| Title: | Propagating Uncertainty in Microarray Analysis(including Affymetrix tranditional 3' arrays and exon arrays and Human Transcriptome Array 2.0) |
|---|---|
| Description: | Most analyses of Affymetrix GeneChip data (including tranditional 3' arrays and exon arrays and Human Transcriptome Array 2.0) are based on point estimates of expression levels and ignore the uncertainty of such estimates. By propagating uncertainty to downstream analyses we can improve results from microarray analyses. For the first time, the puma package makes a suite of uncertainty propagation methods available to a general audience. In additon to calculte gene expression from Affymetrix 3' arrays, puma also provides methods to process exon arrays and produces gene and isoform expression for alternative splicing study. puma also offers improvements in terms of scope and speed of execution over previously available uncertainty propagation methods. Included are summarisation, differential expression detection, clustering and PCA methods, together with useful plotting functions. |
| Authors: | Richard D. Pearson, Xuejun Liu, Magnus Rattray, Marta Milo, Neil D. Lawrence, Guido Sanguinetti, Li Zhang |
| Maintainer: | Xuejun Liu <[email protected]> |
| License: | LGPL |
| Version: | 3.47.0 |
| Built: | 2024-07-05 02:32:53 UTC |
| Source: | https://github.com/bioc/puma |
Most analyses of Affymetrix GeneChip data (including tranditional 3' arrays and exon arrays) are based on point estimates of expression levels and ignore the uncertainty of such estimates. By propagating uncertainty to downstream analyses we can improve results from microarray analyses. For the first time, the puma package makes a suite of uncertainty propagation methods available to a general audience. In additon to calculte gene expression from Affymetrix 3' arrays, puma also provides methods to process exon arrays and produces gene and isoform expression for alternative splicing study. puma also offers improvements in terms of scope and speed of execution over previously available uncertainty propagation methods. Included are summarisation, differential expression detection, clustering and PCA methods, together with useful plotting functions.
| Package: | puma |
| Type: | Package |
| Version: | 3.4.3 |
| Date: | 2013-11-04 |
| License: | LGPL excluding donlp2 |
For details of using the package please refer to the Vignette
Richard Pearson, Xuejun Liu, Guido Sanguinetti, Marta Milo, Neil D. Lawrence, Magnus Rattray, Li Zhang
Maintainer: Richard Pearson <[email protected]>, Li Zhang <[email protected]>
Milo, M., Niranjan, M., Holley, M. C., Rattray, M. and Lawrence, N. D. (2004) A probabilistic approach for summarising oligonucleotide gene expression data, technical report available upon request.
Liu, X., Milo, M., Lawrence, N. D. and Rattray, M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics, 21(18):3637-3644.
Sanguinetti, G., Milo, M., Rattray, M. and Lawrence, N. D. (2005) Accounting for probe-level noise in principal component analysis of microarray data, Bioinformatics, 21(19):3748-3754.
Rattray, M., Liu, X., Sanguinetti, G., Milo, M. and Lawrence, N. D. (2006) Propagating uncertainty in Microarray data analysis, Briefings in Bioinformatics, 7(1):37-47.
Liu, X., Milo, M., Lawrence, N. D. and Rattray, M. (2006) Probe-level measurement error improves accuracy in detecting differential gene expression, Bioinformatics, 22(17):2107-2113.
Liu, X. Lin, K., Andersen, B. Rattray, M. (2007) Including probe-level uncertainty in model-based gene expression clustering, BMC Bioinformatics, 8(98).
Pearson, R. D., Liu, X., Sanguinetti, G., Milo, M., Lawrence, N. D., Rattray, M. (2008) puma: a Bioconductor package for Propagating Uncertainty in Microarray Analysis, BMC Bioinformatics, 2009, 10:211.
Zhang,L. and Liu,X. (2009) An improved probabilistic model for finding differential gene expression, the 2nd BMEI 17-19 oct. 2009. Tianjin. China.
Liu,X. and Rattray,M. (2009) Including probe-level measurement error in robust mixture clustering of replicated microarray gene expression, Statistical Application in Genetics and Molecular Biology, 9(1), Article 42.
puma 3.0: improved uncertainty propagation methods for gene and transcript expression analysis, Liu et al. BMC Bioinformatics, 2013, 14:39.
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) pumapca_mmgmos <- pumaPCA(eset_mmgmos) plot(pumapca_mmgmos) eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) eset_combImproved <- pumaCombImproved(eset_mmgmos_100) esetDE <- pumaDE(eset_comb) esetDEImproved <- pumaDE(eset_combImproved)# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) pumapca_mmgmos <- pumaPCA(eset_mmgmos) plot(pumapca_mmgmos) eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) eset_combImproved <- pumaCombImproved(eset_mmgmos_100) esetDE <- pumaDE(eset_comb) esetDEImproved <- pumaDE(eset_combImproved)
This function calculates the combined signal for each condition from replicates using Bayesian models. The inputs are gene expression levels and the probe-level standard deviation associated with expression measurement for each gene on each chip. The outputs include gene expression levels and standard deviation for each condition.
This function was originally part of the pplr package. Although this function can be called directly, it is recommended to use the pumaComb function instead, which can work directly on ExpressionSet objects, and can automatically determine which arrays are replicates.
bcomb(e, se, replicates, method=c("map","em"), gsnorm=FALSE, nsample=1000, eps=1.0e-6)bcomb(e, se, replicates, method=c("map","em"), gsnorm=FALSE, nsample=1000, eps=1.0e-6)
e |
a data frame containing the expression level for each gene on each chip. |
se |
a data frame containing the standard deviation of gene expression levels. |
replicates |
a vector indicating which chip belongs to which condition. |
method |
character specifying the method algorithm used. |
gsnorm |
logical specifying whether do global scaling normalisation or not. |
nsample |
integer. The number of sampling in parameter estimation. |
eps |
a numeric, optimisation parameter. |
Each element in replicate represents the condition of the chip which is in the same column order as in the expression and standard deviation matrix files.
Method "map" uses MAP of a hierarchical Bayesion model with Gamma prior on the between-replicate variance (Gelman et.al. p.285) and shares the same variance across conditions. This method is fast and suitable for the case where there are many conditions.
Method "em" uses variational inference of the same hierarchical Bayesion model as in method "map" but with conjugate prior on between-replicate variance and shares the variance across conditions.
The parameter nsample should be large enough to ensure stable parameter estimates. Should be at least 1000.
The result is a data frame with components named 'M1', 'M2', and so on, which represent the mean expression values for condition 1, condition 2, and so on. It also has components named 'Std1', 'Std2', and so on, which represent the standard deviation of the gene expression values for condition 1, condtion 2, and so on.
Xuejun Liu, Marta Milo, Neil D. Lawrence, Magnus Rattray
Gelman,A., Carlin,J.B., Stern,H.S., Rubin,D.B., Bayesian data analysis. London: Chapman & Hall; 1995.
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2006) Probe-level variances improve accuracy in detecting differential gene expression, Bioinformatics, 22:2107-2113.
Related methods pumaComb, mmgmos and pplr
# data(exampleE) # data(exampleStd) # r<-bcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2),method="map")# data(exampleE) # data(exampleStd) # r<-bcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2),method="map")
Calculates the AUC values for one or more ROC plots.
calcAUC(scores, truthValues, includedProbesets = 1:length(truthValues))calcAUC(scores, truthValues, includedProbesets = 1:length(truthValues))
scores |
A vector of scores. This could be, e.g. one of the columns of the statistics of a |
truthValues |
A boolean vector indicating which scores are True Positives. |
includedProbesets |
A vector of indices indicating which scores (and truthValues) are to be used in the calculation. The default is to use all, but a subset can be used if, for example, you only want a subset of the probesets which are not True Positives to be treated as False Positives. |
A single number which is the AUC value.
Richard D. Pearson
Related methods plotROC and numFP.
#class1a <- rnorm(1000,0.2,0.1) #class2a <- rnorm(1000,0.6,0.2) #class1b <- rnorm(1000,0.3,0.1) #class2b <- rnorm(1000,0.5,0.2) #scores_a <- c(class1a, class2a) #scores_b <- c(class1b, class2b) #classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) #print(calcAUC(scores_a, classElts)) #print(calcAUC(scores_b, classElts))#class1a <- rnorm(1000,0.2,0.1) #class2a <- rnorm(1000,0.6,0.2) #class1b <- rnorm(1000,0.3,0.1) #class2b <- rnorm(1000,0.5,0.2) #scores_a <- c(class1a, class2a) #scores_b <- c(class1b, class2b) #classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) #print(calcAUC(scores_a, classElts)) #print(calcAUC(scores_b, classElts))
Automatically creates design and contrast matrices if not specified. This function is useful for comparing fold change results with those of other differential expression (DE) methods such as pumaDE.
calculateFC( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )calculateFC( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )
eset |
An object of class |
design.matrix |
A design matrix |
contrast.matrix |
A contrast matrix |
The eset argument must be supplied, and must be a valid ExpressionSet object. Design and contrast matrices can be supplied, but if not, default matrices will be used. These should usually be sufficient for most analyses.
An object of class DEResult.
Richard D. Pearson
Related methods pumaDE, calculateLimma, calculateTtest, createDesignMatrix and createContrastMatrix and class DEResult
#if (require(affydata)) { # data(Dilution) # eset_rma <- affy:::rma(Dilution) # # Next line used so eset_rma only has information about the liver factor # # The scanner factor will thus be ignored, and the two arrays of each level # # of the liver factor will be treated as replicates # pData(eset_rma) <- pData(eset_rma)[,1, drop=FALSE] # FCRes <- calculateFC(eset_rma) # topGeneIDs(FCRes,numberOfGenes=6) # plotErrorBars(eset_rma, topGenes(FCRes)) #}#if (require(affydata)) { # data(Dilution) # eset_rma <- affy:::rma(Dilution) # # Next line used so eset_rma only has information about the liver factor # # The scanner factor will thus be ignored, and the two arrays of each level # # of the liver factor will be treated as replicates # pData(eset_rma) <- pData(eset_rma)[,1, drop=FALSE] # FCRes <- calculateFC(eset_rma) # topGeneIDs(FCRes,numberOfGenes=6) # plotErrorBars(eset_rma, topGenes(FCRes)) #}
Runs a default analysis using the limma package. Automatically creates design and contrast matrices if not specified. This function is useful for comparing limma results with those of other differential expression (DE) methods such as pumaDE.
calculateLimma( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )calculateLimma( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )
eset |
An object of class |
design.matrix |
A design matrix |
contrast.matrix |
A contrast matrix |
The eset argument must be supplied, and must be a valid ExpressionSet object. Design and contrast matrices can be supplied, but if not, default matrices will be used. These should usually be sufficient for most analyses.
An object of class DEResult.
Richard D. Pearson
Related methods pumaDE, calculateTtest, calculateFC, createDesignMatrix and createContrastMatrix and class DEResult
#if (require(affydata)) { # data(Dilution) # eset_rma <- affy:::rma(Dilution) # # Next line used so eset_rma only has information about the liver factor # # The scanner factor will thus be ignored, and the two arrays of each level # # of the liver factor will be treated as replicates # pData(eset_rma) <- pData(eset_rma)[,1, drop=FALSE] # limmaRes <- calculateLimma(eset_rma) # topGeneIDs(limmaRes,numberOfGenes=6) # plotErrorBars(eset_rma, topGenes(limmaRes)) #}#if (require(affydata)) { # data(Dilution) # eset_rma <- affy:::rma(Dilution) # # Next line used so eset_rma only has information about the liver factor # # The scanner factor will thus be ignored, and the two arrays of each level # # of the liver factor will be treated as replicates # pData(eset_rma) <- pData(eset_rma)[,1, drop=FALSE] # limmaRes <- calculateLimma(eset_rma) # topGeneIDs(limmaRes,numberOfGenes=6) # plotErrorBars(eset_rma, topGenes(limmaRes)) #}
Automatically creates design and contrast matrices if not specified. This function is useful for comparing T-test results with those of other differential expression (DE) methods such as pumaDE.
calculateTtest( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )calculateTtest( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )
eset |
An object of class |
design.matrix |
A design matrix |
contrast.matrix |
A contrast matrix |
The eset argument must be supplied, and must be a valid ExpressionSet object. Design and contrast matrices can be supplied, but if not, default matrices will be used. These should usually be sufficient for most analyses.
An object of class DEResult.
Richard D. Pearson
Related methods pumaDE, calculateLimma, calculateFC, createDesignMatrix and createContrastMatrix and class DEResult
# eset_test <- new("ExpressionSet", exprs=matrix(rnorm(400,8,2),100,4)) # pData(eset_test) <- data.frame("class"=c("A", "A", "B", "B")) # TtestRes <- calculateTtest(eset_test) # plotErrorBars(eset_test, topGenes(TtestRes))# eset_test <- new("ExpressionSet", exprs=matrix(rnorm(400,8,2),100,4)) # pData(eset_test) <- data.frame("class"=c("A", "A", "B", "B")) # TtestRes <- calculateTtest(eset_test) # plotErrorBars(eset_test, topGenes(TtestRes))
This data is an artificial example of the mean gene expression levels.
data(Clust.exampleE)data(Clust.exampleE)
A 700x20 matrix including 700 genes and 20 chips. Every 100 genes belong to one cluster from the first gene. There are 7 clusters.
Liu, X. Lin, K., Andersen, B. Rattray, M. (2007) Including probe-level uncertainty in model-based gene expression clustering, BMC Bioinformatics, 8(98).
This data is an artificial example of the standard deviation for gene expression levels.
data(Clust.exampleStd)data(Clust.exampleStd)
A 700x20 matrix including 700 genes and 20 chips. Every 100 genes belong to one cluster from the first gene. There are 7 true clusters.
Liu, X. Lin, K., Andersen, B. Rattray, M. (2007) Including probe-level uncertainty in model-based gene expression clustering, BMC Bioinformatics, 8(98).
This is basically the clusterApplyLB
function from the snow package, but with dots displayed to
indicate progress.
clusterApplyLBDots(cl, x, fun, ...)clusterApplyLBDots(cl, x, fun, ...)
cl |
cluster object |
x |
array |
fun |
function or character string naming a function |
... |
additional arguments to pass to standard function |
Richard D. Pearson (modified from original snow function)
This function normalise the data vector to have zero mean.
clusterNormE(x)clusterNormE(x)
x |
a vector which contains gene expression level on log2 scale. |
Vector x is related to a gene and each element is related to a chip.
The return vector is in the same format as the input x.
Xuejun Liu, Magnus Rattray
See Also as pumaClust and pumaClustii
#data(Clust.exampleE) #Clust.exampleE.centered<-t(apply(Clust.exampleE, 1, clusterNormE))#data(Clust.exampleE) #Clust.exampleE.centered<-t(apply(Clust.exampleE, 1, clusterNormE))
This function adjusts the variance of the gene expression according to the zero-centered normalisation.
clusterNormVar(x)clusterNormVar(x)
x |
a vector which contains the variance of gene expression level on log2 scale. |
Vector x is related to a gene and each element is related to a chip.
The return vector is in the same format as the input x.
Xuejun Liu, Magnus Rattray
See Also as pumaClust and pumaClustii
#data(Clust.exampleE) #data(Clust.exampleStd) #Clust.exampleVar<-Clust.exampleStd^2 #Clust.exampleStd.centered<-t(apply(cbind(Clust.exampleE,Clust.exampleVar), 1, clusterNormVar))#data(Clust.exampleE) #data(Clust.exampleStd) #Clust.exampleVar<-Clust.exampleStd^2 #Clust.exampleStd.centered<-t(apply(cbind(Clust.exampleE,Clust.exampleVar), 1, clusterNormVar))
This data is an artificial example of the mean gene exapression levels generated
by package mmgmos.
data(Clustii.exampleE)data(Clustii.exampleE)
A 600x80 matrix including 600 genes and 20 conditions. Each condition has 4 replicates. Every 100 genes belong to one cluster from the first gene. There are 6 clusters.
Liu,X. and Rattray,M. (2009) Including probe-level measurement error in robust mixture clustering of replicated microarray gene expression, Statistical Application in Genetics and Molecular Biology, 9(1), Article 42.
This data is an artificial example of the standard deviation for gene exapression levels generated
by package mmgmos.
data(Clustii.exampleStd)data(Clustii.exampleStd)
A 600x80 matrix including 600 genes and 20 conditions. Each condition has 4 replicates. Every 100 genes belong to one cluster from the first gene. There are 6 clusters.
Liu,X. and Rattray,M. (2009) Including probe-level measurement error in robust mixture clustering of replicated microarray gene expression, technical report available upon request.
This function compares the identification of differentially expressed (DE) genes using the pumaDE function and the limma package.
compareLimmapumaDE( eset_mmgmos , eset_comb = NULL , eset_other = eset_mmgmos , limmaRes = calculateLimma(eset_other) , pumaDERes = pumaDE(eset_comb) , contrastMatrix = createContrastMatrix(eset_mmgmos) , numberToCompareForContrasts = 3 , numberToCompareForVenn = 100 , plotContrasts = TRUE , contrastsFilename = NULL , plotOther = FALSE , otherFilename = "other" , plotBcombContrasts = FALSE , bcombContrastsFilename = "bcomb_contrasts" , plotVenn = FALSE , vennFilename = "venn.pdf" , showTopMatches = FALSE , returnResults = FALSE )compareLimmapumaDE( eset_mmgmos , eset_comb = NULL , eset_other = eset_mmgmos , limmaRes = calculateLimma(eset_other) , pumaDERes = pumaDE(eset_comb) , contrastMatrix = createContrastMatrix(eset_mmgmos) , numberToCompareForContrasts = 3 , numberToCompareForVenn = 100 , plotContrasts = TRUE , contrastsFilename = NULL , plotOther = FALSE , otherFilename = "other" , plotBcombContrasts = FALSE , bcombContrastsFilename = "bcomb_contrasts" , plotVenn = FALSE , vennFilename = "venn.pdf" , showTopMatches = FALSE , returnResults = FALSE )
eset_mmgmos |
An object of class |
eset_comb |
An object of class |
eset_other |
An object of class |
limmaRes |
A list with two elements, usually created using the function |
pumaDERes |
A list with two elements, usually created using the function |
contrastMatrix |
A contrast matrix. If not supplied this will be created from |
numberToCompareForContrasts |
An integer specifying the number of most differentially expressed probe sets (genes) that will be used in comparison charts. |
numberToCompareForVenn |
An integer specifying the number of most differentially expressed probe sets (genes) that will be used for comparison in the Venn diagram. |
plotContrasts |
A boolean specifying whether or not to plot the most differentially expressed probe sets (genes) for each contrast for the |
contrastsFilename |
A character string specifying a file name stem for the PDF files which will be created to hold the contrast plots for the |
plotOther |
A boolean specifying whether or not to plot the most differentially expressed probe sets (genes) for each contrast for the |
otherFilename |
A character string specifying a file name stem for the PDF files which will be created to hold the contrast plots for the |
plotBcombContrasts |
A boolean specifying whether or not to plot the most differentially expressed probe sets (genes) for each contrast for the |
bcombContrastsFilename |
A character string specifying a file name stem for the PDF files which will be created to hold the contrast plots for the |
plotVenn |
A boolean specifying whether or not to plot a Venn diagram showing the overlap in the most differentially expressed probe sets (genes) as identified from the two different methods being compared. |
vennFilename |
A character string specifying the filename for the PDF file which will hold the Venn diagram showing the overlap in the most differentially expressed probe sets (genes) as identified from the two different methods being compared. |
showTopMatches |
A boolean specifying whether or not to show the probe sets which are deemed most likely to be differentially expressed. |
returnResults |
A boolean specifying whether or not to return a list containing results generated. |
The main outputs from this function are a number of PDF files.
The function only returns results if returnResults=TRUE
Richard D. Pearson
Related methods pumaDE and calculateLimma
This is really an internal function called from pumaComb. It is used to create an ExpressionSet object from the output of the bcomb function (which was originally part of the pplr package. Don't worry about it!
create_eset_r( eset , r , design.matrix=createDesignMatrix(eset) )create_eset_r( eset , r , design.matrix=createDesignMatrix(eset) )
eset |
An object of class |
r |
A data frame with components named 'M1', 'M2', and so on, which represent the mean expression values for condition 1, condition 2, and so on. It also has components named 'Std1', 'Std2', and so on, which represent the standard deviation of the gene expression values for condition 1, condtion 2, and so on. This type of data frame is output by function |
design.matrix |
A design matrix. |
An object of class ExpressionSet.
Richard D. Pearson
Related methods bcomb, hcomb, pumaComb and pumaCombImproved
To appear
createContrastMatrix(eset, design=NULL)createContrastMatrix(eset, design=NULL)
eset |
An object of class |
design |
A design matrix |
The puma package has been designed to be as easy to use as possible, while not compromising on power and flexibility. One of the most difficult tasks for many users, particularly those new to microarray analysis, or statistical analysis in general, is setting up design and contrast matrices. The puma package will automatically create such matrices, and we believe the way this is done will suffice for most users' needs.
It is important to recognise that the automatic creation of design and contrast matrices will only happen if appropriate information about the levels of each factor is available for each array in the experimental design. This data should be held in an AnnotatedDataFrame class. The easiest way of doing this is to ensure that the AnnotatedDataFrame object holding the raw CEL file data has an appropriate phenoData slot. This information will then be passed through to any ExpressionSet object created, for example through the use of mmgmos. The phenoData slot of an ExpressionSet object can also be manipulated directly if necessary.
Design and contrast matrices are dependent on the experimental design. The simplest experimental designs have just one factor, and hence the phenoData slot will have a matrix with just one column. In this case, each unique value in that column will be treated as a distinct level of the factor, and hence pumaComb will group arrays according to these levels. If there are just two levels of the factor, e.g. A and B, the contrast matrix will also be very simple, with the only contrast of interest being A vs B. For factors with more than two levels, a contrast matrix will be created which reflects all possible combinations of levels. For example, if we have three levels A, B and C, the contrasts of interest will be A vs B, A vs C and B vs C. In addition, if the others argument is set to TRUE, the following additional contrasts will be created: A vs other (i.e. A vs B \& C), B vs other and C vs other. Note that these additional contrasts are experimental, and not currently recommended for use in calculating differential expression.
If we now consider the case of two or more factors, things become more complicated. There are now two cases to be considered: factorial experiments, and non-factorial experiments. A factorial experiment is one where all the combinations of the levels of each factor are tested by at least one array (though ideally we would have a number of biological replicates for each combination of factor levels). The estrogen case study from the package vignette is an example of a factorial experiment.
A non-factorial experiment is one where at least one combination of levels is not tested. If we treat the example used in the puma-package help page as a two-factor experiment (with factors “level” and “batch”), we can see that this is not a factorial experiment as we have no array to test the conditions “level=ten” and “batch=B”. We will treat the factorial and non-factorial cases separately in the following sections.
Factorial experiments
For factorial experiments, the design matrix will use all columns from the phenoData slot. This will mean that pumaComb will group arrays according to a combination of the levels of all the factors.
Non-factorial designs
For non-factorial designed experiments, we will simply ignore columns (right to left) from the phenoData slot until we have a factorial design or a single factor. We can see this in the example used in the puma-package help page. Here we have ignored the “batch” factor, and modelled the experiment as a single-factor experiment (with that single factor being “level”).
The result is a matrix. See the code below for an example.
Richard D. Pearson
Related methods createDesignMatrix and pumaDE
if(FALSE){ # This is a simple example based on a real data set. Note that this is an "unbalanced" design, the "level" factor has two replicates of the "twenty" condition, but only one replicate of the "ten" condition. Also note that the second factor, "batch" is not used in the design or contrast matrices, as we don't have every combination of the levels of "level" and "batch" (there is no array for level=twenty and batch=B). # Next 4 lines commented out to save time in package checks, and saved version used if (require(affydata)) { data(Dilution) eset_mmgmos <- mmgmos(Dilution) } data(eset_mmgmos) createContrastMatrix(eset_mmgmos) # The following shows a set of 15 synthetic data sets with increasing complexity. We first create the data sets, then look at the contrast matrices. # single 2-level factor eset1 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset1) <- data.frame("class"=c(1,1,2,2)) # single 2-level factor - unbalanced design eset2 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset2) <- data.frame("class"=c(1,2,2,2)) # single 3-level factor eset3 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset3) <- data.frame("class"=c(1,1,2,2,3,3)) # single 4-level factor eset4 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset4) <- data.frame("class"=c(1,1,2,2,3,3,4,4)) # 2x2 factorial eset5 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset5) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b"), "fac2"=c(1,1,2,2,1,1,2,2)) # 2x2 factorial - unbalanced design eset6 <- new("ExpressionSet", exprs=matrix(0,100,10)) pData(eset6) <- data.frame("fac1"=c("a","a","a","b","b","b","b","b","b","b"), "fac2"=c(1,2,2,1,1,1,2,2,2,2)) # 3x2 factorial eset7 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset7) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c"), "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2)) # 2x3 factorial eset8 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset8) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) ) # 2x2x2 factorial eset9 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset9) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset10 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset10) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c") , "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset11 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset11) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset12 <- new("ExpressionSet", exprs=matrix(0,100,18)) pData(eset12) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b","c","c","c","c","c","c") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 2x2x2x2 factorial eset13 <- new("ExpressionSet", exprs=matrix(0,100,16)) pData(eset13) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a","b","b","b","b","b","b","b","b") , "fac2"=c(0,0,0,0,1,1,1,1,0,0,0,0,1,1,1,1) , "fac3"=c(2,2,3,3,2,2,3,3,2,2,3,3,2,2,3,3) , "fac4"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # "Un-analysable" data set - all arrays are from the same class eset14 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset14) <- data.frame("class"=c(1,1,1,1)) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. eset15 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset15) <- data.frame("fac1"=c("a","a","a","a","b","b"), "fac2"=c(1,1,2,2,1,1)) createContrastMatrix(eset1) createContrastMatrix(eset2) createContrastMatrix(eset3) createContrastMatrix(eset4) createContrastMatrix(eset5) createContrastMatrix(eset6) createContrastMatrix(eset7) createContrastMatrix(eset8) createContrastMatrix(eset9) # For the last 4 data sets, the contrast matrices get pretty big, so we'll just show the names of each contrast colnames(createContrastMatrix(eset10)) colnames(createContrastMatrix(eset11)) # Note that the number of contrasts can rapidly get very large for multi-factorial experiments! colnames(createContrastMatrix(eset12)) # For this final data set, note that the puma package does not currently create interaction terms for data sets with 4 of more factors. E-mail the author if you would like to do this. colnames(createContrastMatrix(eset13)) # "Un-analysable" data set - all arrays are from the same class - gives an error. Note that we've commented this out so that we don't get errors which would make the package fail the Bioconductor checks! createContrastMatrix(eset14) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. createContrastMatrix(eset15) }if(FALSE){ # This is a simple example based on a real data set. Note that this is an "unbalanced" design, the "level" factor has two replicates of the "twenty" condition, but only one replicate of the "ten" condition. Also note that the second factor, "batch" is not used in the design or contrast matrices, as we don't have every combination of the levels of "level" and "batch" (there is no array for level=twenty and batch=B). # Next 4 lines commented out to save time in package checks, and saved version used if (require(affydata)) { data(Dilution) eset_mmgmos <- mmgmos(Dilution) } data(eset_mmgmos) createContrastMatrix(eset_mmgmos) # The following shows a set of 15 synthetic data sets with increasing complexity. We first create the data sets, then look at the contrast matrices. # single 2-level factor eset1 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset1) <- data.frame("class"=c(1,1,2,2)) # single 2-level factor - unbalanced design eset2 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset2) <- data.frame("class"=c(1,2,2,2)) # single 3-level factor eset3 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset3) <- data.frame("class"=c(1,1,2,2,3,3)) # single 4-level factor eset4 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset4) <- data.frame("class"=c(1,1,2,2,3,3,4,4)) # 2x2 factorial eset5 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset5) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b"), "fac2"=c(1,1,2,2,1,1,2,2)) # 2x2 factorial - unbalanced design eset6 <- new("ExpressionSet", exprs=matrix(0,100,10)) pData(eset6) <- data.frame("fac1"=c("a","a","a","b","b","b","b","b","b","b"), "fac2"=c(1,2,2,1,1,1,2,2,2,2)) # 3x2 factorial eset7 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset7) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c"), "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2)) # 2x3 factorial eset8 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset8) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) ) # 2x2x2 factorial eset9 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset9) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset10 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset10) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c") , "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset11 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset11) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset12 <- new("ExpressionSet", exprs=matrix(0,100,18)) pData(eset12) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b","c","c","c","c","c","c") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 2x2x2x2 factorial eset13 <- new("ExpressionSet", exprs=matrix(0,100,16)) pData(eset13) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a","b","b","b","b","b","b","b","b") , "fac2"=c(0,0,0,0,1,1,1,1,0,0,0,0,1,1,1,1) , "fac3"=c(2,2,3,3,2,2,3,3,2,2,3,3,2,2,3,3) , "fac4"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # "Un-analysable" data set - all arrays are from the same class eset14 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset14) <- data.frame("class"=c(1,1,1,1)) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. eset15 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset15) <- data.frame("fac1"=c("a","a","a","a","b","b"), "fac2"=c(1,1,2,2,1,1)) createContrastMatrix(eset1) createContrastMatrix(eset2) createContrastMatrix(eset3) createContrastMatrix(eset4) createContrastMatrix(eset5) createContrastMatrix(eset6) createContrastMatrix(eset7) createContrastMatrix(eset8) createContrastMatrix(eset9) # For the last 4 data sets, the contrast matrices get pretty big, so we'll just show the names of each contrast colnames(createContrastMatrix(eset10)) colnames(createContrastMatrix(eset11)) # Note that the number of contrasts can rapidly get very large for multi-factorial experiments! colnames(createContrastMatrix(eset12)) # For this final data set, note that the puma package does not currently create interaction terms for data sets with 4 of more factors. E-mail the author if you would like to do this. colnames(createContrastMatrix(eset13)) # "Un-analysable" data set - all arrays are from the same class - gives an error. Note that we've commented this out so that we don't get errors which would make the package fail the Bioconductor checks! createContrastMatrix(eset14) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. createContrastMatrix(eset15) }
Automatically create a design matrix from an ExpressionSet.
createDesignMatrix(eset)createDesignMatrix(eset)
eset |
An object of class |
The puma package has been designed to be as easy to use as possible, while not compromising on power and flexibility. One of the most difficult tasks for many users, particularly those new to microarray analysis, or statistical analysis in general, is setting up design and contrast matrices. The puma package will automatically create such matrices, and we believe the way this is done will suffice for most users' needs.
It is important to recognise that the automatic creation of design and contrast matrices will only happen if appropriate information about the levels of each factor is available for each array in the experimental design. This data should be held in an AnnotatedDataFrame class. The easiest way of doing this is to ensure that the AnnotatedDataFrame object holding the raw CEL file data has an appropriate phenoData slot. This information will then be passed through to any ExpressionSet object created, for example through the use of mmgmos. The phenoData slot of an ExpressionSet object can also be manipulated directly if necessary.
Design and contrast matrices are dependent on the experimental design. The simplest experimental designs have just one factor, and hence the phenoData slot will have a matrix with just one column. In this case, each unique value in that column will be treated as a distinct level of the factor, and hence pumaComb will group arrays according to these levels. If there are just two levels of the factor, e.g. A and B, the contrast matrix will also be very simple, with the only contrast of interest being A vs B. For factors with more than two levels, a contrast matrix will be created which reflects all possible combinations of levels. For example, if we have three levels A, B and C, the contrasts of interest will be A vs B, A vs C and B vs C.
If we now consider the case of two or more factors, things become more complicated. There are now two cases to be considered: factorial experiments, and non-factorial experiments. A factorial experiment is one where all the combinations of the levels of each factor are tested by at least one array (though ideally we would have a number of biological replicates for each combination of factor levels). The estrogen case study from the package vignette is an example of a factorial experiment.
A non-factorial experiment is one where at least one combination of levels is not tested. If we treat the example used in the puma-package help page as a two-factor experiment (with factors “level” and “batch”), we can see that this is not a factorial experiment as we have no array to test the conditions “level=ten” and “batch=B”. We will treat the factorial and non-factorial cases separately in the following sections.
Factorial experiments
For factorial experiments, the design matrix will use all columns from the phenoData slot. This will mean that pumaComb will group arrays according to a combination of the levels of all the factors.
Non-factorial designs
For non-factorial designed experiments, we will simply ignore columns (right to left) from the phenoData slot until we have a factorial design or a single factor. We can see this in the example used in the puma-package help page. Here we have ignored the “batch” factor, and modelled the experiment as a single-factor experiment (with that single factor being “level”).
The result is a matrix. See the code below for an example.
Richard D. Pearson
Related methods createContrastMatrix, pumaComb, pumaDE and pumaCombImproved
if(FALSE){ # This is a simple example based on a real data set. Note that this is an "unbalanced" design, the "level" factor has two replicates of the "twenty" condition, but only one replicate of the "ten" condition. Also note that the second factor, "batch" is not used in the design or contrast matrices, as we don't have every combination of the levels of "level" and "batch" (there is no array for level=twenty and batch=B). # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) createDesignMatrix(eset_mmgmos) # The following shows a set of 15 synthetic data sets with increasing complexity. We first create the data sets, then look at the design matrices. # single 2-level factor eset1 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset1) <- data.frame("class"=c(1,1,2,2)) # single 2-level factor - unbalanced design eset2 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset2) <- data.frame("class"=c(1,2,2,2)) # single 3-level factor eset3 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset3) <- data.frame("class"=c(1,1,2,2,3,3)) # single 4-level factor eset4 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset4) <- data.frame("class"=c(1,1,2,2,3,3,4,4)) # 2x2 factorial eset5 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset5) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b"), "fac2"=c(1,1,2,2,1,1,2,2)) # 2x2 factorial - unbalanced design eset6 <- new("ExpressionSet", exprs=matrix(0,100,10)) pData(eset6) <- data.frame("fac1"=c("a","a","a","b","b","b","b","b","b","b"), "fac2"=c(1,2,2,1,1,1,2,2,2,2)) # 3x2 factorial eset7 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset7) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c"), "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2)) # 2x3 factorial eset8 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset8) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) ) # 2x2x2 factorial eset9 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset9) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset10 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset10) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c") , "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset11 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset11) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset12 <- new("ExpressionSet", exprs=matrix(0,100,18)) pData(eset12) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b","c","c","c","c","c","c") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 2x2x2x2 factorial eset13 <- new("ExpressionSet", exprs=matrix(0,100,16)) pData(eset13) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a","b","b","b","b","b","b","b","b") , "fac2"=c(0,0,0,0,1,1,1,1,0,0,0,0,1,1,1,1) , "fac3"=c(2,2,3,3,2,2,3,3,2,2,3,3,2,2,3,3) , "fac4"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # "Un-analysable" data set - all arrays are from the same class eset14 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset14) <- data.frame("class"=c(1,1,1,1)) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. eset15 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset15) <- data.frame("fac1"=c("a","a","a","a","b","b"), "fac2"=c(1,1,2,2,1,1)) # "pseduo 2 factor" data set - second factor is informative eset16 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset16) <- data.frame("fac1"=c("a","a","b","b"), "fac2"=c(1,1,1,1)) # "pseduo 2 factor" data set - first factor is informative eset17 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset17) <- data.frame("fac1"=c("a","a","a","a"), "fac2"=c(1,1,2,2)) # "pseudo 3 factor" data set - first factor is uninformative so actually a 2x2 factorial eset18 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset18) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y") ) # "pseudo 3 factor" data set - first and third factors are uninformative so actually a single factor eset19 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset19) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","X","X","X","X","X","X","X") ) createDesignMatrix(eset1) createDesignMatrix(eset2) createDesignMatrix(eset3) createDesignMatrix(eset4) createDesignMatrix(eset5) createDesignMatrix(eset6) createDesignMatrix(eset7) createDesignMatrix(eset8) createDesignMatrix(eset9) createDesignMatrix(eset10) createDesignMatrix(eset11) createDesignMatrix(eset12) createDesignMatrix(eset13) # "Un-analysable" data set - all arrays are from the same class - gives an error. Note that we've commented this out so that we don't get errors which would make the package fail the Bioconductor checks! # createDesignMatrix(eset14) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. createDesignMatrix(eset15) }if(FALSE){ # This is a simple example based on a real data set. Note that this is an "unbalanced" design, the "level" factor has two replicates of the "twenty" condition, but only one replicate of the "ten" condition. Also note that the second factor, "batch" is not used in the design or contrast matrices, as we don't have every combination of the levels of "level" and "batch" (there is no array for level=twenty and batch=B). # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) createDesignMatrix(eset_mmgmos) # The following shows a set of 15 synthetic data sets with increasing complexity. We first create the data sets, then look at the design matrices. # single 2-level factor eset1 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset1) <- data.frame("class"=c(1,1,2,2)) # single 2-level factor - unbalanced design eset2 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset2) <- data.frame("class"=c(1,2,2,2)) # single 3-level factor eset3 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset3) <- data.frame("class"=c(1,1,2,2,3,3)) # single 4-level factor eset4 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset4) <- data.frame("class"=c(1,1,2,2,3,3,4,4)) # 2x2 factorial eset5 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset5) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b"), "fac2"=c(1,1,2,2,1,1,2,2)) # 2x2 factorial - unbalanced design eset6 <- new("ExpressionSet", exprs=matrix(0,100,10)) pData(eset6) <- data.frame("fac1"=c("a","a","a","b","b","b","b","b","b","b"), "fac2"=c(1,2,2,1,1,1,2,2,2,2)) # 3x2 factorial eset7 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset7) <- data.frame("fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c"), "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2)) # 2x3 factorial eset8 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset8) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) ) # 2x2x2 factorial eset9 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset9) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset10 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset10) <- data.frame( "fac1"=c("a","a","a","a","b","b","b","b","c","c","c","c") , "fac2"=c(1,1,2,2,1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset11 <- new("ExpressionSet", exprs=matrix(0,100,12)) pData(eset11) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 3x2x2 factorial eset12 <- new("ExpressionSet", exprs=matrix(0,100,18)) pData(eset12) <- data.frame( "fac1"=c("a","a","a","a","a","a","b","b","b","b","b","b","c","c","c","c","c","c") , "fac2"=c(1,1,2,2,3,3,1,1,2,2,3,3,1,1,2,2,3,3) , "fac3"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # 2x2x2x2 factorial eset13 <- new("ExpressionSet", exprs=matrix(0,100,16)) pData(eset13) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a","b","b","b","b","b","b","b","b") , "fac2"=c(0,0,0,0,1,1,1,1,0,0,0,0,1,1,1,1) , "fac3"=c(2,2,3,3,2,2,3,3,2,2,3,3,2,2,3,3) , "fac4"=c("X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y","X","Y") ) # "Un-analysable" data set - all arrays are from the same class eset14 <- new("ExpressionSet", exprs=matrix(0,100,4)) pData(eset14) <- data.frame("class"=c(1,1,1,1)) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. eset15 <- new("ExpressionSet", exprs=matrix(0,100,6)) pData(eset15) <- data.frame("fac1"=c("a","a","a","a","b","b"), "fac2"=c(1,1,2,2,1,1)) # "pseduo 2 factor" data set - second factor is informative eset16 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset16) <- data.frame("fac1"=c("a","a","b","b"), "fac2"=c(1,1,1,1)) # "pseduo 2 factor" data set - first factor is informative eset17 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset17) <- data.frame("fac1"=c("a","a","a","a"), "fac2"=c(1,1,2,2)) # "pseudo 3 factor" data set - first factor is uninformative so actually a 2x2 factorial eset18 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset18) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","Y","X","Y","X","Y","X","Y") ) # "pseudo 3 factor" data set - first and third factors are uninformative so actually a single factor eset19 <- new("ExpressionSet", exprs=matrix(0,100,8)) pData(eset19) <- data.frame( "fac1"=c("a","a","a","a","a","a","a","a") , "fac2"=c(1,1,2,2,1,1,2,2) , "fac3"=c("X","X","X","X","X","X","X","X") ) createDesignMatrix(eset1) createDesignMatrix(eset2) createDesignMatrix(eset3) createDesignMatrix(eset4) createDesignMatrix(eset5) createDesignMatrix(eset6) createDesignMatrix(eset7) createDesignMatrix(eset8) createDesignMatrix(eset9) createDesignMatrix(eset10) createDesignMatrix(eset11) createDesignMatrix(eset12) createDesignMatrix(eset13) # "Un-analysable" data set - all arrays are from the same class - gives an error. Note that we've commented this out so that we don't get errors which would make the package fail the Bioconductor checks! # createDesignMatrix(eset14) # "Non-factorial" data set - there are no arrays for fac1="b" and fac2=2. In this case only the first factor (fac1) is used. createDesignMatrix(eset15) }
Class to contain and describe results of a differential expression (DE) analysis. The main components are statistic which hold the results of any statistic (e.g. p-values, PPLR values, etc.), and FC which hold the fold changes.
DEResult objects will generally be created using one of the functions pumaDE, calculateLimma, calculateFC or calculateTtest.
Objects can also be created from scratch:
new("DEResult")
new("DEResult",
statistic=matrix()
, FC=matrix()
, statisticDescription="unknown"
, DEMethod="unknown"
)
statistic:Object of class "matrix" holding the statistics returned by the DE method.
FC:Object of class "matrix" holding the fold changes returned by the DE method.
statisticDescription:A text description of the contents of the statistic slot.
DEMethod:A string indicating which DE method was used to create the object.
Class-specific methods.
statistic(DEResult), statistic(DEResult,matrix)<-
Access and
set the statistic slot.
FC(DEResult), FC(DEResult,matrix)<-
Access and
set the FC slot.
statisticDescription(DEResult), statisticDescription(DEResult,character)<-
Access and
set the statisticDescription slot.
DEMethod(DEResult), DEMethod(DEResult,character)<-
Access and
set the DEMethod slot.
pLikeValues(object, contrast=1, direction="either")Access the statistics of an object of class DEResult, converted to "p-like values". If the object holds information on more than one contrast, only the values of the statistic for contrast number contrast are given. Direction can be "either" (meaning we want order genes by probability of being either up- or down-regulated), "up" (meaning we want to order genes by probability of being up-regulated), or "down" (meaning we want to order genes by probability of being down-regulated). "p-like values" are defined as values between 0 and 1, where 0 identifies the highest probability of being differentially expressed, and 1 identifies the lowest probability of being differentially expressed. We use this so that we can easily compare results from methods that provide true p-values (e.g. calculateLimma) and methods methods that do not provide p-values (e.g. pumaDE). For objects created using pumaDE, this returns 1-PPLR if the direction is "up", PPLR if direction is "down", and 1-abs(2*(PPLR-0.5)) if direction is "either". For objects created using calculateLimma or calculateTtest, this returns the p-value if direction is "either", ((p-1 * sign(FC))/2)+ 0.5, if the direction is "up", and ((1-p * sign(FC))/2)+ 0.5 if the direction is "down". For all other methods, this returns the rank of the appropriate statistic, scaled to lie between 0 and 1. contrast will be returned.
topGenes(object, numberOfGenes=1, contrast=1, direction="either")Returns the index numbers (row numbers) of the genes determined to be most likely to be differentially expressed. numberOfGenes specifies the number of genes to be returned by the function. If the object holds information on more than one contrast, only the values of the statistic for contrast number contrast are given. Direction can be "either" (meaning we want order genes by probability of being either up- or down-regulated), "up" (meaning we want to order genes by probability of being up-ragulated), or "down" (meaning we want to order genes by probability of being down-regulated). Note that genes are ordered by "p-like values" (see pLikeValues). object is an object of class DEResult.
topGeneIDs(object, numberOfGenes=1, contrast=1, direction="either")Returns the Affy IDs (row names) of the genes determined to be most likely to be differentially expressed. numberOfGenes specifies the number of genes to be returned by the function. If the object holds information on more than one contrast, only the values of the statistic for contrast number contrast are given. Direction can be "either" (meaning we want order genes by probability of being either up- or down-regulated), "up" (meaning we want to order genes by probability of being up-ragulated), or "down" (meaning we want to order genes by probability of being down-regulated). Note that genes are ordered by "p-like values" (see pLikeValues). object is an object of class DEResult.
numberOfProbesets(object)Returns the number of probesets (number of rows) in an object of class DEResult. This method is synonymous with numberOfGenes.
numberOfGenes(object)Returns the number of probesets (number of rows) in an object of class DEResult. This method is synonymous with numberOfProbesets.
numberOfContrasts(object)Returns the number of contrasts (number of columns) in an object of class DEResult.
write.reslts(object)signature(x = "DEResult"): writes the statistics and related fold changes (FCs) to
files. It takes the same arguments as write.table. The argument "file" does not need to set any
extension. The different file marks and extension "csv" will be added automatically. The default file name is "tmp".
In the final results, statistics are in the file "tmp\_statistics.csv", and FCs are in
"tmp\_FCs.csv" respectively.
Standard generic methods:
show(object)Informatively display object contents.
Richard D. Pearson
Related methods pumaDE, calculateLimma, calculateFC or calculateTtest.
if(FALSE){ ## Create an example DEResult object # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) esetDE <- pumaDE(eset_comb) ## Use some of the methods statisticDescription(esetDE) DEMethod(esetDE) numberOfProbesets(esetDE) numberOfContrasts(esetDE) topGenes(esetDE) topGenes(esetDE, 3) pLikeValues(esetDE)[topGenes(esetDE,3)] topGeneIDs(esetDE, 3) topGeneIDs(esetDE, 3, direction="down") ## save the expression results into files write.reslts(esetDE, file="example") }if(FALSE){ ## Create an example DEResult object # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) esetDE <- pumaDE(eset_comb) ## Use some of the methods statisticDescription(esetDE) DEMethod(esetDE) numberOfProbesets(esetDE) numberOfContrasts(esetDE) topGenes(esetDE) topGenes(esetDE, 3) pLikeValues(esetDE)[topGenes(esetDE,3)] topGeneIDs(esetDE, 3) topGeneIDs(esetDE, 3, direction="down") ## save the expression results into files write.reslts(esetDE, file="example") }
This function calculates the complementary error function of an input x.
erfc(x)erfc(x)
x |
a numeric, the input. |
erfc is implemented using the function qnorm.
The return is a numeric.
Xuejun Liu
erfc(0.5)erfc(0.5)
This data is created by applying mmgmos to the Dilution AffyBatch object from the affydata package.
data(eset_mmgmos)data(eset_mmgmos)
An object of class ExpressionSet.
see Dilution
This data is an artificial example of the mean gene expression levels from golden spike-in data set in Choe et al. (2005).
data(exampleE)data(exampleE)
A 200x6 matrix including 200 genes and 6 chips. The first 3 chips are replicates for C condition and the last 3 chips are replicates for S conditon.
Choe,S.E., Boutros,M., Michelson,A.M., Church,G.M., Halfon,M.S.: Preferred analysis methods for Affymetrix GeneChips revealed by a wholly defined control dataset. Genome Biology, 6 (2005) R16.
This data is an artificial example of the standard deviation for gene exapression levels from golden spike-in data set in Choe et al. (2005).
data(exampleStd)data(exampleStd)
A 200x6 matrix including 200 genes and 6 chips. The first 3 chips are replicates for C condition and the last 3 chips are replicates for S conditon.
Choe,S.E., Boutros,M., Michelson,A.M., Church,G.M., Halfon,M.S.: Preferred analysis methods for Affymetrix GeneChips revealed by a wholly defined control dataset. Genome Biology, 6 (2005) R16.
This is a class representation for Affymetrix GeneChip probe level data.
The main component are the intensities, estimated expression levels and the confidence
of expression levels from multiple arrays
of the same CDF type. In extends ExpressionSet.
Objects can be created by calls of the form new("exprReslt", ...).
prcfive:Object of class "matrix" representing the 5 percentile of the observed expression levels. This is a matrix with columns representing patients or cases and rows representing genes.
prctwfive:Object of class "matrix" representing the 25 percentile of the observed expression levels. This is a matrix with columns representing patients or cases and rows representing genes.
prcfifty:Object of class "matrix" representing the 50 percentile of the observed expression levels. This is a matrix with columns representing patients or cases and rows representing genes.
prcsevfive:Object of class "matrix" representing the 75 percentile of the observed expression levels. This is a matrix with columns representing patients or cases and rows representing genes.
prcninfive:Object of class "matrix" representing the 95 percentile of the observed expression levels. This is a matrix with columns representing patients or cases and rows representing genes.
phenoData:Object of class "phenoData" inherited from
ExpressionSet.
annotation:A character string identifying the
annotation that may be used for the ExpressionSet instance.
Class "ExpressionSet", directly.
signature(object = "exprReslt"): obtains the standard error of the estimated
expression levels.
signature(object = "exprReslt"): replaces the standard error of the estimated
expression levels.
signature(object = "exprReslt"): obtains the 50 percentile of the estimated
expression levels.
signature(object = "exprReslt"): replaces the 50 percentile of the estimated
expression levels.
signature(object = "exprReslt"): obtains the 5 percentile of the estimated
expression levels.
signature(object = "exprReslt"): replaces the 5 percentile of the estimated
expression levels.
signature(object = "exprReslt"): obtains the 95 percentile of the estimated
expression levels.
signature(object = "exprReslt"): replaces the 95 percentile of the estimated
expression levels.
signature(object = "exprReslt"): obtains the 75 percentile of the estimated
expression levels.
signature(object = "exprReslt"): replaces the 75 percentile of the estimated
expression levels.
signature(object = "exprReslt"): obtains the 25 percentile of the estimated
expression levels.
signature(object = "exprReslt"): replaces the 25 percentile of the estimated
expression levels.
signature(object = "exprReslt"): renders information about the exprReslt in a concise
way on stdout.
signature(x = "exprReslt"): writes the expression levels and related confidences to
files. It takes the same arguments as write.table. The argument "file" does not need to set any
extension. The different file marks and extension "csv" will be added automatically. The default file name is "tmp".
In the final results, expression levels are in the file "tmp\_exprs.csv", standard deviations in
"tmp\_se.csv", 5 percentiles in "tmp\_prctile5.csv", likewise, 25, 50, 75 and 95 percentiles in "tmp\_prctile25.csv",
"tmp\_prctile50.csv", "tmp\_prctile75.csv" and "tmp\_prctile95.csv" respectively.
Xuejun Liu, Magnus Rattray, Marta Milo, Neil D. Lawrence, Richard D. Pearson
Related method mmgmos and related class ExpressionSet.
if(FALSE){ ## load example data from package affydata # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) ## save the expression results into files write.reslts(eset_mmgmos, file="example") }if(FALSE){ ## load example data from package affydata # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) ## save the expression results into files write.reslts(eset_mmgmos, file="example") }
This function converts an object of FeatureSet into an object of class
exprReslt using the gamma model for hta2.0 chips.
This function obtains confidence of measures, standard deviation and 5,
25, 50, 75 and 95 percentiles, as well as the estimated expression levels.
gmhta( object ,background=FALSE ,gsnorm=c("median", "none", "mean", "meanlog") ,savepar=FALSE ,eps=1.0e-6 ,addConstant = 0 ,cl=NULL ,BatchFold=10 )gmhta( object ,background=FALSE ,gsnorm=c("median", "none", "mean", "meanlog") ,savepar=FALSE ,eps=1.0e-6 ,addConstant = 0 ,cl=NULL ,BatchFold=10 )
object |
an object of |
background |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
addConstant |
numeric. This is an experimental feature and should not generally be changed from the default value. |
cl |
This function can be parallelised by setting parameter cl. For more details, please refer to the vignette. |
BatchFold |
we divide tasks into BatchFold*n jobs where n is the number of cluster nodes. The first n jobs are placed on the n nodes. When the first job is completed,the next job is placed on the available node. This continues until all jobs are completed. The default value is ten. The user also can change the value according to the number of cluster nodes n. We suggest that for bigger n BatchFold should be smaller. |
The obtained expression measures are in log base 2 scale. Using the known relationships between genes, transcripts and probes, we propose a gamma model for hta2.0 data to calculate transcript and gene expression levels. The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied. This function can be parallelised by setting parameter cl. For more details, please refer to the vignette.
A list of two object of class exprReslt.
Xuejun Liu,WuJun Zhang,Zhenzhu gao, Magnus Rattray
XueJun Liu. (2013) puma 3.0: improved uncertainty propagation methods for gene and transcript expression analysis, BMC Bioinformatics, 14:39.
Manhong Dai, Pinglang Wang,Andrew D. Boyd. (2005) Evolving gene/transcript definitions significantly alter the interpretation of GeneChip data,Nucleic Acid Research 33(20):e175.
Related class exprReslt-class
if(FALSE){ ## The following scripts show the use of the method. #library(puma) ## load CEL files # object<-read.celfiles("celnames") #eset<-gmhta(object,gsnorm="none",cl=cl) }if(FALSE){ ## The following scripts show the use of the method. #library(puma) ## load CEL files # object<-read.celfiles("celnames") #eset<-gmhta(object,gsnorm="none",cl=cl) }
This function converts an object of FeatureSet into an object of class
exprReslt using the gamma model for exon chips.
This function obtains confidence of measures, standard deviation and 5,
25, 50, 75 and 95 percentiles, as well as the estimated expression levels.
gmoExon( object ,exontype = c("Human", "Mouse", "Rat") ,background=FALSE ,gsnorm=c("median", "none", "mean", "meanlog") ,savepar=FALSE ,eps=1.0e-6 ,addConstant = 0 ,cl=NULL ,BatchFold=10 )gmoExon( object ,exontype = c("Human", "Mouse", "Rat") ,background=FALSE ,gsnorm=c("median", "none", "mean", "meanlog") ,savepar=FALSE ,eps=1.0e-6 ,addConstant = 0 ,cl=NULL ,BatchFold=10 )
object |
an object of |
exontype |
character. specifying the type of exon chip. |
background |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
addConstant |
numeric. This is an experimental feature and should not generally be changed from the default value. |
cl |
This function can be parallelised by setting parameter cl. For more details, please refer to the vignette. |
BatchFold |
we divide tasks into BatchFold*n jobs where n is the number of cluster nodes. The first n jobs are placed on the n nodes. When the first job is completed,the next job is placed on the available node. This continues until all jobs are completed. The default value is ten. The user also can change the value according to the number of cluster nodes n. We suggest that for bigger n BatchFold should be smaller. |
The obtained expression measures are in log base 2 scale. Using the known relationships between genes, transcripts and probes, we propose a gamma model for exon array data to calculate transcript and gene expression levels. The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied. This function can be parallelised by setting parameter cl. For more details, please refer to the vignette.
A list of two object of class exprReslt.
Xuejun Liu, Zhenzhu gao, Magnus Rattray, Marta Milo, Neil D. Lawrence
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics, 21:3637-3644.
Milo,M., Niranjan,M., Holley,M.C., Rattray,M. and Lawrence,N.D. (2004) A probabilistic approach for summarising oligonucleotide gene expression data, technical report available upon request.
Milo,M., Fazeli,A., Niranjan,M. and Lawrence,N.D. (2003) A probabilistic model for the extractioin of expression levels from oligonucleotide arrays, Biochemical Society Transactions, 31: 1510-1512.
Peter Spellucci. DONLP2 code and accompanying documentation. Electronically available via http://plato.la.asu.edu/donlp2.html
Risueno A, Fontanillo C, Dinger ME, De Las Rivas J. GATExplorer: genomic and transcriptomic explorer; mapping expression probes to gene loci, transcripts, exons and ncRNAs. BMC Bioinformatics.2010.
Related class exprReslt-class
if(FALSE){ ## The following scripts show the use of the method. ## load CEL files # celFiles<-c("SR20070419HEX01.CEL", "SR20070419HEX02.CEL","SR20070419HEX06.CEL","SR20070419HEX07.CEL) #oligo_object.exon<-read.celfiles(celFiles); ## use method gmoExon to calculate the expression levels and related confidence ## of the measures for the example data #eset_gmoExon<-gmoExon(oligo_object.exon,exontype="Human",gsnorm="none",cl=cl) }if(FALSE){ ## The following scripts show the use of the method. ## load CEL files # celFiles<-c("SR20070419HEX01.CEL", "SR20070419HEX02.CEL","SR20070419HEX06.CEL","SR20070419HEX07.CEL) #oligo_object.exon<-read.celfiles(celFiles); ## use method gmoExon to calculate the expression levels and related confidence ## of the measures for the example data #eset_gmoExon<-gmoExon(oligo_object.exon,exontype="Human",gsnorm="none",cl=cl) }
This function calculates the combined (from replicates) signal for each condition using Bayesian models, which are added a hidden variable to represent the true expression for each gene on each chips. The inputs are gene expression levels and the probe-level standard deviations associated with expression measurements for each gene on each chip. The outputs include gene expression levels and standard deviation for each condition.
hcomb(e, se, replicates, max_num=c(200,500,1000),gsnorm=FALSE, eps=1.0e-6)hcomb(e, se, replicates, max_num=c(200,500,1000),gsnorm=FALSE, eps=1.0e-6)
e |
a data frame containing the expression level for each gene on each chip. |
se |
a data frame containing the standard deviation of gene expression levels. |
replicates |
a vector indicating which chip belongs to which condition. |
max_num |
integer. The maximum number of iterations controls the convergence. |
gsnorm |
logical specifying whether do global scaling normalisation or not. |
eps |
a numeric, optimisation parameter. |
Each element in replicate represents the condition of the chip which is in the same column order as in the expression and standard deviation matrix files.
The max\_num is used to control the maximum number of the iterations in the EM algorithm. The best value of the max\_num is from 200 to 1000, and should be set 200 at least. The default value is 200.
The result is a data frame with components named 'M1', 'M2', and so on, which represent the mean expression values for condition 1, condition 2, and so on. It also has components named 'Std1', 'Std2', and so on, which represent the standard deviation of the gene expression values for condition 1, condtion 2, and so on.
Li Zhang, Xuejun Liu
Gelman,A., Carlin,J.B., Stern,H.S., Rubin,D.B., Bayesian data analysis. London: Chapman & Hall; 1995.
Zhang,L. and Liu,X. (2009) An improved probabilistic model for finding differential gene expression, technical report available request.
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2006) Probe-level variances improve accuracy in detecting differential gene expression, Bioinformatics, 22(17):2107-13.
Related method pumaCombImproved, mmgmos and pplr
if(FALSE){ data(exampleE) data(exampleStd) r<-hcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2)) }if(FALSE){ data(exampleE) data(exampleStd) r<-hcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2)) }
The pre-estimated parameters of log normal distribution of ,
which is the fraction of specific signal binding to mismatch probe.
data(hgu95aphis)data(hgu95aphis)
The format is: num [1:3] 0.171 -1.341 0.653
The current values of hgu95aphis are estimated from Affymetrix spike-in data sets.
It was loaded in the method "mmgmos".
hgu95aphis[1:3] is respectively the mode, mean and variance of the log normal distribution
of , and hgu95aphis[1] is also the intial value of in the model optimisation.
The principle of this function is as same as the function gmoExon.This function separately calculates gene expression values by the conditions and then combined every condition's results, and normalises them finally.
igmoExon( cel.path ,SampleNameTable ,exontype = c("Human", "Mouse", "Rat") ,background=FALSE ,gsnorm=c("median", "none", "mean", "meanlog") ,savepar=FALSE ,eps=1.0e-6 ,addConstant = 0 ,condition=c("Yes","No") ,cl=NULL ,BatchFold=10 )igmoExon( cel.path ,SampleNameTable ,exontype = c("Human", "Mouse", "Rat") ,background=FALSE ,gsnorm=c("median", "none", "mean", "meanlog") ,savepar=FALSE ,eps=1.0e-6 ,addConstant = 0 ,condition=c("Yes","No") ,cl=NULL ,BatchFold=10 )
cel.path |
The directory where you put the CEL files. |
SampleNameTable |
It is a tab-separated table with two columns,ordered by "Celnames","Condition" |
exontype |
character. specifying the type of exon chip. |
background |
Logical value. If |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
addConstant |
numeric. This is an experimental feature and should not generally be changed from the default value. |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
condition |
Yes or No. “Yes” means the |
cl |
This function can be parallelised by setting parameter cl. For more details, please refer to the vignette. |
BatchFold |
we divide tasks into BatchFold*n jobs where n is the number of cluster nodes. The first n jobs are placed on the n nodes. When the first job is completed,the next job is placed on the available node. This continues until all jobs are completed. The default value is ten. The user also can change the value according to the number of cluster nodes n. We suggest that for bigger n BatchFold should be smaller. |
The obtained expression measures are in log base 2 scale. Using the known relationships between genes, transcripts and probes, we propose a gamma model for exon array data to calculate transcript and gene expression levels. The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied.
A list of two object of class exprReslt.
Xuejun Liu, Zhenzhu gao, Magnus Rattray, Marta Milo, Neil D. Lawrence
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics, 21:3637-3644.
Milo,M., Niranjan,M., Holley,M.C., Rattray,M. and Lawrence,N.D. (2004) A probabilistic approach for summarising oligonucleotide gene expression data, technical report available upon request.
Milo,M., Fazeli,A., Niranjan,M. and Lawrence,N.D. (2003) A probabilistic model for the extractioin of expression levels from oligonucleotide arrays, Biochemical Society Transactions, 31: 1510-1512.
Peter Spellucci. DONLP2 code and accompanying documentation. Electronically available via http://plato.la.asu.edu/donlp2.html
Risueno A, Fontanillo C, Dinger ME, De Las Rivas J. GATExplorer: genomic and transcriptomic explorer; mapping expression probes to gene loci, transcripts, exons and ncRNAs. BMC Bioinformatics.2010.
Related class exprReslt-class
if(FALSE){ ## The following scripts show the use of the method. ## load CEL files # cel.path<-cel.path; # SampleNameTable<-"SampleNameTable" #eset_igmoExon<-igmoExon(cel.path="cel.path" # , SampleNameTable="SampleNameTable" # , exontype="Human" # , gsnorm="none", condition="Yes",cl=cl) }if(FALSE){ ## The following scripts show the use of the method. ## load CEL files # cel.path<-cel.path; # SampleNameTable<-"SampleNameTable" #eset_igmoExon<-igmoExon(cel.path="cel.path" # , SampleNameTable="SampleNameTable" # , exontype="Human" # , gsnorm="none", condition="Yes",cl=cl) }
This function converts CEL files into an exprReslt using mgmos.
justmgMOS(..., filenames=character(0), widget=getOption("BioC")$affy$use.widgets, compress=getOption("BioC")$affy$compress.cel, celfile.path=getwd(), sampleNames=NULL, phenoData=NULL, description=NULL, notes="", background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6) just.mgmos(..., filenames=character(0), phenoData=new("AnnotatedDataFrame"), description=NULL, notes="", compress=getOption("BioC")$affy$compress.cel, background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6)justmgMOS(..., filenames=character(0), widget=getOption("BioC")$affy$use.widgets, compress=getOption("BioC")$affy$compress.cel, celfile.path=getwd(), sampleNames=NULL, phenoData=NULL, description=NULL, notes="", background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6) just.mgmos(..., filenames=character(0), phenoData=new("AnnotatedDataFrame"), description=NULL, notes="", compress=getOption("BioC")$affy$compress.cel, background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6)
... |
file names separated by comma. |
filenames |
file names in a character vector. |
widget |
a logical specifying if widgets should be used. |
compress |
are the CEL files compressed? |
celfile.path |
a character denoting the path where cel files locate. |
sampleNames |
a character vector of sample names to be used in
the |
phenoData |
an
|
description |
a |
notes |
notes. |
background |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
This method should require much less RAM than the conventional
method of first creating an FeatureSet and then running
mgmos.
Note that this expression measure is given to you in log base 2 scale. This differs from most of the other expression measure methods.
The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied.
An exprReslt.
Related class exprReslt-class and related method mgmos
This function converts CEL files into an exprReslt using mmgmos.
justmmgMOS(..., filenames=character(0), widget=getOption("BioC")$affy$use.widgets, compress=getOption("BioC")$affy$compress.cel, celfile.path=getwd(), sampleNames=NULL, phenoData=NULL, description=NULL, notes="", background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6) just.mmgmos(..., filenames=character(0), phenoData=new("AnnotatedDataFrame"), description=NULL, notes="", compress=getOption("BioC")$affy$compress.cel, background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6)justmmgMOS(..., filenames=character(0), widget=getOption("BioC")$affy$use.widgets, compress=getOption("BioC")$affy$compress.cel, celfile.path=getwd(), sampleNames=NULL, phenoData=NULL, description=NULL, notes="", background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6) just.mmgmos(..., filenames=character(0), phenoData=new("AnnotatedDataFrame"), description=NULL, notes="", compress=getOption("BioC")$affy$compress.cel, background=TRUE, gsnorm=c("median", "none", "mean", "meanlog"), savepar=FALSE, eps=1.0e-6)
... |
file names separated by comma. |
filenames |
file names in a character vector. |
widget |
a logical specifying if widgets should be used. |
compress |
are the CEL files compressed? |
celfile.path |
a character denoting the path where cel files locate. |
sampleNames |
a character vector of sample names to be used in
the |
phenoData |
an
|
description |
a |
notes |
notes |
background |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
This method should require much less RAM than the conventional
method of first creating an FeatureSet and then running
mmgmos.
Note that this expression measure is given to you in log base 2 scale. This differs from most of the other expression measure methods.
The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied.
An exprReslt.
Related class exprReslt-class and related method mmgmos
This function can be used to add legends to plots. This is almost identical to the legend function, accept it has an extra parameter, seg.len which allows the user to change the lengths of lines shown in legends.
legend2(x, y = NULL, legend, fill = NULL, col = par("col"), lty, lwd, pch, angle = 45, density = NULL, bty = "o", bg = par("bg"), box.lwd = par("lwd"), box.lty = par("lty"), pt.bg = NA, cex = 1, pt.cex = cex, pt.lwd = lwd, xjust = 0, yjust = 1, x.intersp = 1, y.intersp = 1, adj = c(0, 0.5), text.width = NULL, text.col = par("col"), merge = do.lines && has.pch, trace = FALSE, plot = TRUE, ncol = 1, horiz = FALSE, title = NULL, inset = 0, seg.len = 2)legend2(x, y = NULL, legend, fill = NULL, col = par("col"), lty, lwd, pch, angle = 45, density = NULL, bty = "o", bg = par("bg"), box.lwd = par("lwd"), box.lty = par("lty"), pt.bg = NA, cex = 1, pt.cex = cex, pt.lwd = lwd, xjust = 0, yjust = 1, x.intersp = 1, y.intersp = 1, adj = c(0, 0.5), text.width = NULL, text.col = par("col"), merge = do.lines && has.pch, trace = FALSE, plot = TRUE, ncol = 1, horiz = FALSE, title = NULL, inset = 0, seg.len = 2)
x, y
|
the x and y co-ordinates to be used to position the legend.
They can be specified by keyword or in any way which is accepted by
|
legend |
a character or expression vector.
of length |
fill |
if specified, this argument will cause boxes filled with the specified colors (or shaded in the specified colors) to appear beside the legend text. |
col |
the color of points or lines appearing in the legend. |
lty, lwd
|
the line types and widths for lines appearing in the legend. One of these two must be specified for line drawing. |
pch |
the plotting symbols appearing in the legend, either as vector of 1-character strings, or one (multi character) string. Must be specified for symbol drawing. |
angle |
angle of shading lines. |
density |
the density of shading lines, if numeric and
positive. If |
bty |
the type of box to be drawn around the legend. The allowed
values are |
bg |
the background color for the legend box. (Note that this is
only used if |
box.lty, box.lwd
|
the line type and width for the legend box. |
pt.bg |
the background color for the |
cex |
character expansion factor relative to current
|
pt.cex |
expansion factor(s) for the points. |
pt.lwd |
line width for the points, defaults to the one for
lines, or if that is not set, to |
xjust |
how the legend is to be justified relative to the legend x location. A value of 0 means left justified, 0.5 means centered and 1 means right justified. |
yjust |
the same as |
x.intersp |
character interspacing factor for horizontal (x) spacing. |
y.intersp |
the same for vertical (y) line distances. |
adj |
numeric of length 1 or 2; the string adjustment for legend
text. Useful for y-adjustment when |
text.width |
the width of the legend text in x ( |
text.col |
the color used for the legend text. |
merge |
logical; if |
trace |
logical; if |
plot |
logical. If |
ncol |
the number of columns in which to set the legend items (default is 1, a vertical legend). |
horiz |
logical; if |
title |
a character string or length-one expression giving a
title to be placed at the top of the legend. Other objects will be
coerced by |
inset |
inset distance(s) from the margins as a fraction of the plot region when legend is placed by keyword. |
seg.len |
numeric specifying length of lines in legend. |
Arguments x, y, legend are interpreted in a non-standard way to
allow the coordinates to be specified via one or two arguments.
If legend is missing and y is not numeric, it is assumed
that the second argument is intended to be legend and that the
first argument specifies the coordinates.
The coordinates can be specified in any way which is accepted by
xy.coords. If this gives the coordinates of one point,
it is used as the top-left coordinate of the rectangle containing the
legend. If it gives the coordinates of two points, these specify
opposite corners of the rectangle (either pair of corners, in any
order).
The location may also be specified by setting x to a single
keyword from the list "bottomright", "bottom",
"bottomleft", "left", "topleft",
"top", "topright", "right" and
"center". This places the legend on the inside of the plot
frame at the given location. Partial argument matching is used. The
optional inset argument specifies how far the legend is inset
from the plot margins. If a single value is given, it is used for
both margins; if two values are given, the first is used for x-
distance, the second for y-distance.
“Attribute” arguments such as col, pch, lty,
etc, are recycled if necessary. merge is not.
Points are drawn after lines in order that they can cover the
line with their background color pt.bg, if applicable.
See the examples for how to right-justify labels.
A list with list components
rect |
a list with components
|
text |
a list with components
|
returned invisibly.
Richard Pearson (modified from original graphics package function.)
Becker, R. A., Chambers, J. M. and Wilks, A. R. (1988) The New S Language. Wadsworth \& Brooks/Cole.
Murrell, P. (2005) R Graphics. Chapman & Hall/CRC Press.
if(FALSE){ x <- seq(-pi, pi, len = 65) plot(x, sin(x), type = "l", ylim = c(-1.2, 1.8), col = 3, lty = 2) points(x, cos(x), pch = 3, col = 4) lines(x, tan(x), type = "b", lty = 1, pch = 4, col = 6) title("legend(..., lty = c(2, -1, 1), pch = c(-1,3,4), merge = TRUE)", cex.main = 1.1) legend2(-1, 1.9, c("sin", "cos", "tan"), col = c(3,4,6), text.col = "green4", lty = c(2, -1, 1), pch = c(-1, 3, 4), merge = TRUE, bg = 'gray90', seg.len=6) }if(FALSE){ x <- seq(-pi, pi, len = 65) plot(x, sin(x), type = "l", ylim = c(-1.2, 1.8), col = 3, lty = 2) points(x, cos(x), pch = 3, col = 4) lines(x, tan(x), type = "b", lty = 1, pch = 4, col = 6) title("legend(..., lty = c(2, -1, 1), pch = c(-1,3,4), merge = TRUE)", cex.main = 1.1) legend2(-1, 1.9, c("sin", "cos", "tan"), col = c(3,4,6), text.col = "green4", lty = c(2, -1, 1), pch = c(-1, 3, 4), merge = TRUE, bg = 'gray90', seg.len=6) }
This function prints the license under which puma is made available.
license.puma()license.puma()
Null.
Richard Pearson (based on the license.cosmo function from the cosmo package)
license.puma()license.puma()
This calculates the mean Euclidean distance between the rows of two matrices. It is used in the function pumaPCA
matrixDistance( matrixA , matrixB )matrixDistance( matrixA , matrixB )
matrixA |
the first matrix |
matrixB |
the second matrix |
A numeric giving the mean distance
Richard D. Pearson
Related class pumaPCA
if(FALSE){ show(matrixDistance(matrix(1,2,2),matrix(2,2,2))) }if(FALSE){ show(matrixDistance(matrix(1,2,2),matrix(2,2,2))) }
This function converts an object of class FeatureSet into an object of class
exprReslt using the modified gamma Model for Oligonucleotide Signal
(multi-mgMOS). This function obtains confidence of measures, standard deviation and 5,
25, 50, 75 and 95 percentiles, as well as the estimated expression levels.
mgmos( object , background=FALSE , replaceZeroIntensities=TRUE , gsnorm=c("median", "none", "mean", "meanlog") , savepar=FALSE , eps=1.0e-6 )mgmos( object , background=FALSE , replaceZeroIntensities=TRUE , gsnorm=c("median", "none", "mean", "meanlog") , savepar=FALSE , eps=1.0e-6 )
object |
an object of |
background |
Logical value. If |
replaceZeroIntensities |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
The obtained expression measures are in log base 2 scale.
The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied.
There are 4*n columns in file par\_mgmos.txt, n is the number of chips. Every 4 columns are parameters for a chip. Among every 4 columns, the first one is for 'alpha' values, the 2nd one is for 'a' values, The 3rd column is for 'c' and the final column is values for 'd'.
An object of class exprReslt.
Xuejun Liu, Magnus Rattray, Marta Milo, Neil D. Lawrence
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics, 21:3637-3644.
Milo,M., Niranjan,M., Holley,M.C., Rattray,M. and Lawrence,N.D. (2004) A probabilistic approach for summarising oligonucleotide gene expression data, technical report available upon request.
Milo,M., Fazeli,A., Niranjan,M. and Lawrence,N.D. (2003) A probabilistic model for the extractioin of expression levels from oligonucleotide arrays, Biochemical Society Transactions, 31: 1510-1512.
Peter Spellucci. DONLP2 code and accompanying documentation. Electronically available via http://plato.la.asu.edu/donlp2.html
Related class exprReslt-class and related method mmgmos
if(FALSE){ ## Code commented out to speed up checks ## load example data from package affydata # if (require(pumadata)&&require(puma)){ # data(oligo.estrogen) # use method mgMOS to calculate the expression levels and related confidence # of the measures for the example data # eset<-mgmos(oligo.estrogen,gsnorm="none") #} }if(FALSE){ ## Code commented out to speed up checks ## load example data from package affydata # if (require(pumadata)&&require(puma)){ # data(oligo.estrogen) # use method mgMOS to calculate the expression levels and related confidence # of the measures for the example data # eset<-mgmos(oligo.estrogen,gsnorm="none") #} }
This function converts an object of class FeatureSet into an object of class
exprReslt using the Multi-chip modified gamma Model for Oligonucleotide Signal
(multi-mgMOS). This function obtains confidence of measures, standard deviation and 5,
25, 50, 75 and 95 percentiles, as well as the estimated expression levels.
mmgmos( object , background=FALSE , replaceZeroIntensities=TRUE , gsnorm=c("median", "none", "mean", "meanlog") , savepar=FALSE , eps=1.0e-6 , addConstant = 0 )mmgmos( object , background=FALSE , replaceZeroIntensities=TRUE , gsnorm=c("median", "none", "mean", "meanlog") , savepar=FALSE , eps=1.0e-6 , addConstant = 0 )
object |
an object of |
background |
Logical value. If |
replaceZeroIntensities |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
addConstant |
numeric. This is an experimental feature and should not generally be changed from the default value. |
The obtained expression measures are in log base 2 scale.
The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied.
There are 2*n+2 columns in file par\_mmgmos.txt, n is the number of chips. The first n columns are 'alpha' values for n chips, the next n columns are 'a' values for n chips, column 2*n+1 is 'c' values and the final column is values for 'd'. The file phi\_mmgmos.txt keeps the final optimal value of 'phi'.
An object of class exprReslt.
Xuejun Liu, Magnus Rattray, Marta Milo, Neil D. Lawrence
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics 21: 3637-3644.
Milo,M., Niranjan,M., Holley,M.C., Rattray,M. and Lawrence,N.D. (2004) A probabilistic approach for summarising oligonucleotide gene expression data, technical report available upon request.
Milo,M., Fazeli,A., Niranjan,M. and Lawrence,N.D. (2003) A probabilistic model for the extractioin of expression levels from oligonucleotide arrays, Biochemical Society Transactions, 31: 1510-1512.
Peter Spellucci. DONLP2 code and accompanying documentation. Electronically available via http://plato.la.asu.edu/donlp2.html
Related class exprReslt-class and related method mgmos
if(FALSE){ ## Code commented out to speed up checks ## load example data from package affydata # if (require(pumadata)&&require(puma)){ # data(oligo.estrogen) ## use method mmgMOS to calculate the expression levels and related confidence ## of the measures for the example data # eset<-mmgmos(oligo.estrogen,gsnorm="none") #} }if(FALSE){ ## Code commented out to speed up checks ## load example data from package affydata # if (require(pumadata)&&require(puma)){ # data(oligo.estrogen) ## use method mmgMOS to calculate the expression levels and related confidence ## of the measures for the example data # eset<-mmgmos(oligo.estrogen,gsnorm="none") #} }
This function is only included for backwards compatibility with the pplr package. This function is now superceded by pumaNormalize.
This function does the global scaling normalisation.
normalisation.gs(x)normalisation.gs(x)
x |
a matrix or data frame which contains gene expression level on log2 scale. |
Each row of x is related to a gene and each column is related to a chip.
The return matrix is in the same format as the input x.
Xuejun Liu, Marta Milo, Neil D. Lawrence, Magnus Rattray
if(FALSE){ data(exampleE) exampleE.normalised<-normalisation.gs(exampleE) data(Clust.exampleE) Clust.exampleE.normalised<-normalisation.gs(Clust.exampleE) }if(FALSE){ data(exampleE) exampleE.normalised<-normalisation.gs(exampleE) data(Clust.exampleE) Clust.exampleE.normalised<-normalisation.gs(Clust.exampleE) }
Often when evaluating a differential expression method, we are interested in how well a classifier performs for very small numbers of false positives. This method gives one way of calculating this, by determining the number of false positives for a set proportion of true positives.
numFP(scores, truthValues, TPRate = 0.5)numFP(scores, truthValues, TPRate = 0.5)
scores |
A vector of scores. This could be, e.g. one of the columns of the statistics of a |
truthValues |
A boolean vector indicating which scores are True Positives. |
TPRate |
A number between 0 and 1 identify the proportion of true positives for which we wish to determine the number of false positives. |
An integer giving the number of false positives.
Richard D. Pearson
Related methods plotROC and calcAUC.
if(FALSE){ class1a <- rnorm(1000,0.2,0.1) class2a <- rnorm(1000,0.6,0.2) class1b <- rnorm(1000,0.3,0.1) class2b <- rnorm(1000,0.5,0.2) scores_a <- c(class1a, class2a) scores_b <- c(class1b, class2b) classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) print(numFP(scores_a, classElts)) print(numFP(scores_b, classElts)) }if(FALSE){ class1a <- rnorm(1000,0.2,0.1) class2a <- rnorm(1000,0.6,0.2) class1b <- rnorm(1000,0.3,0.1) class2b <- rnorm(1000,0.5,0.2) scores_a <- c(class1a, class2a) scores_b <- c(class1b, class2b) classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) print(numFP(scores_a, classElts)) print(numFP(scores_b, classElts)) }
This is really an internal function used to determine how many factors to use in design and contrast matrices
numOfFactorsToUse(eset)numOfFactorsToUse(eset)
eset |
An object of class
|
An integer denoting the number of factors to be used.
Richard D. Pearson
Related methods createDesignMatrix and createContrastMatrix
if(FALSE){ # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) numOfFactorsToUse(eset_mmgmos) }if(FALSE){ # Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) numOfFactorsToUse(eset_mmgmos) }
Often when evaluating a differential expression method, we are interested in how well a classifier performs for very small numbers of true positives. This method gives one way of calculating this, by determining the number of true positives for a set proportion of false positives.
numTP(scores, truthValues, FPRate = 0.5)numTP(scores, truthValues, FPRate = 0.5)
scores |
A vector of scores. This could be, e.g. one of the columns of the statistics of a |
truthValues |
A boolean vector indicating which scores are True Positives. |
FPRate |
A number between 0 and 1 identify the proportion of flase positives for which we wish to determine the number of true positives. |
An integer giving the number of true positives.
Richard D. Pearson
Related methods numFP, plotROC and calcAUC.
if(FALSE){ class1a <- rnorm(1000,0.2,0.1) class2a <- rnorm(1000,0.6,0.2) class1b <- rnorm(1000,0.3,0.1) class2b <- rnorm(1000,0.5,0.2) scores_a <- c(class1a, class2a) scores_b <- c(class1b, class2b) classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) print(numTP(scores_a, classElts)) print(numTP(scores_b, classElts)) }if(FALSE){ class1a <- rnorm(1000,0.2,0.1) class2a <- rnorm(1000,0.6,0.2) class1b <- rnorm(1000,0.3,0.1) class2b <- rnorm(1000,0.5,0.2) scores_a <- c(class1a, class2a) scores_b <- c(class1b, class2b) classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) print(numTP(scores_a, classElts)) print(numTP(scores_b, classElts)) }
This is the original version of the pplr function as found in the pplr package. This should give exactly the same results as the pplr function. This function is only included for testing purposes and is not intended to be used. It will not be available in future versions of puma.
This function calculates the probability of positive log-ratio (PPLR) between any two specified conditions in the input data, mean and standard deviation of gene expression level for each condition.
orig_pplr(e, control, experiment)orig_pplr(e, control, experiment)
e |
a data frame containing the mean and standard deviation of gene expression levels for each condition. |
control |
an integer denoting the control condition. |
experiment |
an integer denoting the experiment condition. |
The input of 'e' should be a data frame comprising of 2*n components, where n is the number of conditions. The first 1,2,...,n components include the mean of gene expression values for conditions 1,2,...,n, and the n+1, n+2,...,2*n components contain the standard deviation of expression levels for condition 1,2,...,n.
The return is a data frame. The description of the components are below.
index |
The original row number of genes. |
cM |
The mean expression levels under control condition. |
sM |
The mean expression levels under experiment condition. |
cStd |
The standard deviation of gene expression levels under control condition. |
sStd |
The standard deviation of gene expression levels under experiment condition. |
LRM |
The mean log-ratio between control and experiment genes. |
LRStd |
The standard deviation of log-ratio between control and experiment genes. |
stat |
A statistic value which is -mean/(sqrt(2)*standard deviation). |
PPLR |
Probability of positive log-ratio. |
Xuejun Liu, Marta Milo, Neil D. Lawrence, Magnus Rattray
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2006) Probe-level variances improve accuracy in detecting differential gene expression, Bioinformatics, 22(17):2107-13.
Related method bcomb
if(FALSE){ data(exampleE) data(exampleStd) r<-bcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2),method="map") p<-orig_pplr(r,1,2) }if(FALSE){ data(exampleE) data(exampleStd) r<-bcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2),method="map") p<-orig_pplr(r,1,2) }
This is the method to plot objects of class pumaPCARes. It will produce a scatter plot of two of the principal components
## S4 method for signature 'pumaPCARes,missing' plot(..., firstComponent = 1, secondComponent = 2, useFilenames = FALSE, phenotype = pData(pumaPCARes@phenoData), legend1pos = "topright", legend2pos = "bottomright")## S4 method for signature 'pumaPCARes,missing' plot(..., firstComponent = 1, secondComponent = 2, useFilenames = FALSE, phenotype = pData(pumaPCARes@phenoData), legend1pos = "topright", legend2pos = "bottomright")
... |
Optional graphical parameters to adjust different components of the plot |
firstComponent |
Integer identifying which principal component to plot on the x-axis |
secondComponent |
Integer identifying which principal component to plot on the x-axis |
useFilenames |
Boolean. If TRUE then use filenames as plot points. Otherwise just use points. |
phenotype |
Phenotype information |
legend1pos |
String indicating where to put legend for first factor |
legend2pos |
String indicating where to put legend for second factor |
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) pumapca_mmgmos <- pumaPCA(eset_mmgmos) plot(pumapca_mmgmos)# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) pumapca_mmgmos <- pumaPCA(eset_mmgmos) plot(pumapca_mmgmos)
This produces plots of probesets of interest.
plotErrorBars( eset , probesets = if(dim(exprs(eset))[1] <= 12) 1:dim(exprs(eset))[1] else 1 , arrays = 1:dim(pData(eset))[1] # default is to use all , xlab = paste(colnames(pData(eset))[1:numOfFactorsToUse(eset)], collapse=":") , ylab = "Expression Estimate" , xLabels = apply( as.matrix(pData(eset)[arrays,1:numOfFactorsToUse(eset)]) , 1 , function(mat){paste(mat, collapse=":")} ) , ylim = NA , numOfSEs = qnorm(0.975) , globalYlim = FALSE # Not yet implemented! , plot_cols = NA , plot_rows = NA , featureNames = NA , showGeneNames = FALSE , showErrorBars = if( length(assayDataElement(eset,"se.exprs"))==0 || length(assayDataElement(eset,"se.exprs")) == sum(is.na(assayDataElement(eset,"se.exprs"))) ) FALSE else TRUE , plotColours = FALSE , log.it = if(max(exprs(eset)) > 32) TRUE else FALSE , eset_comb = NULL , jitterWidth = NA , qtpcrData = NULL , ... )plotErrorBars( eset , probesets = if(dim(exprs(eset))[1] <= 12) 1:dim(exprs(eset))[1] else 1 , arrays = 1:dim(pData(eset))[1] # default is to use all , xlab = paste(colnames(pData(eset))[1:numOfFactorsToUse(eset)], collapse=":") , ylab = "Expression Estimate" , xLabels = apply( as.matrix(pData(eset)[arrays,1:numOfFactorsToUse(eset)]) , 1 , function(mat){paste(mat, collapse=":")} ) , ylim = NA , numOfSEs = qnorm(0.975) , globalYlim = FALSE # Not yet implemented! , plot_cols = NA , plot_rows = NA , featureNames = NA , showGeneNames = FALSE , showErrorBars = if( length(assayDataElement(eset,"se.exprs"))==0 || length(assayDataElement(eset,"se.exprs")) == sum(is.na(assayDataElement(eset,"se.exprs"))) ) FALSE else TRUE , plotColours = FALSE , log.it = if(max(exprs(eset)) > 32) TRUE else FALSE , eset_comb = NULL , jitterWidth = NA , qtpcrData = NULL , ... )
eset |
An object of class |
probesets |
A vector of integers indicating the probesets to be plotted. These integers refer to the row numbers of the |
arrays |
A vector of integers indicating the arrays to be shown on plots. |
xlab |
Character string of title to appear on x-axis |
ylab |
Character string of title to appear on y-axis |
xLabels |
Vector of strings for labels of individual points on x-axis. |
ylim |
2-element numeric vector showing minimum and maximum values for y-axis. |
numOfSEs |
Numeric indicating the scaling for the error bars. The default value give error bars that include 95% of expected values. |
globalYlim |
Not yet implemented! |
plot_cols |
Integer specifying number of columns for multi-figure plot. |
plot_rows |
Integer specifying number of rows for multi-figure plot. |
featureNames |
A vector of strings for |
showGeneNames |
Boolean indicating whether to use Affy IDs as titles for each plot. |
showErrorBars |
Boolean indicating whether error bars should be shown on plots. |
plotColours |
A vector of colours to plot. |
log.it |
Boolean indicating whether expression values should be logged. |
eset_comb |
An object of class |
jitterWidth |
Numeric indicating the x-axis distance between replicates of the same condition. |
qtpcrData |
A 2-column matrix of qRT-PCR values (or other data to be plotted on the same charts). |
... |
Additional arguments to be passed to |
This function has no return value. The output is the plot created.
Richard D. Pearson
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) plotErrorBars(eset_mmgmos) plotErrorBars(eset_mmgmos,1:6)# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) plotErrorBars(eset_mmgmos) plotErrorBars(eset_mmgmos,1:6)
Stacked histogram plot of two different classes
plotHistTwoClasses( scores , class1Elements , class2Elements , space=0 , col=c("white", "grey40") , xlab="PPLR" , ylab="Number of genes" , ylim=NULL , las=0 # axis labels all perpendicular to axes , legend=c("non-spike-in genes", "spike-in genes") , inset=0.05 , minScore=0 , maxScore=1 , numOfBars=20 , main=NULL )plotHistTwoClasses( scores , class1Elements , class2Elements , space=0 , col=c("white", "grey40") , xlab="PPLR" , ylab="Number of genes" , ylim=NULL , las=0 # axis labels all perpendicular to axes , legend=c("non-spike-in genes", "spike-in genes") , inset=0.05 , minScore=0 , maxScore=1 , numOfBars=20 , main=NULL )
scores |
A numeric vector of scores (e.g. from the output of |
class1Elements |
Boolean vector, TRUE if element is in first class |
class2Elements |
Boolean vector, TRUE if element is in second class |
space |
Numeric. x-axis distance between bars |
col |
Colours for the two different classes |
xlab |
Title for the x-axis |
ylab |
Title for the y-axis |
ylim |
2-element numeric vector showing minimum and maximum values for y-axis. |
las |
See |
legend |
2-element string vector giving text to appear in legend for the two classes. |
inset |
See |
minScore |
Numeric. Minimum score to plot. |
maxScore |
Numeric. Maximum score to plot. |
numOfBars |
Integer. Number of bars to plot. |
main |
String. Main title for the plot. |
This function has no return value. The output is the plot created.
Richard D. Pearson
class1 <- rnorm(1000,0.2,0.1) class2 <- rnorm(1000,0.6,0.2) class1[which(class1<0)] <- 0 class1[which(class1>1)] <- 1 class2[which(class2<0)] <- 0 class2[which(class2>1)] <- 1 scores <- c(class1, class2) class1elts <- c(rep(TRUE,1000), rep(FALSE,1000)) class2elts <- c(rep(FALSE,1000), rep(TRUE,1000)) plotHistTwoClasses(scores, class1elts, class2elts, ylim=c(0,300))class1 <- rnorm(1000,0.2,0.1) class2 <- rnorm(1000,0.6,0.2) class1[which(class1<0)] <- 0 class1[which(class1>1)] <- 1 class2[which(class2<0)] <- 0 class2[which(class2>1)] <- 1 scores <- c(class1, class2) class1elts <- c(rep(TRUE,1000), rep(FALSE,1000)) class2elts <- c(rep(FALSE,1000), rep(TRUE,1000)) plotHistTwoClasses(scores, class1elts, class2elts, ylim=c(0,300))
Plots a Receiver Operator Characteristic (ROC) curve.
plotROC( scoresList , truthValues , includedProbesets=1:length(truthValues) , legendTitles=1:length(scoresList) , main = "PUMA ROC plot" , lty = 1:length(scoresList) , col = rep(1,length(scoresList)) , lwd = rep(1,length(scoresList)) , yaxisStat = "tpr" , xaxisStat = "fpr" , downsampling = 100 , showLegend = TRUE , showAUC = TRUE , ... )plotROC( scoresList , truthValues , includedProbesets=1:length(truthValues) , legendTitles=1:length(scoresList) , main = "PUMA ROC plot" , lty = 1:length(scoresList) , col = rep(1,length(scoresList)) , lwd = rep(1,length(scoresList)) , yaxisStat = "tpr" , xaxisStat = "fpr" , downsampling = 100 , showLegend = TRUE , showAUC = TRUE , ... )
scoresList |
A list, each element of which is a numeric vector of scores. |
truthValues |
A boolean vector indicating which scores are True Positives. |
includedProbesets |
A vector of indices indicating which scores (and truthValues) are to be used in the calculation. The default is to use all, but a subset can be used if, for example, you only want a subset of the probesets which are not True Positives to be treated as False Positives. |
legendTitles |
Vector of names to appear in legend. |
main |
Main plot title |
lty |
Line types. |
col |
Colours. |
lwd |
Line widths. |
yaxisStat |
Character string identifying what is to be plotted on the y-axis. The default is "tpr" for True Positive Rate. See |
xaxisStat |
Character string identifying what is to be plotted on the x-axis. The default is "fpr" for False Positive Rate. See |
downsampling |
See details for
|
showLegend |
Boolean. Should legend be displayed? |
showAUC |
Boolean. Should AUC values be included in legend? |
... |
Other parameters to be passed to |
This function has no return value. The output is the plot created.
Richard D. Pearson
Related method calcAUC
if(FALSE){ class1a <- rnorm(1000,0.2,0.1) class2a <- rnorm(1000,0.6,0.2) class1b <- rnorm(1000,0.3,0.1) class2b <- rnorm(1000,0.5,0.2) scores_a <- c(class1a, class2a) scores_b <- c(class1b, class2b) scores <- list(scores_a, scores_b) classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) plotROC(scores, classElts) }if(FALSE){ class1a <- rnorm(1000,0.2,0.1) class2a <- rnorm(1000,0.6,0.2) class1b <- rnorm(1000,0.3,0.1) class2b <- rnorm(1000,0.5,0.2) scores_a <- c(class1a, class2a) scores_b <- c(class1b, class2b) scores <- list(scores_a, scores_b) classElts <- c(rep(FALSE,1000), rep(TRUE,1000)) plotROC(scores, classElts) }
A plot showing error bars for genes of interest.
plotWhiskers( eset , comparisons=c(1,2) , sortMethod = c("logRatio", "PPLR") , numGenes=50 , xlim , main = "PUMA Whiskers plot" , highlightedGenes=NULL )plotWhiskers( eset , comparisons=c(1,2) , sortMethod = c("logRatio", "PPLR") , numGenes=50 , xlim , main = "PUMA Whiskers plot" , highlightedGenes=NULL )
eset |
An object of class |
comparisons |
A 2-element integer vector specifying the columns of data to be compared. |
sortMethod |
The method used to sort the genes. "logRatio" is fold change. PPLR is Probability of Positive Log Ratio (as determined by the |
numGenes |
Integer. Number of probesets to plot. |
xlim |
The x limits of the plot. See |
main |
A main title for the plot. See |
highlightedGenes |
Row numbers of probesets to highlight with an asterisk. |
This function has no return value. The output is the plot created.
Richard D. Pearson
Related method pumaDE
This function converts an object of class FeatureSet into an object of class
exprReslt using the Multi-chip modified gamma Model for Oligonucleotide Signal
(PMmulti-mgMOS). This method uses only PM probe intensites. This function obtains confidence of measures, standard deviation and 5,
25, 50, 75 and 95 percentiles, as well as the estimated expression levels.
PMmmgmos( object , background=TRUE , replaceZeroIntensities=TRUE , gsnorm=c("median", "none", "mean", "meanlog") , savepar=FALSE , eps=1.0e-6 , addConstant = 0 )PMmmgmos( object , background=TRUE , replaceZeroIntensities=TRUE , gsnorm=c("median", "none", "mean", "meanlog") , savepar=FALSE , eps=1.0e-6 , addConstant = 0 )
object |
an object of |
background |
Logical value. If |
replaceZeroIntensities |
Logical value. If |
gsnorm |
character. specifying the algorithm of global scaling normalisation. |
savepar |
Logical value. If |
eps |
Optimisation termination criteria. |
addConstant |
numeric. This is an experimental feature and should not generally be changed from the default value. |
The obtained expression measures are in log base 2 scale.
The algorithms of global scaling normalisation can be one of "median", "none", "mean", "meanlog". "mean" and "meanlog" are mean-centered normalisation on raw scale and log scale respectively, and "median" is median-centered normalisation. "none" will result in no global scaling normalisation being applied.
There are n+2 columns in file par\_pmmmgmos.txt, n is the number of chips. The first n columns are 'alpha' values for n chips, column n+1 is 'c' values and the final column is values for 'd'.
An object of class exprReslt.
Xuejun Liu, Zhenzhu Gao, Magnus Rattray, Marta Milo, Neil D. Lawrence
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics 21: 3637-3644.
Milo,M., Niranjan,M., Holley,M.C., Rattray,M. and Lawrence,N.D. (2004) A probabilistic approach for summarising oligonucleotide gene expression data, technical report available upon request.
Milo,M., Fazeli,A., Niranjan,M. and Lawrence,N.D. (2003) A probabilistic model for the extractioin of expression levels from oligonucleotide arrays, Biochemical Society Transactions, 31: 1510-1512.
Peter Spellucci. DONLP2 code and accompanying documentation. Electronically available via http://plato.la.asu.edu/donlp2.html
Related class exprReslt-class and related method mgmos
## Code commented out to speed up checks ## load example data from package pumadata #if (require(pumadata)&&require(puma)){ # data(oligo.estrogen) ## use method PMmmgMOS to calculate the expression levels and related confidence ##of the measures for the example data # eset<-PMmmgmos(oligo.estrogen,gsnorm="none") #}## Code commented out to speed up checks ## load example data from package pumadata #if (require(pumadata)&&require(puma)){ # data(oligo.estrogen) ## use method PMmmgMOS to calculate the expression levels and related confidence ##of the measures for the example data # eset<-PMmmgmos(oligo.estrogen,gsnorm="none") #}
WARNING - this function is generally not expected to be used, but is intended as an internal function. It is included for backwards compatibility with the pplr package, but may be deprecated and then hidden in future. Users should generally use pumaDE instead.
This function calculates the probability of positive log-ratio (PPLR) between any two specified conditions in the input data, mean and standard deviation of gene expression level for each condition.
pplr(e, control, experiment, sorted=TRUE)pplr(e, control, experiment, sorted=TRUE)
e |
a data frame containing the mean and standard deviation of gene expression levels for each condition. |
control |
an integer denoting the control condition. |
experiment |
an integer denoting the experiment condition. |
sorted |
Boolean. Should PPLR values be sorted by value? If FALSE, PPLR values are returned in same order as supplied. |
The input of 'e' should be a data frame comprising of 2*n components, where n is the number of conditions. The first 1,2,...,n components include the mean of gene expression values for conditions 1,2,...,n, and the n+1, n+2,...,2*n components contain the standard deviation of expression levels for condition 1,2,...,n.
The return is a data frame. The description of the components are below.
index |
The original row number of genes. |
cM |
The mean expression levels under control condition. |
sM |
The mean expression levels under experiment condition. |
cStd |
The standard deviation of gene expression levels under control condition. |
sStd |
The standard deviation of gene expression levels under experiment condition. |
LRM |
The mean log-ratio between control and experiment genes. |
LRStd |
The standard deviation of log-ratio between control and experiment genes. |
stat |
A statistic value which is -mean/(sqrt(2)*standard deviation). |
PPLR |
Probability of positive log-ratio. |
Xuejun Liu, Marta Milo, Neil D. Lawrence, Magnus Rattray
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2006) Probe-level variances improve accuracy in detecting differential gene expression, Bioinformatics, 22(17):2107-13.
Related methods pumaDE, bcomb and hcomb
data(exampleE) data(exampleStd) r<-bcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2),method="map") p<-pplr(r,1,2)data(exampleE) data(exampleStd) r<-bcomb(exampleE,exampleStd,replicates=c(1,1,1,2,2,2),method="map") p<-pplr(r,1,2)
Returns the output from pplr as an unsorted matrix (i.e. sorted according to the original sorting in the original matrix)
pplrUnsorted(p)pplrUnsorted(p)
p |
A matrix as output by |
A matrix of PPLR values
Richard D. Pearson
Related method pplr
This function clusters gene expression using a Gaussian mixture model including probe-level measurement error. The inputs are gene expression levels and the probe-level standard deviation associated with expression measurement for each gene on each chip. The outputs is the clustering results.
pumaClust(e=NULL, se=NULL, efile=NULL, sefile=NULL, subset=NULL, gsnorm=FALSE, clusters, iter.max=100, nstart=10, eps=1.0e-6, del0=0.01)pumaClust(e=NULL, se=NULL, efile=NULL, sefile=NULL, subset=NULL, gsnorm=FALSE, clusters, iter.max=100, nstart=10, eps=1.0e-6, del0=0.01)
e |
either a valid |
se |
data frame containing the standard deviation of gene expression levels. |
efile |
character, the name of the file which contains gene expression measurements. |
sefile |
character, the name of the file which contains the standard deviation of gene expression measurements. |
subset |
vector specifying the row number of genes which are clustered on. |
gsnorm |
logical specifying whether do global scaling normalisation or not. |
clusters |
integer, the number of clusters. |
iter.max |
integer, the maximum number of iterations allowed in the parameter initialisation. |
nstart |
integer, the number of random sets chosen in the parameter initialisation. |
eps |
numeric, optimisation parameter. |
del0 |
numeric, optimisation parameter. |
The input data is specified either as an ExpressionSet object (in which case se, efile and sefile will be ignored), or by e and se, or by efile and sefile.
The result is a list with components
cluster: vector, containing the membership of clusters for each gene; centers: matrix, the center of each cluster; centersigs: matrix, the center variance of each cluster; likelipergene: matrix, the likelihood of belonging to each cluster for each gene; bic: numeric, the Bayesian Information Criterion score.
Xuejun Liu, Magnus Rattray
Liu,X., Lin,K.K., Andersen,B., and Rattray,M. (2006) Propagating probe-level uncertainty in model-based gene expression clustering, BMC Bioinformatics, 8(98).
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics, 21(18):3637-3644.
Related method mmgmos and pumaClustii
data(Clust.exampleE) data(Clust.exampleStd) pumaClust.example<-pumaClust(Clust.exampleE,Clust.exampleStd,clusters=7)data(Clust.exampleE) data(Clust.exampleStd) pumaClust.example<-pumaClust(Clust.exampleE,Clust.exampleStd,clusters=7)
This function clusters gene expression by including uncertainties of gene expression measurements from probe-level analysis models and replicate information into a robust t mixture clustering model. The inputs are gene expression levels and the probe-level standard deviation associated with expression measurement for each gene on each chip. The outputs is the clustering results.
pumaClustii(e=NULL, se=NULL, efile=NULL, sefile=NULL, subset=NULL, gsnorm=FALSE, mincls, maxcls, conds, reps, verbose=FALSE, eps=1.0e-5, del0=0.01)pumaClustii(e=NULL, se=NULL, efile=NULL, sefile=NULL, subset=NULL, gsnorm=FALSE, mincls, maxcls, conds, reps, verbose=FALSE, eps=1.0e-5, del0=0.01)
e |
data frame containing the expression level for each gene on each chip. |
se |
data frame containing the standard deviation of gene expression levels. |
efile |
character, the name of the file which contains gene expression measurements. |
sefile |
character, the name of the file which contains the standard deviation of gene expression measurements. |
subset |
vector specifying the row number of genes which are clustered on. |
gsnorm |
logical specifying whether do global scaling normalisation or not. |
mincls |
integer, the minimum number of clusters. |
maxcls |
integer, the maximum number of clusters. |
conds |
integer, the number of conditions. |
reps |
vector, specifying which condition each column of the input data matrix belongs to. |
verbose |
logical value. If 'TRUE' messages about the progress of the function is printed. |
eps |
numeric, optimisation parameter. |
del0 |
numeric, optimisation parameter. |
The input data is specified either by e and se, or by efile and sefile.
The result is a list with components
cluster: vector, containing the membership of clusters for each gene; centers: matrix, the center of each cluster; centersigs: matrix, the center variance of each cluster; likelipergene: matrix, the likelihood of belonging to each cluster for each gene; optK: numeric, the optimal number of clusters. optF: numeric, the maximised value of target function.
Xuejun Liu
Liu,X. and Rattray,M. (2009) Including probe-level measurement error in robust mixture clustering of replicated microarray gene expression, Statistical Application in Genetics and Molecular Biology, 9(1), Article 42.
Liu,X., Lin,K.K., Andersen,B., and Rattray,M. (2007) Propagating probe-level uncertainty in model-based gene expression clustering, BMC Bioinformatics, 8:98.
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable probabilistic model for Affymetrix probe-level analysis across multiple chips, Bioinformatics, 21(18):3637-3644.
Related method mmgmos and pumaclust
data(Clustii.exampleE) data(Clustii.exampleStd) r<-vector(mode="integer",0) for (i in c(1:20)) for (j in c(1:4)) r<-c(r,i) cl<-pumaClustii(Clustii.exampleE,Clustii.exampleStd,mincls=6,maxcls=6,conds=20,reps=r,eps=1e-3)data(Clustii.exampleE) data(Clustii.exampleStd) r<-vector(mode="integer",0) for (i in c(1:20)) for (j in c(1:4)) r<-c(r,i) cl<-pumaClustii(Clustii.exampleE,Clustii.exampleStd,mincls=6,maxcls=6,conds=20,reps=r,eps=1e-3)
This function calculates the combined (from replicates) signal for each condition using Bayesian models. The inputs are gene expression levels and the probe-level standard deviations associated with expression measurements for each gene on each chip. The outputs include gene expression levels and standard deviation for each condition.
pumaComb( eset , design.matrix=NULL , method="em" , numOfChunks=1000 , save_r=FALSE , cl=NULL , parallelCompute=FALSE )pumaComb( eset , design.matrix=NULL , method="em" , numOfChunks=1000 , save_r=FALSE , cl=NULL , parallelCompute=FALSE )
eset |
An object of class |
design.matrix |
A design matrix. |
method |
Method "map" uses MAP of a hierarchical Bayesion model with Gamma prior on the between-replicate variance (Gelman et.al. p.285) and shares the same variance across conditions. This method is fast and suitable for the case where there are many conditions. Method "em" uses variational inference of the same hierarchical Bayesian model as in method "map" but with conjugate prior on between-replicate variance and shares the variance across conditions. This is generaly much slower than "map", but is recommended where there are few conditions (as is usually the case). |
numOfChunks |
An integer defining how many chunks the data is divided into before processing. There is generally no need to change the default value. |
save_r |
Will save an internal variable |
cl |
A "cluster" object. See
|
parallelCompute |
Boolean identifying whether processing in parallel should occur. |
It is generally recommended that data is normalised prior to using this function. Note that the default behaviour of mmgmos is to normalise data so this shouldn't generally be an issue. See the function pumaNormalize for more details on normalisation.
The result is an ExpressionSet object.
Xuejun Liu, Marta Milo, Neil D. Lawrence, Magnus Rattray
Gelman,A., Carlin,J.B., Stern,H.S., Rubin,D.B., Bayesian data analysis. London: Chapman & Hall; 1995.
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2006) Probe-level variances improve accuracy in detecting differential gene expression, Bioinformatics, 22:2107-2113.
Related methods pumaNormalize, bcomb, mmgmos and pumaDE
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line shows that eset_mmgmos has 4 arrays, each of which is a different # condition (the experimental design is a 2x2 factorial, with both liver and # scanner factors) pData(eset_mmgmos) # Next line shows expression levels of first 3 probe sets exprs(eset_mmgmos)[1:3,] # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) # We can see that the resulting ExpressionSet object has just two conditions # and 1 expression level for each condition pData(eset_comb) exprs(eset_comb)[1:3,]# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line shows that eset_mmgmos has 4 arrays, each of which is a different # condition (the experimental design is a 2x2 factorial, with both liver and # scanner factors) pData(eset_mmgmos) # Next line shows expression levels of first 3 probe sets exprs(eset_mmgmos)[1:3,] # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) # We can see that the resulting ExpressionSet object has just two conditions # and 1 expression level for each condition pData(eset_comb) exprs(eset_comb)[1:3,]
This function calculates the combined (from replicates) signal for each condition using Bayesian models, which are added a hidden variable to represent the true expression for each gene on each chips. The inputs are gene expression levels and the probe-level standard deviations associated with expression measurements for each gene on each chip. The outputs include gene expression levels and standard deviation for each condition.
pumaCombImproved( eset , design.matrix=NULL , numOfChunks=1000 , maxOfIterations=200 , save_r=FALSE , cl=NULL , parallelCompute=FALSE )pumaCombImproved( eset , design.matrix=NULL , numOfChunks=1000 , maxOfIterations=200 , save_r=FALSE , cl=NULL , parallelCompute=FALSE )
eset |
An object of class |
design.matrix |
A design matrix. |
numOfChunks |
An integer defining how many chunks the data is divided into before processing. There is generally no need to change the default value. |
maxOfIterations |
The maximum number of iterations controls the convergence. |
save_r |
Will save an internal variable |
cl |
A "cluster" object. See
|
parallelCompute |
Boolean identifying whether processing in parallel should occur. |
It is generally recommended that data is normalised prior to using this function. Note that the default behaviour of mmgmos is to normalise data so this shouldn't generally be an issue. See the function pumaNormalize for more details on normalisation.
The maxOfIterations is used to control the maximum number of the iterations in the EM algorithm. You can change the number of maxOfIterations, but the best value of the maxOfIterations is from 200 to 1000, and should be set 200 at least. The default value is 200.
The result is an ExpressionSet object.
Li Zhang, Xuejun Liu
Gelman,A., Carlin,J.B., Stern,H.S., Rubin,D.B., Bayesian data analysis. London: Chapman & Hall; 1995.
Zhang,L. and Liu,X. (2009) An improved probabilistic model for finding differential gene expression, technical report available request. the 2nd BMEI 17-19 oct. 2009. Tianjin. China.
Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2006) Probe-level variances improve accuracy in detecting differential gene expression, Bioinformatics, 22(17):2107-13.
Related methods pumaNormalize, hcomb, mmgmos and pumaDE
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line shows that eset_mmgmos has 4 arrays, each of which is a different # condition (the experimental design is a 2x2 factorial, with both liver and # scanner factors) pData(eset_mmgmos) # Next line shows expression levels of first 3 probe sets exprs(eset_mmgmos)[1:3,] # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_combimproved <- pumaCombImproved(eset_mmgmos_100) # We can see that the resulting ExpressionSet object has just two conditions # and 1 expression level for each condition pData(eset_combimproved) exprs(eset_combimproved)[1:3,]# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line shows that eset_mmgmos has 4 arrays, each of which is a different # condition (the experimental design is a 2x2 factorial, with both liver and # scanner factors) pData(eset_mmgmos) # Next line shows expression levels of first 3 probe sets exprs(eset_mmgmos)[1:3,] # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_combimproved <- pumaCombImproved(eset_mmgmos_100) # We can see that the resulting ExpressionSet object has just two conditions # and 1 expression level for each condition pData(eset_combimproved) exprs(eset_combimproved)[1:3,]
The function generates lists of genes ranked by probability of differential expression (DE). This uses the PPLR method.
pumaDE( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )pumaDE( eset , design.matrix = createDesignMatrix(eset) , contrast.matrix = createContrastMatrix(eset) )
eset |
An object of class |
design.matrix |
A design matrix |
contrast.matrix |
A contrast matrix |
A separate list of genes will be created for each contrast of interest.
Note that this class returns a DEResult-class object. This object contains information on both the PPLR statistic values (which should generally be used to rank genes for differential expression), as well as fold change values (which are generally not recommended for ranking genes, but which might be useful, for example, to use as a filter). To understand more about the object returned see DEResult-class, noting that when created a DEResult object with the pumaDE function, the statistic method should be used to return PPLR values. Also note that the pLikeValues method can be used on the returned object to create values which can more readily be compared with p-values returned by other methods such as variants of t-tests (limma, etc.).
While it is possible to run this function on data from individual arrays, it is generally recommended that this function is run on the output of the function pumaComb (which combines information from replicates).
An object of class DEResult-class.
Richard D. Pearson
Related methods calculateLimma, calculateFC, calculateTtest, pumaComb, pumaCombImproved, mmgmos, pplr, createDesignMatrix and createContrastMatrix
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line shows that eset_mmgmos has 4 arrays, each of which is a different # condition (the experimental design is a 2x2 factorial, with both liver and # scanner factors) pData(eset_mmgmos) # Next line shows expression levels of first 3 probe sets exprs(eset_mmgmos)[1:3,] # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) eset_combimproved <- pumaCombImproved(eset_mmgmos_100) pumaDEResults <- pumaDE(eset_comb) pumaDEResults_improved <- pumaDE(eset_combimproved) topGeneIDs(pumaDEResults,6) # Gives probeset identifiers topGeneIDs(pumaDEResults_improved,6) topGenes(pumaDEResults,6) # Gives row numbers topGenes(pumaDEResults_improved,6) statistic(pumaDEResults)[topGenes(pumaDEResults,6),] # PPLR scores of top six genes statistic(pumaDEResults_improved)[topGenes(pumaDEResults_improved,6),] FC(pumaDEResults)[topGenes(pumaDEResults,6),] # Fold-change of top six genes FC(pumaDEResults_improved)[topGenes(pumaDEResults_improved,6),]# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) # Next line shows that eset_mmgmos has 4 arrays, each of which is a different # condition (the experimental design is a 2x2 factorial, with both liver and # scanner factors) pData(eset_mmgmos) # Next line shows expression levels of first 3 probe sets exprs(eset_mmgmos)[1:3,] # Next line used so eset_mmgmos only has information about the liver factor # The scanner factor will thus be ignored, and the two arrays of each level # of the liver factor will be treated as replicates pData(eset_mmgmos) <- pData(eset_mmgmos)[,1,drop=FALSE] # To save time we'll just use 100 probe sets for the example eset_mmgmos_100 <- eset_mmgmos[1:100,] eset_comb <- pumaComb(eset_mmgmos_100) eset_combimproved <- pumaCombImproved(eset_mmgmos_100) pumaDEResults <- pumaDE(eset_comb) pumaDEResults_improved <- pumaDE(eset_combimproved) topGeneIDs(pumaDEResults,6) # Gives probeset identifiers topGeneIDs(pumaDEResults_improved,6) topGenes(pumaDEResults,6) # Gives row numbers topGenes(pumaDEResults_improved,6) statistic(pumaDEResults)[topGenes(pumaDEResults,6),] # PPLR scores of top six genes statistic(pumaDEResults_improved)[topGenes(pumaDEResults_improved,6),] FC(pumaDEResults)[topGenes(pumaDEResults,6),] # Fold-change of top six genes FC(pumaDEResults_improved)[topGenes(pumaDEResults_improved,6),]
Returns the output from pumaDE as an unsorted matrix (i.e. sorted according to the original sorting in the ExpressionSet)
pumaDEUnsorted(pp)pumaDEUnsorted(pp)
pp |
A list as output by |
A matrix of PPLR values
Richard D. Pearson
Related method pumaDE
Full analysis including pumaPCA and mmgmos/pumaDE vs rma/limma comparison
pumaFull ( ExpressionFeatureSet = NULL , data_dir = getwd() , load_ExpressionFeatureSet = FALSE , calculate_eset = TRUE , calculate_pumaPCAs = TRUE , calculate_bcomb = TRUE , mmgmosComparisons = FALSE )pumaFull ( ExpressionFeatureSet = NULL , data_dir = getwd() , load_ExpressionFeatureSet = FALSE , calculate_eset = TRUE , calculate_pumaPCAs = TRUE , calculate_bcomb = TRUE , mmgmosComparisons = FALSE )
ExpressionFeatureSet |
An object of class |
data_dir |
A character string specifying where data files are stored. |
load_ExpressionFeatureSet |
Boolean. Load a pre-existing ExpressionFeatureSet object? Note that this has to be named "ExpressionFeatureSet.rda" and be in the |
calculate_eset |
Boolean. Calculate ExpressionSet from |
calculate_pumaPCAs |
Boolean. Calculate pumaPCA from |
calculate_bcomb |
Boolean. Calculate pumaComb from |
mmgmosComparisons |
Boolean. If TRUE, will compare mmgmos with default settings, with mmgmos used with background correction. |
No return values. Various objects are saved as .rda files during the execution of this function, and various PDF files are created.
Richard D. Pearson
Related methods pumaDE, createDesignMatrix and createContrastMatrix
## Code commented out to ensure checks run quickly # if (require(pumadata)) data(oligo.estrogen) # pumaFull(oligo.estrogen)## Code commented out to ensure checks run quickly # if (require(pumadata)) data(oligo.estrogen) # pumaFull(oligo.estrogen)
This is used to apply a scaling normalization to set of arrays. This normalization can be at the array scale (thus giving all arrays the same mean or median), or at the probeset scale (thus giving all probesets the same mean or median).
It is generally recommended that the default option (median array scaling) is used after running mmgmos and before running pumaComb and/or pumaDE. There are however, situations where this might not be the recommended, for example in time series experiments where it is expected than there will be general up-regulation or down-regulation in overall gene expression levels between time points.
pumaNormalize( eset , arrayScale = c("median", "none", "mean", "meanlog") , probesetScale = c("none", "mean", "median") , probesetNormalisation = NULL , replicates = list(1:dim(exprs(eset))[2]) )pumaNormalize( eset , arrayScale = c("median", "none", "mean", "meanlog") , probesetScale = c("none", "mean", "median") , probesetNormalisation = NULL , replicates = list(1:dim(exprs(eset))[2]) )
eset |
An object of class |
arrayScale |
A method of scale normalisation at the array level. |
probesetScale |
A method of scale normalisation at the probe set level. |
probesetNormalisation |
If not NULL normalises the expression levels to have zero mean and adjusts the variance of the gene expression according to the zero-centered normalisation. |
replicates |
List of integer vectors indicating which arrays are replicates. |
An object of class ExpressionSet holding the normalised data.
Richard D. Pearson
Methods mmgmos, pumaComb and pumaDE
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) apply(exprs(eset_mmgmos),2,median) eset_mmgmos_normd <- pumaNormalize(eset_mmgmos) apply(exprs(eset_mmgmos_normd),2,median)# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) apply(exprs(eset_mmgmos),2,median) eset_mmgmos_normd <- pumaNormalize(eset_mmgmos) apply(exprs(eset_mmgmos_normd),2,median)
This function carries out principal components analysis (PCA), taking into account not only the expression levels of genes, but also the variability in these expression levels.
The various other pumaPCA... functions are called during the execution of pumaPCA
pumaPCA( eset , latentDim = if(dim(exprs(eset))[2] <= 3) dim(exprs(eset))[[2]]-1 else 3 , sampleSize = if(dim(exprs(eset))[1] <= 1000) dim(exprs(eset))[[1]] else 1000 ## Set to integer or FALSE for all , initPCA = TRUE ## Initialise parameters with PCA , randomOrder = FALSE ## Update parameters in random order , optimMethod = "BFGS" ## ?optim for details of methods , stoppingCriterion = "deltaW"## can also be "deltaL" , tol = 1e-3 ## Stop when delta update < this , stepChecks = FALSE ## Check likelihood after each update? , iterationNumbers = TRUE ## Show iteration numbers? , showUpdates = FALSE ## Show values after each update? , showTimings = FALSE ## Show timings after each update? , showPlot = FALSE ## Show projection plot after each update? , maxIters = 500 ## Number of EM iterations. , transposeData = FALSE ## Transpose eset matrices? , returnExpectations = FALSE , returnData = FALSE , returnFeedback = FALSE , pumaNormalize = TRUE )pumaPCA( eset , latentDim = if(dim(exprs(eset))[2] <= 3) dim(exprs(eset))[[2]]-1 else 3 , sampleSize = if(dim(exprs(eset))[1] <= 1000) dim(exprs(eset))[[1]] else 1000 ## Set to integer or FALSE for all , initPCA = TRUE ## Initialise parameters with PCA , randomOrder = FALSE ## Update parameters in random order , optimMethod = "BFGS" ## ?optim for details of methods , stoppingCriterion = "deltaW"## can also be "deltaL" , tol = 1e-3 ## Stop when delta update < this , stepChecks = FALSE ## Check likelihood after each update? , iterationNumbers = TRUE ## Show iteration numbers? , showUpdates = FALSE ## Show values after each update? , showTimings = FALSE ## Show timings after each update? , showPlot = FALSE ## Show projection plot after each update? , maxIters = 500 ## Number of EM iterations. , transposeData = FALSE ## Transpose eset matrices? , returnExpectations = FALSE , returnData = FALSE , returnFeedback = FALSE , pumaNormalize = TRUE )
eset |
An object of class |
latentDim |
An integer specifying the number of latent dimensions (kind of like the number of principal components). |
sampleSize |
An integer specifying the number of probesets to sample (default is 1000), or FALSE, meaning use all the data. |
initPCA |
A boolean indicating whether to initialise using standard PCA (the default, and generally quicker and recommended). |
randomOrder |
A boolean indicating whether the parameters should be updated in a random order (this is generally not recommended, and the default is FALSE). |
optimMethod |
See ?optim for details of methods. |
stoppingCriterion |
If set to "deltaW" will stop when W changes by less than |
tol |
Tolerance value for |
stepChecks |
Boolean. Check likelihood after each update? |
iterationNumbers |
Boolean. Show iteration numbers? |
showUpdates |
Boolean. Show values after each update? |
showTimings |
Boolean. Show timings after each update? |
showPlot |
Boolean. Show projection plot after each update? |
maxIters |
Integer. Maximum number of EM iterations. |
transposeData |
Boolean. Transpose eset matrices? |
returnExpectations |
Boolean. Return expectation values? |
returnData |
Boolean. Return expectation data? |
returnFeedback |
Boolean. Return feedback on progress of optimisation? |
pumaNormalize |
Boolean. Normalise data prior to running algorithm (recommended)? |
An object of class pumaPCARes
Richard D. Pearson
Related methods pumaDE, createDesignMatrix and createContrastMatrix
# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) pumapca_mmgmos <- pumaPCA(eset_mmgmos) plot(pumapca_mmgmos)# Next 4 lines commented out to save time in package checks, and saved version used # if (require(affydata)) { # data(Dilution) # eset_mmgmos <- mmgmos(Dilution) # } data(eset_mmgmos) pumapca_mmgmos <- pumaPCA(eset_mmgmos) plot(pumapca_mmgmos)
This is a class representation for storing a set of expectations from a pumaPCA model. It is an internal representation and shouldn't normally be instantiated.
Objects can be created by calls of the form new("pumaPCAExpectations", ...).
x:Object of class "matrix" representing x
xxT:Object of class "array" representing xxT
logDetCov:Object of class "numeric" representing logDetCov
This class has no methods defined
Richard D. Pearson
Related method pumaPCA and related class pumaPCARes.
This is a class representation for storing a pumaPCA model. It is an internal representation and shouldn't normally be instantiated.
Objects can be created by calls of the form new("pumaPCAModel", ...).
sigma:Object of class "numeric" representing sigma
m:Object of class "matrix" representing m
Cinv:Object of class "matrix" representing Cinv
W:Object of class "matrix" representing W
mu:Object of class "matrix" representing mu
This class has no methods defined
Richard D. Pearson
Related method pumaPCA and related class pumaPCARes.
This is a class representation for storing the outputs of the pumaPCA function. Objects of this class should usually only be created through the pumaPCA function.
Objects can be created by calls of the form new("pumaPCARes", ...).
model:Object of class "pumaPCAModel" representing the model parameters
expectations:Object of class "pumaPCAExpectations" representing the model expectations
varY:Object of class "matrix" representing the variance in the expression levels
Y:Object of class "matrix" representing the expression levels
phenoData:Object of class "AnnotatedDataFrame" representing the phenotype information
timeToCompute:Object of class "numeric" representing the time it took pumaPCA to run
numberOfIterations:Object of class "numeric" representing the number of iterations it took pumaPCA to converge
likelihoodHistory:Object of class "list" representing the history of likelihood values while pumaPCA was running
timingHistory:Object of class "list" representing the history of how long each iteration took while pumaPCA was running
modelHistory:Object of class "list" representing the history of how the model was changing while pumaPCA was running
exitReason:Object of class "character" representing the reason pumaPCA halted. Can take the values "Update of Likelihood less than tolerance x", "Update of W less than tolerance x", "Iterations exceeded", "User interrupt", "unknown exit reason"
signature(x="pumaPCARes-class"): plots two principal components on a scatter plot.
signature(x = "pumaPCARes-class"): writes the principal components for each array to a file. It takes the same arguments as write.table. The argument "file" does not need to set any extension. The file name and extension "csv" will be added automatically. The default file name is "tmp".
Richard D. Pearson
Related method pumaPCA and related class pumaPCARes.
This is really an internal function used to remove uninformative factors from the phenotype data. Uninformative factors here are defined as those which have the same value for all arrays in the ExpressionSet.
removeUninformativeFactors(eset)removeUninformativeFactors(eset)
eset |
An object of class |
An ExpressionSet object with the same data as the input, except for a new phenoData slot.
Richard D. Pearson
Related methods createDesignMatrix and createContrastMatrix
eset_test <- new("ExpressionSet", exprs=matrix(rnorm(400,8,2),100,4)) pData(eset_test) <- data.frame("informativeFactor"=c("A", "A", "B", "B"), "uninformativeFactor"=c("X","X","X","X")) eset_test2 <- removeUninformativeFactors(eset_test) pData(eset_test) pData(eset_test2)eset_test <- new("ExpressionSet", exprs=matrix(rnorm(400,8,2),100,4)) pData(eset_test) <- data.frame("informativeFactor"=c("A", "A", "B", "B"), "uninformativeFactor"=c("X","X","X","X")) eset_test2 <- removeUninformativeFactors(eset_test) pData(eset_test) pData(eset_test2)