFit(eset,design). This will get the treatment estimates for each microRNA in the eset object:
treatmentA treatmentB treatmentC subject2 hsa-miR-152 7.5721 7.656 7.566 -0.1157 hsa-miR-15a* 0.9265 1.066 1.211 -0.2242 hsa-miR-337-5p 6.2448 7.298 7.084 -0.4489
We can define the contrasts of interest using a contrast matrix as in CM=cbind(MSC\_AvsMSC\_B=c(1,-1,0), MSC\_AvsMSC\_C=c(1,0,-1))
And then, we can estimate those contrats using fit2=contrasts.fit(fit,CM). Finally, we can obtain moderated statistics using fit2=eBayes(fit2).
The function 'basicLimma' implemented in AgiMicroRna produces the last fit2 object, that has in fit2\$coeff the M values, in fit\$t the moderated-t statistic of the contrasts, and in fit2\$p.value the corresponding p value of each particular contrasts. Be aware that these p values must be corrected by multiple testing.
MSC\_AvsMSC\_B MSC\_AvsMSC\_C hsa-miR-152 0.67567761 0.977326746 hsa-miR-15a* 0.68019442 0.413657270 hsa-miR-337-5p 0.03737814 0.075248741
See limmaUsersGuide() for a complete description of the limma package.
An MArrayLM object of the package limma
Pedro Lopez-Romero
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397–420.
Smyth, G. K. (2004). Linear models and empirical Bayes methods for assessing diferential expression in microarray experiments. Statistical Applications in Genetics and Molecular Biology, Vol. 3, No. 1, Article 3. http://www.bepress.com/sagmb/vol3/iss1/art3
An 'RGList' example containing proccesed data is in
ddPROC and an overview of how the processed data is produced
is given in filterMicroRna. The ExpressionSet object can be
generated using esetMicroRna
## Not run: data(targets.micro) data(ddPROC) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE,verbose=FALSE) levels.treatment=levels(factor(targets.micro$Treatment)) treatment=factor(as.character(targets.micro$Treatment), levels=levels.treatment) levels.subject=levels(factor(targets.micro$Subject)) subject=factor(as.character(targets.micro$Subject), levels=levels.subject) design=model.matrix(~ -1 + treatment + subject ) CM=cbind(MSC_AvsMSC_B=c(1,-1,0,0), MSC_AvsMSC_C=c(1,0,-1,0)) fit2=basicLimma(esetPROC,design,CM,verbose=TRUE) names(fit2) head(fit2$coeff) head(fit2$p.value) plot(fit2\$Amean,fit2$coeff[,1],xlab="A",ylab="M") abline(h=0) abline(h=c(-1,1),col="red") plot(fit2$coeff[,1],fit2$p.value[,1], xlab="M",ylab="p value") ## End(Not run)## Not run: data(targets.micro) data(ddPROC) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE,verbose=FALSE) levels.treatment=levels(factor(targets.micro$Treatment)) treatment=factor(as.character(targets.micro$Treatment), levels=levels.treatment) levels.subject=levels(factor(targets.micro$Subject)) subject=factor(as.character(targets.micro$Subject), levels=levels.subject) design=model.matrix(~ -1 + treatment + subject ) CM=cbind(MSC_AvsMSC_B=c(1,-1,0,0), MSC_AvsMSC_C=c(1,0,-1,0)) fit2=basicLimma(esetPROC,design,CM,verbose=TRUE) names(fit2) head(fit2$coeff) head(fit2$p.value) plot(fit2\$Amean,fit2$coeff[,1],xlab="A",ylab="M") abline(h=0) abline(h=c(-1,1),col="red") plot(fit2$coeff[,1],fit2$p.value[,1], xlab="M",ylab="p value") ## End(Not run)
It creates a Boxplot using the matrix columns as input
boxplotMicroRna(object, maintitle, colorfill, xlab, ylab)boxplotMicroRna(object, maintitle, colorfill, xlab, ylab)
object |
A matrix containing by columns the expression arrays in log2 scale |
maintitle |
character to indicate the title of the graph |
colorfill |
color to fill the boxplot |
xlab |
title for the x axe |
ylab |
title for the y axe |
Pedro Lopez-Romero
data(dd.micro) MMM=log2(dd.micro$meanS) boxplotMicroRna(MMM, maintitle="log2 Mean Signal", colorfill="green", xlab="Samples", ylab="expression")data(dd.micro) MMM=log2(dd.micro$meanS) boxplotMicroRna(MMM, maintitle="log2 Mean Signal", colorfill="green", xlab="Samples", ylab="expression")
Identifies replicated features at probe and at gene level and computes the coefficient of variation of the array
cvArray(ddDUP, foreground = c("MeanSignal", "ProcessedSignal"),targets,verbose=FALSE)cvArray(ddDUP, foreground = c("MeanSignal", "ProcessedSignal"),targets,verbose=FALSE)
ddDUP |
uRNAList, containing the output from |
foreground |
Specifies the signal used, only "MeanSignal" or "ProcessedSignal" can be used |
targets |
data.frame with the target structure |
verbose |
logical, if |
In the Agilent microRNA platforms the features are replicated at a probe level and normally, a single microRNA is interrogated by either two or four sets of replicated probes. The replication of the probes allows computing the coefficient of variation (CV) for each array as a measure of the reproducibility of the array. The CV is computed for every set of replicated probes and the CV median is reported as the array CV. A lower array CV indicates a better array reproducibility.
It prints out the results of the replication for the NON CONTROL FEATURES at a probe and gene level.
Pedro Lopez-Romero
## Not run: data(dd.micro) data(targets.micro) cvArray(dd.micro,"MeanSignal",targets.micro,verbose=TRUE) graphics.off() ## End(Not run)## Not run: data(dd.micro) data(targets.micro) cvArray(dd.micro,"MeanSignal",targets.micro,verbose=TRUE) graphics.off() ## End(Not run)
Data, extracted from scanned images using Agilent Feature Extraction Software, are stored in a uRNAList object.
data(dd.micro)data(dd.micro)
A data example is provided. The data example includes 3 experimental conditions with two replicates.
For these data, chips were scanned using the Agilent G2567AA Microarray Scanner System (Agilent Technologies) Image analysis and data collection were carried out using the Agilent Feature Extraction 9.1.3.1. (AFE).
Data, colected with the Agilent Feature Extraction Software, are stored in a uRNAList object with the following components:
matrix, 'gTotalGeneSignal'
matrix, 'gTotalProbeSignal'
matrix, 'gMeanSignal'
matrix, 'gProcessedSignal'
data.frame, 'FileName'
vector of characters, 'AGilent Probe Name'
vector of characters, 'microRNA Name'
vector of integers, '0'= Feature, '1'= Positive control, '-1'= Negative control
matrix, FLAG to classify signal if 'IsGeneDetected=1' or 'not=0'
matrix, FLAG to classify signal if 'IsSaturated = 1' or 'not=0'
matrix, FLAG to classify signal if 'IsFeatPopnOL = 0' or 'not=1'
matrix, FLAG to classify signal if 'gIsFeatNonUnifOL = 0' or 'not=1'
matrix, gBGMedianSignal
matrix, gBGUsed
Pedro Lopez-Romero
readMicroRnaAFE.Rd
Filtered and Normalized miRNA data stored in a uRNAList object.
data(ddPROC)data(ddPROC)
ddPROC is originated after the processing of the dd.micro raw data.
Pedro Lopez-Romero
An overview of how ddPROC is obtained is given in
filterMicroRna
Retrieve the number of rows (genes) and columns (arrays) for an uRNAList object.
## S3 method for class 'uRNAList' dim(x) ## S3 method for class 'uRNAList' length(x)## S3 method for class 'uRNAList' dim(x) ## S3 method for class 'uRNAList' length(x)
x |
an object of class |
This function and this file, has been borrowed from the files created by Gordon Smyth for the limma package.
Numeric vector of length 2. The first element is the number of rows (genes) and the second is the number of columns (arrays).
Pedro Lopez-Romero
Retrieve the dimension names of a microarray data object.
## S3 method for class 'uRNAList' dimnames(x) ## S3 replacement method for class 'uRNAList' dimnames(x) <- value## S3 method for class 'uRNAList' dimnames(x) ## S3 replacement method for class 'uRNAList' dimnames(x) <- value
x |
an object of class |
value |
a possible value for |
The dimension names of a microarray object are the same as those of the most important matrix component of that object.
A consequence is that rownames and colnames will work as expected.
This function and this file, has been borrowed from the files
created by Gordon Smyth for the limma package.
Either NULL or a list of length 2.
Pedro Lopez-Romero
It creates an 'ExpressionSet' object from a 'uRNAList' with unique probe names Tipically, the 'uRNAList object' contains the Total Gene Processed data
esetMicroRna(uRNAList, targets, makePLOT=FALSE,verbose=FALSE)esetMicroRna(uRNAList, targets, makePLOT=FALSE,verbose=FALSE)
uRNAList |
An uRNAList containing normally the processed data |
targets |
data.frame with the targets structure |
makePLOT |
logical, if |
verbose |
logical, if |
It creates an ExpressionSet object from a uRNAList. Usually this function is applied to an uRNAList object containing the Total Gene Processed data.
An ExpressionSet object
Pedro Lopez-Romero
An 'uRNAList' example containing proccesed data is in
ddPROC and an overview of how the processed data is produced
is given in filterMicroRna
Filter genes out according to their Quality Flag
filterMicroRna(ddNORM, dd, control, IsGeneDetected, wellaboveNEG, limIsGeneDetected, limNEG, makePLOT, targets, verbose, writeout)filterMicroRna(ddNORM, dd, control, IsGeneDetected, wellaboveNEG, limIsGeneDetected, limNEG, makePLOT, targets, verbose, writeout)
ddNORM |
uRNAList with the Total Gene Signal in log2 scale to be FILTERED out according to a Quality FLAG |
dd |
uRNAList, containing the output from |
control |
logical, if |
IsGeneDetected |
logical, if |
wellaboveNEG |
logical, if |
limIsGeneDetected |
for a given feature xi accros samples, is the minimum in at least one experimental condition with a IsGeneDetected-FLAG = 1 (Is Detected) |
limNEG |
for a given feature xi accros samples, is the minimum in at least one experimental condition with intensity > Limit established for negative controls (Mean + 1.5 x SD) |
makePLOT |
logical, if |
targets |
data.frame with the targets structure |
verbose |
logical, if |
writeout |
logical, if |
Agilent Feature Extraction software provides a flag for each spot that identifies different quantification errors of the signal. Quantification flags were used to filter out signals that did not reach a minimum established criterion of quality.
The function returns a uRNAList containing the FILTERED data. In order to allow the tracking of those microRNAs that may have been filtered out from the original raw data, the following files are given:
NOCtrl\_exprs.txt: Log2 Normalized Total Gene Signals for the Non Control Genes NOCtrl\_FlagIsGeneDetected.txt: IsGeneDetected Flag for the Non Control Genes, 1 = detected IsNOTGeneDetected.txt: Genes that not are not detected according to IsGeneDetected Flag
Pedro Lopez-Romero
Agilent Feature Extraction Reference Guide http://www.Agilent.com
data(dd.micro,verbose=FALSE) data(targets.micro,verbose=FALSE) ddTGS=tgsMicroRna(dd.micro,half=TRUE,makePLOT=FALSE,verbose=FALSE) ddNORM=tgsNormalization(ddTGS,'quantile', makePLOTpre=FALSE,makePLOTpost=TRUE,targets.micro,verbose=FALSE) ddPROC=filterMicroRna(ddNORM, dd.micro, control=TRUE, IsGeneDetected=TRUE, wellaboveNEG=FALSE, limIsGeneDetected=50, limNEG=25, makePLOT=FALSE, targets.micro, verbose=FALSE, writeout=FALSE)data(dd.micro,verbose=FALSE) data(targets.micro,verbose=FALSE) ddTGS=tgsMicroRna(dd.micro,half=TRUE,makePLOT=FALSE,verbose=FALSE) ddNORM=tgsNormalization(ddTGS,'quantile', makePLOTpre=FALSE,makePLOTpost=TRUE,targets.micro,verbose=FALSE) ddPROC=filterMicroRna(ddNORM, dd.micro, control=TRUE, IsGeneDetected=TRUE, wellaboveNEG=FALSE, limIsGeneDetected=50, limNEG=25, makePLOT=FALSE, targets.micro, verbose=FALSE, writeout=FALSE)
It Uses the decideTests function of the 'limma' package to
classify the list of genes as up, down or not significant after
correcting by the multiplicity of the tests.
getDecideTests(fit2, DEmethod, MTestmethod, PVcut,verbose=FALSE)getDecideTests(fit2, DEmethod, MTestmethod, PVcut,verbose=FALSE)
fit2 |
MArrayLM object |
DEmethod |
method for |
MTestmethod |
method for multiple test, choices are 'none','BH', 'BY', ...
see |
PVcut |
p value threshold to declare significant features |
verbose |
logical, if |
A 'TestResults' object of the 'limma' package It prints out the number of UP and DOWN genes for every contrasts according to the p value limit specified
Pedro Lopez-Romero
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397–420.
An overview of miRNA differential expression analysis is given in
basicLimma
## Not run: DE=getDecideTests(fit2, DEmethod="separate", MTestmethod="BH", PVcut=0.10, verbose=TRUE) ## End(Not run)## Not run: DE=getDecideTests(fit2, DEmethod="separate", MTestmethod="BH", PVcut=0.10, verbose=TRUE) ## End(Not run)
Creates a HeatMap graph using the 'heatmap.2' function
HeatMapMicroRna(object, size, maintitle)HeatMapMicroRna(object, size, maintitle)
object |
A expression Matrix |
size |
number of highest variance genes to be considered in the plot |
maintitle |
title of the plot |
Pedro Lopez-Romero
heatmap.2
data(ddPROC) HeatMapMicroRna(ddPROC$TGS, size=100, maintitle="100 High Var genes")data(ddPROC) HeatMapMicroRna(ddPROC$TGS, size=100, maintitle="100 High Var genes")
Hierarchical cluster of samples using the 'hclust' function
hierclusMicroRna(object, GErep, methdis, methclu,sel, size)hierclusMicroRna(object, GErep, methdis, methclu,sel, size)
object |
An expression Matrix |
GErep |
Numerical vector that relates each sample with its experimental condition |
methdis |
the distance measure to be used. Options are 'euclidean' and 'pearson'. see 'dist' function |
methclu |
the agglomeration method to be used by the 'hclust' function |
sel |
logical, if |
size |
selects the 'size' highest variance genes for the plot if 'sel=TRUE' |
Pedro Lopez-Romero
hclust,dist
data(targets.micro) data(ddPROC) hierclusMicroRna(ddPROC$TGS,GErep, methdis="euclidean", methclu="complete", sel=FALSE,100)data(targets.micro) data(ddPROC) hierclusMicroRna(ddPROC$TGS,GErep, methdis="euclidean", methclu="complete", sel=FALSE,100)
For each array, the M value is computed for every spot as the difference between the spot intensity in the array and the averaged intensity for that feature over the whole set of arrays. It does not make a distintion between the different kind of features in the array as the mvaMicroRna() does.
mvaBASIC(object, colorfill, maintitle)mvaBASIC(object, colorfill, maintitle)
object |
An expression matrix in log2 scale |
colorfill |
color of the plot |
maintitle |
title of the plot |
Pedro Lopez-Romero
data(dd.micro) op=par(mfrow=c(1,1),ask=TRUE) mvaBASIC(log2(dd.micro$meanS), colorfill="red", maintitle=" log2 Mean Signal") par(op)data(dd.micro) op=par(mfrow=c(1,1),ask=TRUE) mvaBASIC(log2(dd.micro$meanS), colorfill="red", maintitle=" log2 Mean Signal") par(op)
For each array, the M value is computed for every spot as the difference between the spot intensity in the array and the median intensity for that feature over the whole set of arrays. Every kind of feature is identified with different color (microRNA genes, positive controls, etc ...) The input must be an uRNAList object created by the user, in such a way that the uRNAList$meanS field contains the expression matrix that we want to use in log2 scale (see example below) The gProcessedSignal computed by the Agilent Feature Extaction software normally contains negative values, so a small constant has to be added to the signals before log tranformation.
mvaMicroRna(uRNAList, maintitle, verbose=FALSE)mvaMicroRna(uRNAList, maintitle, verbose=FALSE)
uRNAList |
A uRNAList object. It uses the expression matrix stored in the uRNAList$meanS slot. Input expression matrix should be in log2 scale |
maintitle |
character to indicate the title of the graph |
verbose |
logical, if |
Pedro Lopez-Romero
data(dd.micro) op=par(mfrow=c(1,1),ask=TRUE) MMM=dd.micro$procS ## gProcessedSignal min=min(MMM) ## transforming gProcessedSignal to positive values for(i in 1:dim(MMM)[2]){ ## before log2 transformation MMM[,i]=MMM[,i]+(abs(min)+ 5) } ddaux=dd.micro ddaux$meanS=log2(MMM) mvaMicroRna(ddaux,maintitle="ProcessedSignal",verbose=FALSE) rm(ddaux) par(op)data(dd.micro) op=par(mfrow=c(1,1),ask=TRUE) MMM=dd.micro$procS ## gProcessedSignal min=min(MMM) ## transforming gProcessedSignal to positive values for(i in 1:dim(MMM)[2]){ ## before log2 transformation MMM[,i]=MMM[,i]+(abs(min)+ 5) } ddaux=dd.micro ddaux$meanS=log2(MMM) mvaMicroRna(ddaux,maintitle="ProcessedSignal",verbose=FALSE) rm(ddaux) par(op)
It is a wrapper for the 'plotPCA' of the 'affycoretools' package
PCAplotMicroRna(eset, targets)PCAplotMicroRna(eset, targets)
eset |
An Expression Set object |
targets |
data.frame with the target structure |
Pedro Lopez-Romero
data(targets.micro) data(ddPROC) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE,verbose=FALSE) PCAplotMicroRna(esetPROC,targets.micro)data(targets.micro) data(ddPROC) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE,verbose=FALSE) PCAplotMicroRna(esetPROC,targets.micro)
Creates a density plot with the arrays intensities
plotDensityMicroRna(object, maintitle)plotDensityMicroRna(object, maintitle)
object |
An expression matrix, in log2 scale |
maintitle |
title of the plot |
Pedro Lopez-Romero
data(dd.micro) plotDensity(log2(dd.micro$meanS),maintitle="log2 Mean Signal")data(dd.micro) plotDensity(log2(dd.micro$meanS),maintitle="log2 Mean Signal")
Creates an histogram of the pvalues. For multiple contrats, creates an histogram for every t.test pvalue (separate) or a single histogram for the F.test pvalue (nestedF). A uniform histogram will indicate no differential expression in the data set, whereas a right skewed histogram, will indicate some significant differential expression
pvalHistogram(fit2, DE, PVcut, DEmethod, MTestmethod, CM,verbose=FALSE)pvalHistogram(fit2, DE, PVcut, DEmethod, MTestmethod, CM,verbose=FALSE)
fit2 |
MArrayLM object |
DE |
TestResults object |
PVcut |
limit p value to declare significant features |
DEmethod |
method for decideTests, only 'separate' or 'nestedF' are implemented |
MTestmethod |
method for multiple test |
CM |
contrast matrix |
verbose |
logical, if |
Pedro Lopez-Romero
An overview of miRNA differential expression analysis is given in
basicLimma
An example of how to get the 'TestResults' object is in
getDecideTests
## Not run: pvalHistogram(fit2,DE,PVcut=0.10, DEmethod="separate",MTestmethod="BH",CM) ## End(Not run)## Not run: pvalHistogram(fit2,DE,PVcut=0.10, DEmethod="separate",MTestmethod="BH",CM) ## End(Not run)
It creates BoxPlots, Density Plots, MA plots, RLE plots and hierachical clustering plots with the sample data set.
qcPlots(dd, offset, MeanSignal=TRUE, ProcessedSignal=FALSE, TotalProbeSignal=FALSE, TotalGeneSignal=FALSE, BGMedianSignal=FALSE, BGUsed=FALSE, targets)qcPlots(dd, offset, MeanSignal=TRUE, ProcessedSignal=FALSE, TotalProbeSignal=FALSE, TotalGeneSignal=FALSE, BGMedianSignal=FALSE, BGUsed=FALSE, targets)
dd |
A uRNAList object containing the ouput from |
offset |
numeric value to add to the intensities before log transforming |
MeanSignal |
logical, if |
ProcessedSignal |
logical, if |
TotalProbeSignal |
logical, if |
TotalGeneSignal |
logical, if |
BGMedianSignal |
logical, if |
BGUsed |
logical, if |
targets |
data.frame with the target structure |
The signals loaded from the AFE data files can be used for the quality
assesment using the graphical utilities included in the qcPlots function.
For the gMeanSignal, the BoxPlots, Density Plots, MA plots, RLE plots
and hierachical clustering plots are done. For the gProcessedSignal the same
plots are done, except the hierarchical clustering. For the gTotalProbeSignal
and the gTotalGeneSignal only the BoxPlots and Density Plots are done, and
finally, for the Background signals only the Boxplots are done.
Pedro Lopez-Romero
Boldstad B.M., Collin F., Brettschneider J., Simpson, K., Cope L., Irizarry R. A., Speed T. P. Quality Assesement of Affymetrix GeneChip Data. In Bioinformatics and Computational Biology Solutions Using R and Bioconductor. (eds.) Gentleman R., Carey V. J., Huber W., Irizarry R. A., Dudoit S. (2005). Springer.
boxplotMicroRna,plotDensityMicroRna,RleMicroRna, mvaMicroRna
and hierclusMicroRna
## Not run: data(dd.micro) qcPlots(dd.micro,offset=5, MeanSignal=TRUE, ProcessedSignal=TRUE, TotalProbeSignal=TRUE, TotalGeneSignal=TRUE, BGMedianSignal=TRUE, BGUsed=TRUE, targets.micro) graphics.off() ## End(Not run)## Not run: data(dd.micro) qcPlots(dd.micro,offset=5, MeanSignal=TRUE, ProcessedSignal=TRUE, TotalProbeSignal=TRUE, TotalGeneSignal=TRUE, BGMedianSignal=TRUE, BGUsed=TRUE, targets.micro) graphics.off() ## End(Not run)
Read the data files generated by the Agilent Feature Extraction image analysis software
readMicroRnaAFE(targets,verbose=FALSE)readMicroRnaAFE(targets,verbose=FALSE)
targets |
A data frame that specifies experimental conditions under which each sample has been obtained. |
verbose |
logical, if |
The function reads the *.txt files generated by the AFE Software using the 'read.maimages' function of 'limma' package.
Data, colected with the Agilent Feature Extraction Software, are stored in a uRNAList object with the following components:
- dd.micro\$TGS 'gTotalGeneSignal' - dd.micro\$TPS 'gTotalProbeSignal' - dd.micro\$meanS 'gMeanSignal' - dd.micro\$procS 'gProcessedSignal' - dd.micro\$targets 'targets' - dd.micro\$genes\$ProbeName 'Probe Name' - dd.micro\$genes\$GeneName 'microRNA Name' - dd.micro\$genes\$ControlType 'FLAG to specify the sort of feature' - dd.micro\$other\$gIsGeneDetected 'FLAG IsGeneDetected' - dd.micro\$other\$gIsSaturated 'FLAG IsSaturated' - dd.micro\$other\$gIsFeatNonUnifOL 'FLAG IsFeatNonUnifOL' - dd.micro\$other\$gIsFeatPopnOL 'FLAG IsFeatPopnOL' - dd.micro\$other\$gBGMedianSignal 'gBGMedianSignal' - dd.micro\$other\$gBGUsed 'gBGUsed'
A uRNAList containing the following elements:
uRNAList\$TGS |
matrix, 'gTotalGeneSignal' |
uRNAList\$TPS |
matrix, 'gTotalProbeSignal' |
uRNAList\$meanS |
matrix, 'gMeanSignal' |
uRNAList\$procS |
matrix, 'gProcessedSignal' |
uRNAList\$targets |
data.frame, 'FileName' |
uRNAList\$genes\$ProbeName |
character, 'AGilent Probe Name' |
uRNAList\$genes\$GeneName |
character, 'microRNA Name' |
uRNAList\$genes\$ControlType |
integer, '0'= Feature, '1'= Positive control, '-1'= Negative control |
uRNAList\$other\$gIsGeneDetected |
matrix, FLAG to classify signal if 'IsGeneDetected=1' or 'not=0' |
uRNAList\$other\$gIsSaturated |
matrix, FLAG to classify signal if 'IsSaturated = 1' or 'not=0' |
uRNAList\$other\$gIsFeatPopnOL |
matrix, FLAG to classify signal if 'IsFeatPopnOL = 0' or 'not=1' |
uRNAList\$other\$gIsFeatNonUnifOL |
matrix, FLAG to classify signal if 'gIsFeatNonUnifOL = 0' or 'not=1' |
uRNAList\$other\$gBGMedianSignal |
matrix, gBGMedianSignal |
uRNAList\$other\$gBGUsed |
matrix, gBGUsed |
Pedro Lopez-Romero
Agilent Feature Extraction Reference Guide http://www.Agilent.com
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397–420.
A data example can be found in dd.micro
See also readTargets to see how to build the target file and
the example given in targets.micro
## Not run: data(targets.micro) dd.micro = readMicroRnaAFE(targets.micro) ## End(Not run)## Not run: data(targets.micro) dd.micro = readMicroRnaAFE(targets.micro) ## End(Not run)
The target file is a txt file created by the user where every input file (array, sample) is attached to a experimental condition
readTargets(infile,verbose=FALSE)readTargets(infile,verbose=FALSE)
infile |
name of the target file, for instance 'targets.micro.txt' |
verbose |
logical, if |
In the 'target' file (see Table 1 in vignette) we specify the experimental conditions under which the data have been generated. The target file MUST contain the following mandatory columns: -FileName : Name of the array data file -Treatment : Treatment effect -GErep : Treatment effect in numeric code, from '1' to 'n', being 'n' the number of the levels of the treatment effect
Other explanatory variables specifying the experimental conditions might be also included.
A 'data.frame' containing by the columns specified in the input file targets.txt. This 'targets.txt' file must be created by the user.
Pedro Lopez-Romero
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397–420.
An example of a target file can be found in targets.micro
RLE: Relative Log Expression
RleMicroRna(object, maintitle, colorfill)RleMicroRna(object, maintitle, colorfill)
object |
An expression matrix |
maintitle |
title of the plot |
colorfill |
color of the plot |
Each Boxplot corresponds to a sample and displays the Relative Log Expression computed for every spot in the array as the difference between the spot intensity and the median intensity for the same feature accros all the arrays. Since majority of the spots are expected not to be differentially expressed, the plot shoud show boxplots centered around zero and all of them having the aproximately the same dispersion. An array showing greater dispersion than the other, or being not centered at zero could have quality problems.
Pedro Lopez-Romero
Boldstad B.M., Collin F., Brettschneider J., Simpson, K., Cope L., Irizarry R. A., Speed T. P. Quality Assesement of Affymetrix GeneChip Data. In Bioinformatics and Computational Biology Solutions Using R and Bioconductor. (eds.) Gentleman R., Carey V. J., Huber W., Irizarry R. A., Dudoit S. (2005). Springer.
data(dd.micro) RleMicroRna(log2(dd.micro$meanS), maintitle="log2 Mean Signal RLE", colorfill="orange")data(dd.micro) RleMicroRna(log2(dd.micro$meanS), maintitle="log2 Mean Signal RLE", colorfill="orange")
The function creates an uRNAList containing the TotalGeneSignal computed by the RMA algorithm. This signal can be used for the statistical analysis.
rmaMicroRna(dd, normalize, background)rmaMicroRna(dd, normalize, background)
dd |
uRNAList, containing the output from |
normalize |
logical, if |
background |
logical, if |
The function creates an uRNAList output that contains in the uRNAList$TGS, uRNAList$TPS, uRNAList$meanS & uRNAList$procS slots the Total Gene Signal (TGS) computed by the RMA algorithm. The function uses the robust multiarray average (RMA) method from the 'affy' package. RMA obtains an estimate of the expression measure for each gene using all the replicated probes for that gene. First, RMA obtains a background corrected intensity by fitting a normal + exponential convolution model to a vector of observed intensities. The normal part represents the background and the exponential part represents the signal intensities. Then the arrays are normalized using 'quantile' normalization. Finally, for each probe set that interrogates the same microRNA, RMA fits a linear model to the background-corrected, normalized and log2 transformed probe intensities. This model produces an estimate of the gene signal taking into account the probe effect. The model parameters estimates are obtained by median polish. The estimates of the gene expression signals are referred as RMA estimates. Normally, each microRNA is interrogated by 16 probes either using 2 different probes, each of them replicated 8 times, or using 4 differnt probes replicated 4 times. First, function 'rmaMicroRna' obtains a background corrected signal using the 'rma.background.correct' function of the package 'preprocessCore' , then the signal is normalized bewtween arrays using the 'limma' function 'normalizeBetweenArrays' with the 'quantile' method. Then, the median of the replicated probes is obtained, leading to either 2 or 4 different measures for each gene. These measures correspond to different probes for the same genes that are summarized into a single RMA linear model described above.
uRNAList containing the Total Gene Signal computed by the RMA algorithm in log 2 scale.
Pedro Lopez-Romero
Irizarry, R., Hobbs,B., Collin,F., Beazer-Barclay,Y., Antonellis,K., Scherf,U., Speed,T. (2003) Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 4, 249-264
Gautier, L., Cope, L., Bolstad, B. M., and Irizarry, R. A.(2004). affy—analysis of Affymetrix GeneChip data at the probe level. Bioinformatics 20, 3, 307-315.
Bolstad B. M. (). preprocessCore: A collection of pre-processing functions. R package version 1.4.0
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397 - 420
data(dd.micro) ddTGS.rma=rmaMicroRna(dd.micro, normalize=TRUE, background=TRUE) dim(ddTGS.rma) RleMicroRna(ddTGS.rma$TGS,"RLE TGS.rma","blue")data(dd.micro) ddTGS.rma=rmaMicroRna(dd.micro, normalize=TRUE, background=TRUE) dim(ddTGS.rma) RleMicroRna(ddTGS.rma$TGS,"RLE TGS.rma","blue")
The function summarizes the results from the differential expression analysis using the different objects that are obtained after 'limma' has been used, such as the 'MArrayLM' object with the statistics and the 'TestResults' object highlighting the significant features.
significantMicroRna(eset, ddset, targets, fit2, CM, DE, DEmethod, MTestmethod, PVcut, Mcut,verbose=FALSE)significantMicroRna(eset, ddset, targets, fit2, CM, DE, DEmethod, MTestmethod, PVcut, Mcut,verbose=FALSE)
eset |
ExpressionSet containing the Total Gene processed data |
ddset |
An uRNAList object containing the Total Gene proceseed data |
targets |
data.frame with the target structure |
fit2 |
MArrayLM object from |
CM |
Contrast matrix |
DE |
TestResults object |
DEmethod |
method used in decideTests, only 'separate' or 'nestedF' are implemented |
MTestmethod |
method for multiple test |
PVcut |
p value threshold to declare significant features |
Mcut |
M value threshold to select within significant features |
verbose |
logical, if |
A list containing the genes with their statistics is generated. The significant genes above the PVcut p values are also given in a html file that links the selected miRNAS to the miRBase http://microrna.sanger.ac.uk/. A MA plots indicating the differentially expressed genes are also displayed.
When multiple contrasts are done, the method for the selection of the significant genes can be either 'separated' or 'nestedF'. See decideTests in package limma limma for a detailed description on these two methods. When 'separated' is used a list with all the genes that have been analized in limma is given. The list includes de following columns:
PROBE - Probe name (one of the probes interrogating the gene) GENE - miRNA name PROBE chr\_coord - Agilent chromosomal location M - Fold change A - Mean of the intensity for that miRNA t - moderated t-statistic pval - p value of the t-statistic adj.pval - p value adjusted by 'MTestmethod' fdr.pval - p value adjusted by fdr
Some times, the user can be set 'MTestmethod = none', in this case, it might be interesting to still see the fdr value, despite of the fact that the user has decided not apply any multiple testing correction.
If the 'nestedF' is used, then two lists are provided for each contrasts. A first containing the selected significant genes, and a second list containing the rest of the genes that have been analyzed. The columns given in this case is:
PROBE - Probe name (one of the probes interrogating the gene) GENE - miRNA name PROBE chr\_coord - Agilent chromosomal location M - Fold change A - Mean of the intensity for that miRNA t - moderated t-statistic t pval - p value of the t-statistic F - F statistic (null hypothesis: Ci = Cj, for all contrasts i, j) adj.F.pval - F p value adjusted by 'MTestmethod' fdr.F.pval - F p value adjusted by fdr
The html files, both for the 'separated' and 'nestedF' method, includes only the selected as significant genes.
Pedro Lopez-Romero
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397–420.
miRBase: the home of microRNA data http://microrna.sanger.ac.uk/
A 'uRNAList' example containing proccesed data is in
ddPROC and an overview of how the processed data is produced
is given in filterMicroRna. The ExpressionSet object can be
generated using esetMicroRna
An overview of miRNA differential expression analysis is given in
basicLimma
An example of how to get the 'TestResults' object is in
getDecideTests
data(targets.micro) data(ddPROC) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE) levels.treatment=levels(factor(targets.micro$Treatment)) treatment=factor(as.character(targets.micro$Treatment), levels=levels.treatment) levels.subject=levels(factor(targets.micro$Subject)) subject=factor(as.character(targets.micro$Subject), levels=levels.subject) design=model.matrix(~ -1 + treatment + subject ) CM=cbind(MSC_AvsMSC_B=c(1,-1,0,0), MSC_AvsMSC_C=c(1,0,-1,0)) fit2=basicLimma(esetPROC,design,CM,verbose=TRUE) DE=getDecideTests(fit2, DEmethod="separate", MTestmethod="BH", PVcut=0.10) significantMicroRna(esetPROC, ddPROC, targets.micro, fit2, CM, DE, DEmethod="separate", MTestmethod="BH", PVcut=0.10, Mcut=0, verbose=TRUE)data(targets.micro) data(ddPROC) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=FALSE) levels.treatment=levels(factor(targets.micro$Treatment)) treatment=factor(as.character(targets.micro$Treatment), levels=levels.treatment) levels.subject=levels(factor(targets.micro$Subject)) subject=factor(as.character(targets.micro$Subject), levels=levels.subject) design=model.matrix(~ -1 + treatment + subject ) CM=cbind(MSC_AvsMSC_B=c(1,-1,0,0), MSC_AvsMSC_C=c(1,0,-1,0)) fit2=basicLimma(esetPROC,design,CM,verbose=TRUE) DE=getDecideTests(fit2, DEmethod="separate", MTestmethod="BH", PVcut=0.10) significantMicroRna(esetPROC, ddPROC, targets.micro, fit2, CM, DE, DEmethod="separate", MTestmethod="BH", PVcut=0.10, Mcut=0, verbose=TRUE)
Briefly summarize microarray data objects.
## S3 method for class 'uRNAList' summary(object, ...)## S3 method for class 'uRNAList' summary(object, ...)
object |
an object of class |
... |
other arguments are not used |
The data objects are summarized as if they were lists, i.e., brief information about the length and type of the components is given. This function and this file, has been borrowed from the files created by Gordon Smyth for the limma package.
A table.
Pedro Lopez-Romero
Example of target file
data(targets.micro)data(targets.micro)
A data frame with 4 observations on the following 5 variables.
FileNamenames of the Files Ast.txt Bst.txt Aunst.txt Bunst.txt
TreatmentAssigns level for Treatment Effect to each File (mandatory)
GErepa numeric vector tha numerates the FACTOR of the Treatment Effect (mandatory)
SubjectAssigns level for Subject Effect to each File
It is a tab-delimited text format file. The target file is created by the user with the intention of carrying out a differential expression analysis in future steps using 'limma'. Here is where the factors that are going to be included in the linear model that is fitted to each gen are specified. The targets file assigns each data file to a particular experimental conditions. First column 'FileName' is mandatory and includes the image data files names. Second column 'Treatment' is also mandatory and includes the treament effect. Third column 'GErep' is also mandatory, and includes the treatment effect in a numeric code, from 1 to n, being n the number of Treatment effect levels.
Pedro Lopez-Romero
Gordon K. Smyth, M. Ritchie, N. Thorne, J. Wettenhall (2007). limma: Linear Models for Microarray Data User's Guide.
readTargets
The function creates an uRNAList containing the TotalGeneSignal computed by the Agilent Feature Extraction software. This signal can be used for the statistical analysis after a possible normalization step.
tgsMicroRna(dd, offset, half, makePLOT=FALSE, verbose=FALSE)tgsMicroRna(dd, offset, half, makePLOT=FALSE, verbose=FALSE)
dd |
uRNAList, containing the output from |
offset |
integer. To use this option set half = FALSE |
half |
logical, if |
makePLOT |
logical, if |
verbose |
logical, if |
The function creates a uRNAList oobject that contains in the uRNAList$TGS, uRNAList$TPS, uRNAList$meanS & uRNAList$procS the Total Gene Signal (TGS) as computed by the Agilent Feature Extraction algorithms. This TGS is not in log2 scale. All the replicated genes have the same estimated TGS, and the function simply picks one gene from each set of replicated genes. To mantain the format of the uRNAList, every selected gene retains a probe name attach to them. This probe name is not meaningful any more, since the signal corresponds to the total gene signal and not to the probe signal. The TGS processed by AFE contains some negative values. To get signals with positive values we can either add a positive small constant to all the signals (offset) or we can select the 'half' option, which set to 0.5 all the values that are smaller than 0.5. To use the offset option we have to set half=FALSE, otherwise the half method is used by default. The offset option, adds to each signal the quantity (abs( min(ddTGS$TGS)) + offset), where ddTGS$TGS is the matrix that contains the TotalGeneSignal.
uRNAList containing the TotalGeneSignal computed by the Agilent Feature Extraction software Optionally, it can generate a boxplot, a density plot and a MA plot with the Total Gene Signal.
Pedro Lopez-Romero
Agilent Feature Extraction Reference Guide http://www.Agilent.com
data(dd.micro) data(targets.micro) ddTGS=tgsMicroRna(dd.micro,half=TRUE,makePLOT=FALSE,verbose=FALSE)data(dd.micro) data(targets.micro) ddTGS=tgsMicroRna(dd.micro,half=TRUE,makePLOT=FALSE,verbose=FALSE)
Normalization between arrays of the Total Gene Signal. The function is a wrapper of the 'limma' 'normalizeBetweenArrays' with ('none','quantile','scale') methods
tgsNormalization(ddTGS, NORMmethod = "quantile", makePLOTpre = FALSE, makePLOTpost = FALSE, targets,verbose=FALSE)tgsNormalization(ddTGS, NORMmethod = "quantile", makePLOTpre = FALSE, makePLOTpost = FALSE, targets,verbose=FALSE)
ddTGS |
uRNAList, containing the output from |
NORMmethod |
character specifying the normalization method, 'none','quantile','scale'. The
default is |
makePLOTpre |
logical, if |
makePLOTpost |
logical, if |
targets |
data.frame with the target structure |
verbose |
logical, if |
A uRNAList object containing the Normalized Total Gene Signal in log 2 scale
Pedro Lopez-Romero
Smyth, G. K. (2005). Limma: linear models for microarray data. In: 'Bioinformatics and Computational Biology Solutions Using R and Bioconductor'. R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds), Springer, New York, pages 397 - 420
Smyth, G. K., and Speed, T. P. (2003). Normalization of cDNA microarray data. Methods 31, 265-273.
## Not run: data(dd.micro) data(targets.micro) ddTGS=tgsMicroRna(dd.micro,half=TRUE,makePLOT=FALSE,verbose=FALSE) ddNORM=tgsNormalization(ddTGS,'quantile', makePLOTpre=FALSE,makePLOTpost=TRUE,targets.micro,verbose=TRUE) graphics.off() ## End(Not run)## Not run: data(dd.micro) data(targets.micro) ddTGS=tgsMicroRna(dd.micro,half=TRUE,makePLOT=FALSE,verbose=FALSE) ddNORM=tgsNormalization(ddTGS,'quantile', makePLOTpre=FALSE,makePLOTpost=TRUE,targets.micro,verbose=TRUE) graphics.off() ## End(Not run)
A list-based class (similar to the RGList class in limma package) for the storing of
Agilent chips microRNA data
uRNAList objects are created by read.agiMicroRna
uRNAList objects are created by new("uRNAList",Newagi) where Newagi is a list.
with the following components:
matrix, 'gTotalGeneSignal'
matrix, 'gTotalProbeSignal'
matrix, 'gMeanSignal'
matrix, 'gProcessedSignal'
data.frame, 'FileName'
vector of characters, 'AGilent Probe Name'
vector of characters, 'microRNA Name'
vector of integers, '0'= Feature, '1'= Positive control, '-1'= Negative control
matrix, FLAG to classify signal if 'IsGeneDetected=1' or 'not=0'
matrix, FLAG to classify signal if 'IsSaturated = 1' or 'not=0'
matrix, FLAG to classify signal if 'IsFeatPopnOL = 0' or 'not=1'
matrix, FLAG to classify signal if 'gIsFeatNonUnifOL = 0' or 'not=1'
matrix, gBGMedianSignal
matrix, gBGUsed
Pedro Lopez-Romero
## Not run: data(dd.micro) ## End(Not run)## Not run: data(dd.micro) ## End(Not run)
Writes the expression data matrix of an ExpressionSet object in a file.
writeEset(eset, ddPROC, targets,verbose=FALSE)writeEset(eset, ddPROC, targets,verbose=FALSE)
eset |
An Expression object, normally containing the procesed data |
ddPROC |
An RGList object, normally containing the processed data |
targets |
data.frame with the targets structure |
verbose |
logical, if |
Writes the expression data matrix of an ExpressionSet object in a file.
Pedro Lopez-Romero
An 'RGList' example containing proccesed data is in
ddPROC and an overview of how the processed data is produced
is given in filterMicroRna. The ExpressionSet object can be
generated using esetMicroRna
## Not run: data(ddPROC) data(targets.micro) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=TRUE,verbose=FALSE) writeEset(esetPROC,ddPROC,targets.micro,verbose=TRUE) ## End(Not run)## Not run: data(ddPROC) data(targets.micro) esetPROC=esetMicroRna(ddPROC,targets.micro,makePLOT=TRUE,verbose=FALSE) writeEset(esetPROC,ddPROC,targets.micro,verbose=TRUE) ## End(Not run)