@@ -48,8 +48,9 @@ information used by CRLMM is stored in the \Rpackage{human370v1cCrlmm} package.

 The required packages can be installed in the usual way using the \Rfunction{biocLite} function.

 <<echo=TRUE, results=hide, eval=FALSE>>=

-source("http://www.bioconductor.org/biocLite.R")

-biocLite(c("crlmm", "hapmap370k", "human370v1cCrlmm"))

+if (!requireNamespace("BiocManager", quietly=TRUE))

+    install.packages("BiocManager")

+BiocManager::install(c("crlmm", "hapmap370k", "human370v1cCrlmm"))

 @

 \section{Reading in data}

@@ -62,8 +62,8 @@ options(width=70)

 @

 <<read, results=hide, eval=TRUE>>=

-library(Biobase)

 library(crlmm)

+library(ff)

 library(hapmap370k)

 data.dir = system.file("idatFiles", package="hapmap370k")

@@ -73,73 +73,31 @@ samples = read.csv(file.path(data.dir, "samples370k.csv"), as.is=TRUE)

 samples[1:5,]

 @

-<<read2, results=hide, cache=TRUE>>=

-# Read in .idats using sampleSheet information

-RG = readIdatFiles(samples, path=data.dir,

-arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),saveDate=TRUE)

-@

-

-Reading in this data takes approximately 100 seconds and peak memory usage

-was 0.8 GB of RAM on our linux system.

-If memory is limiting, load the \Rpackage{ff} package and run the same command.

-When this package is available, the objects are stored using disk rather then RAM.

-The \Robject{RG} object is an \Rclass{NChannelSet} which stores the

-Red and Green intensities, the number of beads and standard errors for

-each bead-type.

-The scanning date of each array is stored in \Robject{protocolData}.

-

-<<explore>>=

-class(RG)

-dim(RG)

-slotNames(RG)

-channelNames(RG)

-exprs(channel(RG, "R"))[1:5,1:5]

-exprs(channel(RG, "G"))[1:5,1:5]

-pd = pData(RG)

-pd[1:5,]

-

-scandatetime = strptime(protocolData(RG)[["ScanDate"]], "%m/%d/%Y %H:%M:%S %p")

-datescanned = substr(scandatetime, 1, 10)

-scanbatch =  factor(datescanned)

-levels(scanbatch) = 1:16

-scanbatch = as.numeric(scanbatch)

-@

-

-If GenCall output is available instead of idat files, the function \Rfunction{readGenCallOutput} can be

-used to read in the data.

-This function assumes the GenCall output is formatted to have samples listed one below the other,

-and that the columns 'X Raw' and 'Y Raw' are available in the file.

-The resulting \Robject{NChannelSet} from this function can be used as input to \Rfunction{crlmmIllumina} via the \Robject{XY} argument (instead of the usual \Rfunction{RG} argument used when the data has been read in from idat files).

-

-Plots of the summarised data can be easily generated to check for arrays with poor signal.

-

-<<boxplots, fig=TRUE, width=8, height=8>>=

-par(mfrow=c(2,1), mai=c(0.4,0.4,0.4,0.1), oma=c(1,1,0,0))

-boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", names=1:40,

-main="Red channel",outline=FALSE,las=2)

-boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", names=1:40,

-main="Green channel",outline=FALSE,las=2)

-mtext(expression(log[2](intensity)), side=2, outer=TRUE)

-mtext("Array", side=1, outer=TRUE)

-@

-

 \section{Genotyping}

 Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing followed by genotyping using the CRLMM algorithm.

+It reads in data from idat files and stores results in a \Rclass{CNSet} object.

 <<genotype, results=hide, cache=TRUE>>=

-crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", returnParams=TRUE)

-@

+crlmmResult = crlmmIllumina(samples, path=data.dir,

+				arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),

+				saveDate=TRUE, cdfName="human370v1c", verbose=FALSE)

-This analysis took 3 minutes to complete and peak memory usage was 1.9 GB on our system.

-The output stored in \Robject{crlmmResult} is a \Rclass{SnpSet} object.

-<<explore2>>=

 class(crlmmResult)

 dim(crlmmResult)

 slotNames(crlmmResult)

 calls(crlmmResult)[1:10, 1:5]

+i2p(confs(crlmmResult)[1:10,1:5])

 @

+If GenCall output is available instead of idat files, the function \Rfunction{readGenCallOutput} can be

+used to read in the data.

+This function assumes the GenCall output is formatted to have samples listed one below the other,

+and that the columns 'X Raw' and 'Y Raw' are available in the file.

+The resulting \Robject{NChannelSet} from this function can be used as input to 

+\Rfunction{crlmmIllumina} (or equivalently \code{genotype.Illumina}) via the \Robject{XY} argument 

+(instead of the usual \Rfunction{RG} argument used when the data has been read in from idat files).

+

 Plotting the {\it SNR} reveals no obvious batch effects in this data set (different symbols are used for

 arrays scanned on different days).

@@ -148,15 +106,16 @@ plot(crlmmResult[["SNR"]], pch=scanbatch, xlab="Array", ylab="SNR",

      main="Signal-to-noise ratio per array",las=2)

 @

-An all-in-one function named \Rfunction{crlmmIlluminaV2} that combines

-reading of idat files with genotyping is also available.

-

-<<readandgenotypeinone, results=hide, cache=TRUE>>=

-crlmmResult2 <- crlmmIlluminaV2(samples, path=data.dir,

-				arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),

-				saveDate=TRUE, cdfName="human370v1c", returnParams=TRUE)

+<<plotsamples, fig=TRUE, width=8, height=8>>=

+par(mfrow=c(2,2))

+plotSamples(crlmmResult, col=1:4)

 @

+<<plotsnps, fig=TRUE, width=8, height=8>>=

+par(mfrow=c(2,2))

+plotSNPs(crlmmResult, row=1:4)

+@ 

+

 \section{System information}

 This analysis was carried out on a linux machine with 32GB of RAM

@@ -29,8 +29,8 @@

 \textwidth=6.2in

-\bibliographystyle{plainnat} 

- 

+\bibliographystyle{plainnat}

+

 \begin{document}

 %\setkeys{Gin}{width=0.55\textwidth}

@@ -40,26 +40,26 @@

 \section{Getting started}

-In this user guide we read in and genotype data from 40 HapMap samples 

+In this user guide we read in and genotype data from 40 HapMap samples

 which have been analyzed using Illumina's 370k Duo BeadChips.

-This data is available in the \Rpackage{hapmap370k} package.  

-Additional chip-specific model parameters and basic SNP annotation 

+This data is available in the \Rpackage{hapmap370k} package.

+Additional chip-specific model parameters and basic SNP annotation

 information used by CRLMM is stored in the \Rpackage{human370v1cCrlmm} package.

 The required packages can be installed in the usual way using the \Rfunction{biocLite} function.

 <<echo=TRUE, results=hide, eval=FALSE>>=

 source("http://www.bioconductor.org/biocLite.R")

 biocLite(c("crlmm", "hapmap370k", "human370v1cCrlmm"))

-@ 

+@

-\section{Reading in data} 

-The function \Rfunction{readIdatFiles} extracts the Red and Green intensities 

+\section{Reading in data}

+The function \Rfunction{readIdatFiles} extracts the Red and Green intensities

 from the binary {\tt idat} files output by Illumina's scanning device.

 The file {\tt samples370k.csv} contains information about each sample.

 <<echo=FALSE, results=hide, eval=TRUE>>=

 options(width=70)

-@ 

+@

 <<read, results=hide, eval=TRUE>>=

 library(Biobase)

@@ -71,21 +71,21 @@ data.dir = system.file("idatFiles", package="hapmap370k")

 # Read in sample annotation info

 samples = read.csv(file.path(data.dir, "samples370k.csv"), as.is=TRUE)

 samples[1:5,]

-@ 

+@

 <<read2, results=hide, cache=TRUE>>=

 # Read in .idats using sampleSheet information

-RG = readIdatFiles(samples, path=data.dir, 

+RG = readIdatFiles(samples, path=data.dir,

 arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),saveDate=TRUE)

 @

-Reading in this data takes approximately 100 seconds and peak memory usage 

+Reading in this data takes approximately 100 seconds and peak memory usage

 was 0.8 GB of RAM on our linux system.

 If memory is limiting, load the \Rpackage{ff} package and run the same command.

 When this package is available, the objects are stored using disk rather then RAM.

-The \Robject{RG} object is an \Rclass{NChannelSet} which stores the 

-Red and Green intensities, the number of beads and standard errors for 

-each bead-type.  

+The \Robject{RG} object is an \Rclass{NChannelSet} which stores the

+Red and Green intensities, the number of beads and standard errors for

+each bead-type.

 The scanning date of each array is stored in \Robject{protocolData}.

 <<explore>>=

@@ -105,19 +105,19 @@ levels(scanbatch) = 1:16

 scanbatch = as.numeric(scanbatch)

 @

-If GenCall output is available instead of idat files, the function \Rfunction{readGenCallOutput} can be 

-used to read in the data.  

+If GenCall output is available instead of idat files, the function \Rfunction{readGenCallOutput} can be

+used to read in the data.

 This function assumes the GenCall output is formatted to have samples listed one below the other,

-and that the columns 'X Raw' and 'Y Raw' are available in the file.  

+and that the columns 'X Raw' and 'Y Raw' are available in the file.

 The resulting \Robject{NChannelSet} from this function can be used as input to \Rfunction{crlmmIllumina} via the \Robject{XY} argument (instead of the usual \Rfunction{RG} argument used when the data has been read in from idat files).

 Plots of the summarised data can be easily generated to check for arrays with poor signal.

 <<boxplots, fig=TRUE, width=8, height=8>>=

 par(mfrow=c(2,1), mai=c(0.4,0.4,0.4,0.1), oma=c(1,1,0,0))

-boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", names=1:40, 

+boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", names=1:40,

 main="Red channel",outline=FALSE,las=2)

-boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", names=1:40, 

+boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", names=1:40,

 main="Green channel",outline=FALSE,las=2)

 mtext(expression(log[2](intensity)), side=2, outer=TRUE)

 mtext("Array", side=1, outer=TRUE)

@@ -129,36 +129,37 @@ Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing

 <<genotype, results=hide, cache=TRUE>>=

 crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", returnParams=TRUE)

-@ 

+@

 This analysis took 3 minutes to complete and peak memory usage was 1.9 GB on our system.

-The output stored in \Robject{crlmmResult} is a \Rclass{SnpSet} object.                                                                                                                                                         

+The output stored in \Robject{crlmmResult} is a \Rclass{SnpSet} object.

 <<explore2>>=

 class(crlmmResult)

 dim(crlmmResult)

 slotNames(crlmmResult)

 calls(crlmmResult)[1:10, 1:5]

-@ 

+@

-Plotting the {\it SNR} reveals no obvious batch effects in this data set (different symbols are used for 

+Plotting the {\it SNR} reveals no obvious batch effects in this data set (different symbols are used for

 arrays scanned on different days).

 <<snr,  fig=TRUE, width=8, height=6>>=

-plot(crlmmResult[["SNR"]], pch=scanbatch, xlab="Array", ylab="SNR", 

-main="Signal-to-noise ratio per array",las=2)

+plot(crlmmResult[["SNR"]], pch=scanbatch, xlab="Array", ylab="SNR",

+     main="Signal-to-noise ratio per array",las=2)

 @

-An all-in-one function named \Rfunction{crlmmIlluminaV2} that combines reading of idat files with genotyping is also available.

+An all-in-one function named \Rfunction{crlmmIlluminaV2} that combines

+reading of idat files with genotyping is also available.

 <<readandgenotypeinone, results=hide, cache=TRUE>>=

-crlmmResult2 = crlmmIlluminaV2(samples, path=data.dir, 

-arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),

-saveDate=TRUE, cdfName="human370v1c", returnParams=TRUE)

-@ 

+crlmmResult2 <- crlmmIlluminaV2(samples, path=data.dir,

+				arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),

+				saveDate=TRUE, cdfName="human370v1c", returnParams=TRUE)

+@

 \section{System information}

-This analysis was carried out on a linux machine with 32GB of RAM 

+This analysis was carried out on a linux machine with 32GB of RAM

 using the following packages:

 <<session>>=

@@ -47,7 +47,7 @@ Additional chip-specific model parameters and basic SNP annotation

 information used by CRLMM is stored in the \Rpackage{human370v1cCrlmm} package.

 The required packages can be installed in the usual way using the \Rfunction{biocLite} function.

-<<<echo=TRUE, results=hide, eval=FALSE>>=

+<<echo=TRUE, results=hide, eval=FALSE>>=

 source("http://www.bioconductor.org/biocLite.R")

 biocLite(c("crlmm", "hapmap370k", "human370v1cCrlmm"))

 @ 

@@ -57,8 +57,8 @@ The function \Rfunction{readIdatFiles} extracts the Red and Green intensities

 from the binary {\tt idat} files output by Illumina's scanning device.

 The file {\tt samples370k.csv} contains information about each sample.

-<<<echo=FALSE, results=hide, eval=TRUE>>=

-options(width=50)

+<<echo=FALSE, results=hide, eval=TRUE>>=

+options(width=70)

 @ 

 <<read, results=hide, eval=TRUE>>=

@@ -75,7 +75,8 @@ samples[1:5,]

 <<read2, results=hide, cache=TRUE>>=

 # Read in .idats using sampleSheet information

-RG = readIdatFiles(samples, path=data.dir, arrayInfoColNames=list(barcode=NULL, position="SentrixPosition"), saveDate=TRUE)

+RG = readIdatFiles(samples, path=data.dir, 

+arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),saveDate=TRUE)

 @

 Reading in this data takes approximately 100 seconds and peak memory usage 

@@ -104,13 +105,20 @@ levels(scanbatch) = 1:16

 scanbatch = as.numeric(scanbatch)

 @

-Plots of the summarised data can be easily generated to check for arrays 

-with poor signal.

+If GenCall output is available instead of idat files, the function \Rfunction{readGenCallOutput} can be 

+used to read in the data.  

+This function assumes the GenCall output is formatted to have samples listed one below the other,

+and that the columns 'X Raw' and 'Y Raw' are available in the file.  

+The resulting \Robject{NChannelSet} from this function can be used as input to \Rfunction{crlmmIllumina} via the \Robject{XY} argument (instead of the usual \Rfunction{RG} argument used when the data has been read in from idat files).

+

+Plots of the summarised data can be easily generated to check for arrays with poor signal.

 <<boxplots, fig=TRUE, width=8, height=8>>=

 par(mfrow=c(2,1), mai=c(0.4,0.4,0.4,0.1), oma=c(1,1,0,0))

-boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", names=1:40, main="Red channel", outline=FALSE, las=2)

-boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", names=1:40, main="Green channel", outline=FALSE, las=2)

+boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", names=1:40, 

+main="Red channel",outline=FALSE,las=2)

+boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", names=1:40, 

+main="Green channel",outline=FALSE,las=2)

 mtext(expression(log[2](intensity)), side=2, outer=TRUE)

 mtext("Array", side=1, outer=TRUE)

 @

@@ -120,25 +128,34 @@ mtext("Array", side=1, outer=TRUE)

 Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing followed by genotyping using the CRLMM algorithm.

 <<genotype, results=hide, cache=TRUE>>=

-crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", sns=pData(RG)$ID, returnParams=TRUE)

+crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", returnParams=TRUE)

 @ 

-This analysis took 18 minutes to complete and peak memory usage was 2.5 GB on our system.

+This analysis took 3 minutes to complete and peak memory usage was 1.9 GB on our system.

 The output stored in \Robject{crlmmResult} is a \Rclass{SnpSet} object.                                                                                                                                                         

 <<explore2>>=

-  class(crlmmResult)

-  dim(crlmmResult)

-  slotNames(crlmmResult)

-  calls(crlmmResult)[1:10, 1:5]

+class(crlmmResult)

+dim(crlmmResult)

+slotNames(crlmmResult)

+calls(crlmmResult)[1:10, 1:5]

 @ 

 Plotting the {\it SNR} reveals no obvious batch effects in this data set (different symbols are used for 

 arrays scanned on different days).

 <<snr,  fig=TRUE, width=8, height=6>>=

-plot(crlmmResult[["SNR"]], pch=scanbatch, xlab="Array", ylab="SNR", main="Signal-to-noise ratio per array", las=2)

+plot(crlmmResult[["SNR"]], pch=scanbatch, xlab="Array", ylab="SNR", 

+main="Signal-to-noise ratio per array",las=2)

 @

+An all-in-one function named \Rfunction{crlmmIlluminaV2} that combines reading of idat files with genotyping is also available.

+

+<<readandgenotypeinone, results=hide, cache=TRUE>>=

+crlmmResult2 = crlmmIlluminaV2(samples, path=data.dir, 

+arrayInfoColNames=list(barcode=NULL,position="SentrixPosition"),

+saveDate=TRUE, cdfName="human370v1c", returnParams=TRUE)

+@ 

+

 \section{System information}

 This analysis was carried out on a linux machine with 32GB of RAM 

@@ -78,8 +78,8 @@ samples[1:5,]

 RG = readIdatFiles(samples, path=data.dir, arrayInfoColNames=list(barcode=NULL, position="SentrixPosition"), saveDate=TRUE)

 @

-Reading in this data takes approximately 90 seconds and peak memory usage 

-was 1.2 GB of RAM on our linux system.

+Reading in this data takes approximately 100 seconds and peak memory usage 

+was 0.8 GB of RAM on our linux system.

 If memory is limiting, load the \Rpackage{ff} package and run the same command.

 When this package is available, the objects are stored using disk rather then RAM.

 The \Robject{RG} object is an \Rclass{NChannelSet} which stores the 

@@ -123,7 +123,7 @@ Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing

 crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", sns=pData(RG)$ID, returnParams=TRUE)

 @ 

-This analysis took 470 seconds to complete and peak memory usage was 3.3 GB on our system.

+This analysis took 18 minutes to complete and peak memory usage was 2.5 GB on our system.

 The output stored in \Robject{crlmmResult} is a \Rclass{SnpSet} object.                                                                                                                                                         

 <<explore2>>=

   class(crlmmResult)

@@ -45,11 +45,11 @@ which have been analyzed using Illumina's 370k Duo BeadChips.

 This data is available in the \Rpackage{hapmap370k} package.  

 Additional chip-specific model parameters and basic SNP annotation 

 information used by CRLMM is stored in the \Rpackage{human370v1cCrlmm} package.

-These packages can be installed in the usual way using the \Rfunction{biocLite} function.

+The required packages can be installed in the usual way using the \Rfunction{biocLite} function.

 <<<echo=TRUE, results=hide, eval=FALSE>>=

 source("http://www.bioconductor.org/biocLite.R")

-biocLite(c("hapmap370k", "human370v1cCrlmm"))

+biocLite(c("crlmm", "hapmap370k", "human370v1cCrlmm"))

 @ 

 \section{Reading in data} 

@@ -80,6 +80,8 @@ RG = readIdatFiles(samples, path=data.dir, arrayInfoColNames=list(barcode=NULL,

 Reading in this data takes approximately 90 seconds and peak memory usage 

 was 1.2 GB of RAM on our linux system.

+If memory is limiting, load the \Rpackage{ff} package and run the same command.

+When this package is available, the objects are stored using disk rather then RAM.

 The \Robject{RG} object is an \Rclass{NChannelSet} which stores the 

 Red and Green intensities, the number of beads and standard errors for 

 each bead-type.  

@@ -44,12 +44,15 @@ In this user guide we read in and genotype data from 40 HapMap samples

 which have been analyzed using Illumina's 370k Duo BeadChips.

 This data is available in the \Rpackage{hapmap370k} package.  

 Additional chip-specific model parameters and basic SNP annotation 

-information used by CRLMM is stored in the \Rpackage{human370v1c} package.

-These can be downloaded from \href{http://rafalab.jhsph.edu/software.html}{http://rafalab.jhsph.edu/software.html}

-and must be installed for the following code to work.

+information used by CRLMM is stored in the \Rpackage{human370v1cCrlmm} package.

+These packages can be installed in the usual way using the \Rfunction{biocLite} function.

-\section{Reading in data}

+<<<echo=TRUE, results=hide, eval=FALSE>>=

+source("http://www.bioconductor.org/biocLite.R")

+biocLite(c("hapmap370k", "human370v1cCrlmm"))

+@ 

+\section{Reading in data} 

 The function \Rfunction{readIdatFiles} extracts the Red and Green intensities 

 from the binary {\tt idat} files output by Illumina's scanning device.

 The file {\tt samples370k.csv} contains information about each sample.

@@ -58,7 +61,7 @@ The file {\tt samples370k.csv} contains information about each sample.

 options(width=50)

 @ 

-<<read>>=

+<<read, results=hide, eval=TRUE>>=

 library(Biobase)

 library(crlmm)

 library(hapmap370k)

@@ -112,8 +115,7 @@ mtext("Array", side=1, outer=TRUE)

 \section{Genotyping}

-Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing followed by genotyping

-using the CRLMM algorithm.

+Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing followed by genotyping using the CRLMM algorithm.

 <<genotype, results=hide, cache=TRUE>>=

 crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", sns=pData(RG)$ID, returnParams=TRUE)

@@ -80,7 +80,7 @@ was 1.2 GB of RAM on our linux system.

 The \Robject{RG} object is an \Rclass{NChannelSet} which stores the 

 Red and Green intensities, the number of beads and standard errors for 

 each bead-type.  

-The scanning date of each array is stored in the \Robject{protocolData} slot.

+The scanning date of each array is stored in \Robject{protocolData}.

 <<explore>>=

 class(RG)

@@ -104,8 +104,8 @@ with poor signal.

 <<boxplots, fig=TRUE, width=8, height=8>>=

 par(mfrow=c(2,1), mai=c(0.4,0.4,0.4,0.1), oma=c(1,1,0,0))

-boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", main="Red channel", outline=FALSE, las=2)

-boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", main="Green channel", outline=FALSE, las=2)

+boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", names=1:40, main="Red channel", outline=FALSE, las=2)

+boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", names=1:40, main="Green channel", outline=FALSE, las=2)

 mtext(expression(log[2](intensity)), side=2, outer=TRUE)

 mtext("Array", side=1, outer=TRUE)

 @

@@ -80,7 +80,7 @@ was 1.2 GB of RAM on our linux system.

 The \Robject{RG} object is an \Rclass{NChannelSet} which stores the 

 Red and Green intensities, the number of beads and standard errors for 

 each bead-type.  

-The scanning date of each array is stored in the \Robject{scanDates} slot.

+The scanning date of each array is stored in the \Robject{protocolData} slot.

 <<explore>>=

 class(RG)

@@ -92,7 +92,7 @@ exprs(channel(RG, "G"))[1:5,1:5]

 pd = pData(RG)

 pd[1:5,]

-scandatetime = strptime(scanDates(RG), "%m/%d/%Y %H:%M:%S %p")

+scandatetime = strptime(protocolData(RG)[["ScanDate"]], "%m/%d/%Y %H:%M:%S %p")

 datescanned = substr(scandatetime, 1, 10)

 scanbatch =  factor(datescanned)

 levels(scanbatch) = 1:16

new file mode 100644

@@ -0,0 +1,147 @@

+%\VignetteIndexEntry{crlmm Vignette - Illumina 370k chip}

+%\VignetteKeywords{genotype, crlmm, Illumina}

+%\VignettePackage{crlmm}

+

+\documentclass[12pt]{article}

+

+\usepackage{amsmath,pstricks}

+\usepackage[authoryear,round]{natbib}

+\usepackage{hyperref}

+\usepackage{Sweave}

+

+\textwidth=6.2in

+\textheight=8.5in

+%\parskip=.3cm

+\oddsidemargin=.1in

+\evensidemargin=.1in

+\headheight=-.3in

+

+\newcommand{\scscst}{\scriptscriptstyle}

+\newcommand{\scst}{\scriptstyle}

+

+

+\newcommand{\Rfunction}[1]{{\texttt{#1}}}

+\newcommand{\Robject}[1]{{\texttt{#1}}}

+\newcommand{\Rpackage}[1]{{\textit{#1}}}

+\newcommand{\Rmethod}[1]{{\texttt{#1}}}

+\newcommand{\Rfunarg}[1]{{\texttt{#1}}}

+\newcommand{\Rclass}[1]{{\textit{#1}}}

+

+\textwidth=6.2in

+

+\bibliographystyle{plainnat} 

+ 

+\begin{document}

+%\setkeys{Gin}{width=0.55\textwidth}

+

+\title{Using \Rpackage{crlmm} to genotype data from Illumina's Infinium BeadChips}

+\author{Matt Ritchie}

+\maketitle

+

+\section{Getting started}

+

+In this user guide we read in and genotype data from 40 HapMap samples 

+which have been analyzed using Illumina's 370k Duo BeadChips.

+This data is available in the \Rpackage{hapmap370k} package.  

+Additional chip-specific model parameters and basic SNP annotation 

+information used by CRLMM is stored in the \Rpackage{human370v1c} package.

+These can be downloaded from \href{http://rafalab.jhsph.edu/software.html}{http://rafalab.jhsph.edu/software.html}

+and must be installed for the following code to work.

+

+\section{Reading in data}

+

+The function \Rfunction{readIdatFiles} extracts the Red and Green intensities 

+from the binary {\tt idat} files output by Illumina's scanning device.

+The file {\tt samples370k.csv} contains information about each sample.

+

+<<<echo=FALSE, results=hide, eval=TRUE>>=

+options(width=50)

+@ 

+

+<<read>>=

+library(Biobase)

+library(crlmm)

+library(hapmap370k)

+

+data.dir = system.file("idatFiles", package="hapmap370k")

+

+# Read in sample annotation info

+samples = read.csv(file.path(data.dir, "samples370k.csv"), as.is=TRUE)

+samples[1:5,]

+@ 

+

+<<read2, results=hide, cache=TRUE>>=

+# Read in .idats using sampleSheet information

+RG = readIdatFiles(samples, path=data.dir, arrayInfoColNames=list(barcode=NULL, position="SentrixPosition"), saveDate=TRUE)

+@

+

+Reading in this data takes approximately 90 seconds and peak memory usage 

+was 1.2 GB of RAM on our linux system.

+The \Robject{RG} object is an \Rclass{NChannelSet} which stores the 

+Red and Green intensities, the number of beads and standard errors for 

+each bead-type.  

+The scanning date of each array is stored in the \Robject{scanDates} slot.

+

+<<explore>>=

+class(RG)

+dim(RG)

+slotNames(RG)

+channelNames(RG)

+exprs(channel(RG, "R"))[1:5,1:5]

+exprs(channel(RG, "G"))[1:5,1:5]

+pd = pData(RG)

+pd[1:5,]

+

+scandatetime = strptime(scanDates(RG), "%m/%d/%Y %H:%M:%S %p")

+datescanned = substr(scandatetime, 1, 10)

+scanbatch =  factor(datescanned)

+levels(scanbatch) = 1:16

+scanbatch = as.numeric(scanbatch)

+@

+

+Plots of the summarised data can be easily generated to check for arrays 

+with poor signal.

+

+<<boxplots, fig=TRUE, width=8, height=8>>=

+par(mfrow=c(2,1), mai=c(0.4,0.4,0.4,0.1), oma=c(1,1,0,0))

+boxplot(log2(exprs(channel(RG, "R"))), xlab="Array", ylab="", main="Red channel", outline=FALSE, las=2)

+boxplot(log2(exprs(channel(RG, "G"))), xlab="Array", ylab="", main="Green channel", outline=FALSE, las=2)

+mtext(expression(log[2](intensity)), side=2, outer=TRUE)

+mtext("Array", side=1, outer=TRUE)

+@

+

+\section{Genotyping}

+

+Next we use the function \Rfunction{crlmmIllumina} which performs preprocessing followed by genotyping

+using the CRLMM algorithm.

+

+<<genotype, results=hide, cache=TRUE>>=

+crlmmResult = crlmmIllumina(RG=RG, cdfName="human370v1c", sns=pData(RG)$ID, returnParams=TRUE)

+@ 

+

+This analysis took 470 seconds to complete and peak memory usage was 3.3 GB on our system.

+The output stored in \Robject{crlmmResult} is a \Rclass{SnpSet} object.                                                                                                                                                         

+<<explore2>>=

+  class(crlmmResult)

+  dim(crlmmResult)

+  slotNames(crlmmResult)

+  calls(crlmmResult)[1:10, 1:5]

+@ 

+

+Plotting the {\it SNR} reveals no obvious batch effects in this data set (different symbols are used for 

+arrays scanned on different days).

+

+<<snr,  fig=TRUE, width=8, height=6>>=

+plot(crlmmResult[["SNR"]], pch=scanbatch, xlab="Array", ylab="SNR", main="Signal-to-noise ratio per array", las=2)

+@

+

+\section{System information}

+

+This analysis was carried out on a linux machine with 32GB of RAM 

+using the following packages:

+

+<<session>>=

+sessionInfo()

+@

+

+\end{document}