#Version 1.5.1(Nov. 17, 2016)
## (1) fixed a bug in quantilePlots function: drop the "..." option
## (2) added "x=", and "y=" in plot(), lines(), and points()
## (3) revised functions 'scatterPlots' and 'boxPlots' so that
## the user can choose not output 'p.adj' in the subtitle
# (4) add input options 'outcomeFlag' and 'fitLineFlag' to 'scatterPlots'
# (5) added Rd files for functions 'scatterPlots', 'boxPlots',
# and 'densityPlots'
#
#Version 1.3.1(Oct. 16, 2016)
# fixed the following bug by replacing vsn package by limma package
#
#* installing the package to build vignettes
# -----------------------------------
#ERROR: dependency ‘vsn’ is not available for package ‘iCheck’
#* removing ‘/tmp/RtmpBEUPpZ/Rinst6e5d1ea3e873/iCheck’
-----------------------------------
ERROR: package installation failed
#Version 1.2.1(Oct. 13, 2016)
# (1) fixed a bug in getPCAFunc function: "prcomp(t(dat2), cor=corFlag)" should be
# prcomp(t(dat2), scale=corFlag)
# (2) added
adding
importFrom("grDevices", "jpeg", "png", "postscript")
importFrom("graphics", "stripchart")
importFrom("utils", "capture.output", "packageDescription",
"read.delim", "write.table")
#
#Version 1.0.1(March 31, 2016)
# (1) added 'Authors@R' field in DESCRIPTION file
# (2) modified 'Authors' field in DESCRIPTION file
# (3) added plot functions: scatterPlots, boxPlots, and densityPlots
# (4) added a note to the Example for 'lmFitPaired'
#
#Version 0.99.7(Sept. 29, 2015)
## deleted read.bgx and readData functions from this release
# will add them on later
#
# (1) thank Valerie for the following suggestions:
#
# 1) man page examples
#
# Neither read.bgx() or readData() have a running example in the man page nor are they used in
# the vignette. Please add examples to these man pages. The issue is that each night the build
# system runs build and check on each package which makes sure all man page examples and the
# code in the vignette are functioning as expected, the proper dependencies are available.
#
# If you need a sample .bgx file, you can include a small subset in inst/extdata.
#
#
# 2) dependencies
#
# If gplots and lumi are required for any of the functions in iCheck they should be in Depends
# or Imports in DESCRIPTION and imported accorrdingly in the NAMESPACE. Warning messages such
# as these should never be needed:
#
# ~/sandbox/iCheck/R >grep library *
# plotFuncs.R: { stop("The library 'gplots' is required to draw heatmap\n") }
# readData.R: stop("library 'lumi' is required to do vsn normalization!\n")
#
# You have both packages in Depends in DESCRIPTION but have selectively imported in the
# NAMESPACE. This doesn't make sense. If a package is in Depends it should be fully imported,
#
# import(lumi)
# import(gplots)
#
# If you only need select functions, then they should be in the Imports field in DESCRIPTION.
#
# ** This goes for all packages in Depends; if in Depends they need to be fully imported in
# the NAMESPACE.
#
#
# 3) unnecessary library() calls
#
# Please remove these library() calls. Even though they are commented out, the are uncessary.
#
# ~/sandbox/iCheck/man >grep library *
# lkhrWrapper.Rd:#library(lmtest)
# lmFitPaired.Rd:#library(limma)
# lmFitWrapper.Rd:#library(limma)
# pca3DPlot.Rd:#library(scatterplot3d)
# plotCurves.Rd: #library(Biobase)
# plotSamplep95p05.Rd: #library(Biobase)
# sortExpressionSet.Rd: #library(Biobase)
#
#
# 4) vignette
#
# i) library() calls
#
# If Biobase and gplots are in Depends and fully imported in the NAMESPACE you do not need to
# load them when you load iCheck. Please remove these from lines 54, 55:
#
# library(gplots)
# library(Biobase)
#
# ii) unevaluated code
#
# Why is the code unevaluated in these sections? They don't look like time-consuming
# operations ...?
#
# \section{Exclude GC arrays}
# \section{Check squared correlations among replicated arrays}
#
#
#
#Version 0.99.6(Sept. 17, 2015)
# (1) thank Valerie for the following suggestions:
#* Checking if other packages can import this one...
# * REQUIRED: Packages (graphics, stats) which provide abline, mad
# (used in readbgx, readbgx2) should be imported in the NAMESPACE
# file, otherwise packages that import iCheck could fail.
#
#* Checking R Version dependency...
# * RECOMMENDED: Update R version dependency from 3.1.0 to 3.2.
#
#This one is up to you. We generate suggested BiocViews - the idea is that they help
#other users find your package.
#
#* Checking for recommended biocViews...
# * CONSIDER: Adding some of these automatically suggested biocViews:
# Microarray, Preprocessing, DNAMethylation, OneChannel,
# TwoChannel, QualityControl
#
#Version 0.99.5(Sept. 9, 2015)
# (1) remove the /dontrun statements in the man pages
#Version 0.99.4(Sept. 4, 2015)
# (2) put 'MASS', 'limma', 'lmtest', 'scatterplot3d', 'beadarray' and 'Biobase' to Dependes
#
# (1) fix problems flagged in the single package build report:
#* Checking exported objects have runnable examples...
# * REQUIRED: At least 80% of man pages documenting exported objects
# must have runnable examples.The following pages do not:
# genExprSet.Rd, genSimData.BayesNormal.Rd, lmFitPaired.Rd,
# LumiBatch2Table.Rd, pca2DPlot.Rd, pca3DPlot.Rd, plotCurves.Rd,
# plotQCCurves.Rd, plotSamplep95p05.Rd, quantilePlot.Rd, R2PlotFunc.Rd,
# readbgx.Rd, readData.Rd
#
#* Checking if other packages can import this one...
# * REQUIRED: Packages (graphics, graphics, graphics, graphics) which
# provide box, axis, box, axis (used in plotSamplep95p05,
# quantilePlot, readbgx, readbgx2) should be imported in the
# NAMESPACE file, otherwise packages that import iCheck could fail.
#Version 0.99.3(Aug 20, 2015)
# (1) to fix error message about rgl after asking Dr. Duncan Murdoch,
# who is the author of the package 'rgl':
#
#If you use something like this:
#
#if (!interactive())
# options(rgl.useNULL = TRUE)
#
#before making any rgl calls, then rgl won't try to open a display, and
#you shouldn't have problems. You might have trouble doing this if you
#don't know why your package is calling rgl, though.
##
# (2) add the following import in NAMESPACE:
# stats: rchisq, rnorm, na.omit, prcomp, gaussian, glm,
# quantile,
# graphics: par, dev.off, points
# rgl: rgl.useNULL
#
#Version 0.99.2(Aug 19, 2015)
# (1) remove assayDataNew to fix one of the error during checking process
# (2) set all examples relating plotting figures as dontrun to avoid
# the error "unable to open X11 display"
# and the error "error in rgl_init"
#
#Version 0.99.1(Aug 18, 2015)
# (1) fix errors occurred in bioconductor checking :
# * Installing package...
# Warning in rgl.init(initValue, onlyNULL) :
# RGL: unable to open X11 display
# Warning in fun(libname, pkgname) : error in rgl_init
# No methods found in "Biobase" for requests: assayDataNew
#
#
# >>>>>>>
# >>>>>>> INSTALLATION WITH 'R CMD INSTALL --preclean --library=libdir iCheck_0.99.0.tar.gz'
# >>>>>>>
#
# * installing *source* package iCheck ...
# ** R
# ** preparing package for lazy loading
# Warning in rgl.init(initValue, onlyNULL) :
# RGL: unable to open X11 display
# Warning in fun(libname, pkgname) : error in rgl_init
# No methods found in "Biobase" for requests: assayDataNew
# ** help
# *** installing help indices
# ** building package indices
# ** testing if installed package can be loaded
# Warning in rgl.init(initValue, onlyNULL) :
# RGL: unable to open X11 display
# Warning in fun(libname, pkgname) : error in rgl_init
# No methods found in "Biobase" for requests: assayDataNew
# * DONE (iCheck)
#
# #
# Checking for bioc-devel mailing list subscription...
# Error in curl::curl_fetch_memory(url, handle = handle) :
# Timeout was reached
# Calls: <Anonymous> ... request_fetch -> request_fetch.write_memory -> <Anonymous> -> .Call
# Execution halted
# #
#
#
# * testing if installed package can be loaded
# Warning in rgl.init(initValue, onlyNULL) :
# RGL: unable to open X11 display
# Warning in fun(libname, pkgname) : error in rgl_init
# No methods found in "Biobase" for requests: assayDataNew
#
# (2) remove unused functions
# (3) revise readData function to explicit required
# phenotype variables and feature variables
# (4) remove the input parameter 'detectPval' in 'genExprsSet' function
#
#Version 0.99.0(Aug 14, 2015)
# (1)Submit to R Bioconductor
#
#Version 0.7.2 (Aug 13, 2015)
# (1) add a function 'genSimData.BayesNormal' to generate simulated data set
# (2) add examples in Rd files
# (3) revise vignettes
# (4) delete 'calDpv' input parameter in read data functions
#
#Version 0.7.1 (Aug 5, 2015)
# (1) Simplify the package to be submitted to R Bioconductors
#
#Version 0.7.0 (Aug 11, 2014)
# (1) fixed a bug in the function 'LumiBatch2Table': should assign column name for the first column
# of 'dat' and 'detect'
# (2) add an input option 'quote' to the function 'LumiBatch2Table'
#
# Version 0.6.9 (Feb 14, 2014)
# (1) fixed a bug in 'getPCAFunc':
# forgot updating 'es' after deleting probes with missing values
#
# Version 0.6.8 (Jan 11, 2014)
# (1) revised 'getPCAFunc': set 'corFlag = FALSE'
#
# Version 0.6.7 (Sept 29, 2013)
# (1) revised 'getPCAFunc': added line
# dat=na.omit(dat)
# right before the calling of the function prcompt
#
#
# Version 0.6.6 (Sept 05, 2013)
# (1) add output of regression coefficients from glmWrapper
# Version 0.6.5 (August 21, 2013)
# # (1) fixed a typo in the warning message of 'lmFitWrapper':
# # lkhRatioWrapper should be lkhWrapper
# (2) remove genderCheck function output since it is not ready yet
# #
# Version 0.6.3 (March 22, 2013)
# (1) fixed 2 bugs in function plotQCCurves: requireLog2 should take FALSE
# when calling 'plotCurves' inside plotQCCurves because data passed
# to plotCurves has been log2 transformed when calling
# getDatColnames; when plot curve other than 'signal',
# need to call 'getDatColnames' to use ' labelVariable'
# (2) delete input arguments 'sortFlag', 'varSort' and 'timeFormat' in functions
# 'hclustPlotFunc' and 'getPCAFunc'
# Version 0.6.2 (March 22, 2013)
# (1) fixed a bug in pca2DPlot and pca3DPlot: did not sort pcs when sort
# expression data. Actually, we do not need to resort the data since
# it was sorted in the function 'getPCAFunc'.
# To fix this bug, I deleted the input arguments:
# 'sortFlag', 'varSort' and 'timeFormat' from these 2 functions
#
# Version 0.6.1 (March 21, 2013)
# (1) updated the manual file for the function 'getPCAFunc' (adding more info for the output 'pcs')
#
# Version 0.6.0 (Jan. 18, 2013)
# (1) rename 'log.txt' to 'NEWS'
# (2) fixed a bug in function 'addfDat':
# 'ff2<-ff[pos]' should be
# 'ff2<-ff'
# (3) commented out genderChecking function and will revise it later
#
# Version 0.5.9 (Dec. 13, 2012)
# (1) in 'myCombine.R', put combining phenotype data step before merging
# feature data and expression data
#
# Version 0.5.8 (Nov. 14, 2012)
# (1) added function 'lkhrWrapper' to do likelihood ratio test
#
# Version 0.5.7 (Oct. 2, 2012)
# (1) add function genExprSet
# (2) remove 'pcaMethods' package
# (3) complete vingette
#
# Version 0.5.6 (July 20, 2012)
# (1) fill in the vingette (to be continued)
#
# Version 0.5.5 (July 20, 2012)
# (1) revised the functions 'pca2DPlot' and 'pca3DPlot':
# R2vec<-egvals/sum(egvals) instead of
# R2vec<-egvals/sum(egvals)[plot.dim]
# Also replace
# round(R2vec[1],2) with
# round(R2vec[plot.dim[1]],2)
#
# Version 0.5.4 (June 1, 2012)
# (1) revised the function getPCAFunc to use 'prcomp' to
# obtain principal components. The inputs 'scale' and 'center' were
# replaced by 'corFlag'
#
# Version 0.5.3 (May 28, 2012)
# (1) revised read data functions to allow arrays' chip versions are
# different
# Version 0.5.2 (March 16, 2012)
# (1) fixed a bug in function 'filterFunc':
# 'subjIDs.all<-subjID' should be
# 'subjIDs.all <- as.character(pDat[, c(subjID)])'
#
# Version 0.5.1 (Jan. 31, 2012)
# (1) fixed a bug in 'outputFunc': sum() should set 'na.rm=TRUE'
#
# Version 0.5.0 (Jan. 6, 2012)
# (1) revised the function 'myCombine': replace 'which' by 'match'
#
# Version 0.4.9 (Dec. 19, 2011)
# (1) revise the description of the argument 'es' in function 'genderCheck'
# (2) change the default value of the argument 'plotFlag' in function
# 'genderCheck' to be 'TRUE'.
# (3) add randomForest classification in function 'genderCheck'
#
# (4) add vignette
#
# Version 0.4.8 (Dec. 16, 2011)
# (1) revised the function 'genderCheck': first remove confounding
# effects; the argment 'formula' was replaced by 'var.gender';
# the default value for the argument 'probeType' changed to 'Yonly'
#
# Version 0.4.7 (Nov. 30, 2011)
# (1) fixed a bug in function 'myfilter': forgot put the filtering
# of probes with empty gene symbols in condition 'filter=TRUE'
#
# Version 0.4.6 (July 27, 2011)
# (1) added 2 arguments 'scale' and 'center' to the function
# 'getPCAFunc'
# (2) fixed bug in function 'extractIDs3': some times, 'Study_Name' can take
# value 'CONTRL' instead of 'CONTROL'
#
# Version 0.4.5 (July 25, 2011)
# (1) fixed a bug in myCombine function
# I forgot check if chipver.x and chipver.y are NULL
#
# Version 0.4.4 (June 6, 2011)
# (1) simplify function 'separatePairs'
#
# Version 0.4.3 (May 18, 2011)
# (1) fixed a bug: the line colors and line types in plot produced
# by quantilePlot
# are not distinguishable
# (2) fixed a bug in 'filterFunc':
# 'cn<-as.character(pDat[,c(subjID)][pos.rep])'
# should be
# 'cn<-sampleNames(es)[pos.rep]'
# (3) fixed a bug in 'lmFitPaired', 'lmFitWrapper': should check if there is only intercept in the model
# if yes, then need to make sure the pvalMat.unsorted and
# statMat.unsorted are matrices, instead of vectors
# (4) fixed a bug in 'vst' of lumi package: min and max should use the
# option 'na.rm=TRUE'
#
#
# Version 0.4.2 (May 17, 2011)
# (1) fixed a bug in separateGCnonGC which caused wrong objects
# when there are no GC arrays in the ExpressionSet object
# (2) added an input option 'sortFlag'. If sortFlag=TRUE, then
# arrays will be sorted by MAD (median absolute deviation)
# (3) fixed a bug in 'plotSamplep95p05': the output should be 'res'
# not 'qMat'
#
# Version 0.4.1 (May 12, 2011)
# (1) fixed a bug in 'filterFunc': if(excludeGC) should be
# if(excludeGC && n.pos)
# (2) fixed a bug in 'readDataLst.default', 'readData1File.default', and
# 'readData.default': should use 'extractIDs'
# instead of 'extractIDs3'
# (3) added 'projLocLst<-unique(projLocLst)' to avoid duplication
# in function readDataLst.default
#
# Version 0.4.0 (April 2, 2011)
# (1) simplify the code for extractIDs
# (2) simplify the code for addfDat
# (3) simplify the code for mergeDuplicates
# (4) output number of significant tests (raw p-value < 0.05)
# (5) simplify the code for functions in 'myCombine.R'
# (6) simplify the code for function 'readData.default'
# (7) simplify the code for function 'readData'
# (8) added option 'stringsAsFactors=FALSE' in functions in readbgx.R
# (9) added option 'plot.dim' to function 'pca2DPlot'
# (10) added function 'pca3DPlot'
# (11) added outputs for functions 'plotSamplep95p05' and 'quantilePlot'
# (12) revised the function 'separateGCnonGC' since
# GC arrays do not always started with "128115", "Hela", or "Brain" in
# 'Hybridization_Name'
# e.g there are 12 GC arrays having the 'Hybridization_Name':
# Hybridization_Name
#141533_GC008700115DR_4158307007_A 141533_GC008700115DR_4158307007_A
#141532_GC008700070DR_4158307007_D 141532_GC008700070DR_4158307007_D
#141531_GC008700099DR_4158307007_H 141531_GC008700099DR_4158307007_H
#141534_GC008700133DR_4158307009_C 141534_GC008700133DR_4158307009_C
#141539_GC00870017XDR_4158307010_A 141539_GC00870017XDR_4158307010_A
#141535_GC008701024DR_4158307010_C 141535_GC008701024DR_4158307010_C
#141536_GC008700151DR_4158307011_A 141536_GC008700151DR_4158307011_A
#141538_GC008700214DR_4158307011_D 141538_GC008700214DR_4158307011_D
#141540_GC008700198DR_4158307011_H 141540_GC008700198DR_4158307011_H
#141542_GC008700250DR_4158307012_F 141542_GC008700250DR_4158307012_F
#141537_GC008700232DR_4158307016_A 141537_GC008700232DR_4158307016_A
#141541_GC008700279DR_4158307016_E 141541_GC008700279DR_4158307016_E
# in the metaData file
# /proj/reilms/reilm00/GX/CAMP/Kelan_CellLines_Nov08/DexNov08_05Nov2010_Metadata.txt
# (13) revised the function 'extractSamples'
# (14) added the functions 'extractIDs2' and 'extractIDs3'
# (15) added the output of detection p-value in function
# LumiBatch2Table.R
#
#
# Version 0.3.9 (March. 30, 2011)
# (1) added input option 'na.del' to preProcess function
# to delete probes with missing values (not implemented yet)
#
# Version 0.3.8 (March. 21, 2011)
# (1) added input option 'ylim' to functions 'boxPlotFunc'
# and 'boxPlotFunc.default'
#
# Version 0.3.7 (Feb. 23, 2011)
# (1) fixed a bug: the rownames of designMat for lmFit should
# be put back to the column names of corresponding data matrix
#
# Version 0.3.6 (Feb. 22, 2011)
# (1) added the option 'sortFlag' to the function 'filterFunc'
# (2) remove 'dpvFunc' from export list
#
# Version 0.3.5 (Feb. 4, 2011)
# (1) fixed a bug in function 'lmFitWrapper': matrix degenerated
# to vector when only one column
# (2) set 'calDPV=FALSE' by default
#
# Version 0.3.4 (Jan. 10, 2011)
# (1) removed the addition of nuID in read functions since it will be too slow
# (2) revised function 'dpvFunc' to use 'mclapply' instead of 'lapply'
# if library multicore is available to improve speed.
# (3) move 'multicore' to 'Suggests' in DESCRIPTION file
# (4) added 'dpvFunc' to NAMESPACE
# (5) added argument 'calDPV' to functions 'readData', 'readData1File',
# and 'readDataLst'
#
# Version 0.3.3 (Jan. 08, 2011)
# (1) added function 'dpvFunc' to calculate detection p-value
# (2) calculate detection p-value in function 'readData'
# (3) added back the exportion of 'dataCleanLst', 'dataClean1File'
# and 'readDataLst'
#
# Version 0.3.2 (Dec. 21, 2010)
# (1) fixed a bug in 'genderCheck': forgot defining
# xlabPca, ylabPca, xlimPca and ylimPca
#
# Version 0.3.1 (Dec. 8, 2010)
# (1) revise 'genderCheck' to incorporate pca plot and hclust plot
# (2) put functions 'readData1File', 'preProcess', and 'LumiBatch2Table'
# back to export list in 'NAMESPACE'
# (3) added nuIDs to feature data in function 'readData.default'
# 'filterFunc', 'normalizeFunc', and 'readData1File.default'
# (4) added inputs 'probeID.var', 'gene.var', and 'chr.var' to function
# 'LumiBatch2Table'
# (5) replace 'ProbeID' by 'nuIDs' in output of statistical analysis results
#
# Version 0.3.0 (Dec. 7, 2010)
# (1) forgot revising function 'genderCheck' after revising function
# 'lmFitWrapper'
#
# Version 0.2.9 (Dec. 3, 2010)
# (1) revise function 'extractSamples' to allow get nonGC samples
# (2) added function 'preProcessWrapper'
# (3) clean NAMESPACE and INDEX to remove some functions from export list
# (4) added function 'separatePairs' to separate paired data
#
# Version 0.2.8 (Dec. 1, 2010)
# (1) revise function 'mysort' so that the user can
# specify which variables will be used to sort the samples
# of an ExpressionSet object
# (2) add inputs 'varSort' and 'timeFormat' to plot functions
# (3) revised functions 'glmWrapper' and 'lmFitWrapper'
# by using some modules to summary the common tasks
#
# Version 0.2.7 (Nov. 21, 2010)
# (1) put the check of the covariate of the interest
# outside the loop for glmWrapper to improve efficiency
# (2) add the check of the covariate of the interest
# to the functions lmFitWrapper and lmFitPaired
# (3) add functions 'inputChecking', 'getCovInterest',
# 'checkCovariate' to modulize my Rcode
#
# Version 0.2.6 (Nov. 20, 2010)
# (1) added input 'pos.var.interest' to function 'glmWrapper'
# to indicate which covariate in the right-hand-side
# of ~ in the 'formula' is of the interest
# pos.var.interest = 0 means intercept is of the interest
#
# Version 0.2.5 (Nov. 19, 2010)
# (1) added input 'probeIDs' for the functions
# 'lmFitWrapper' and 'glmWrapper'
# (2) added output 'pvalMat' and 'statMat' for the functions
# 'lmFitWrapper' and 'glmWrapper'
# (3) added function 'lmFitPaired'
# (4) replace the statement 'resMat2<-do.call(rbind, t4)'
# in glmWrapper by 'resMat<-t(sapply(t4, function(x) {x}))'
# It is much faster now and without error
#
# Version 0.2.4 (Nov. 6, 2010)
# (1) fixed a bug in myCombine: it will produce more rows than
# expected when merging controlData since row names are not unique.
# In function 'readData', it is okay since 'controlData' slot is
# NULL until we set its value after reading and combining
# all SampleProbeProfile and QCProbeProfile.
# But if we want to combine 2 LumiBatch object created by
# iCheck, it will cause error. No good fix yet. Just temporary fix:
# if two controlData have the same row names, then use cbind.
# if not the same, then use the rows with common row names, then use
# cbind. Otherwise, set combined LumiBatch object have NULL
# controlData.
# (2) fixed a bug in 'readData1File' and 'readData1File.default':
# no need to use option 'combineRule' and no need to use variable count
#
# Version 0.2.3 (Nov. 5, 2010)
# (1) fixed a bug in pca2DPlot: forgot define 'sortFlag' in arguments
#
# Version 0.2.2 (Nov. 5, 2010)
# (1) To make sure sorting Batch_Run_Date correctly,
# Batch_Run_Date should have the format m/d/yyyy
# Note that year should contain century. e.g. 2010, not 10
#
# Version 0.2.1 (Nov. 4, 2010)
# (1) the function 'scores' is from library 'pls', not 'pcaMethods'
# (2) add 'pcaMethods' to line 'Depends' in file 'DESCRIPTION'
# (3) add option 'sortFlag' to related plot functions
# (4) split function 'preProcess' to 2 functions:
# 'preProcess' and 'normalizeFunc'
# (5) added function 'glmWrapper'
#
# Version 0.2.0 (Oct. 29, 2010)
# (1) fixed a bug in 'lmFitWrapper': sometimes the rownames of
# 'designMat' is not numeric. Hence, we added an temporary pDat2
# to generate designMat
# (2) fixed a bug in print out top 20 genes in 'lmFitWrapper'
#
# Version 0.1.9 (Oct. 6, 2010)
# (1) fixed a bug in 'lmFitWrapper': the number of rows of
# 'designMat' might be less than that of pDat due to
# missing values. Hence, we need to reduce the dimension of
# 'dat'
# (2) added function 'pcaPlotFunc' to draw scatter plot of
# first 2 principal components of subjects. (very slow)
# (3) added function 'quantilePlot' to draw quantiles across arrays
# (4) revised function 'hclustPlotFunc' to add color to labels
# (5) added function 'pca2DPlot'
# (6) using grDevices::rainbow to define colors
# (7) added function 'replacePheno2es'
# (8) added functions 'getDatColnames' and 'extractREplicates'
# (9) revised 'plotSample95p05'
#
# Version 0.1.7 (Sept. 24, 2010)
# (1) fixed a bug: should add ... when calling function 'axis'
# in function plotCurves and when calling boxplot in plotFuncs.R
#
# Version 0.1.6 (Sept. 24, 2010)
# (1) fixed a bug: The object 'beadNum.y' in the command
# 'colnames(beadNum.y) <- sn.y' in the file 'myCombine.R'
# should be 'beadNum.dat.y'
#
# Version 0.1.5 (Sept. 10, 2010)
# (1) fixed a bug: Some times, Probe_Sequence is not unique,
# that is, The same Probe_Sequence might correspond to mulitple
# probe ids. In this case, if we use Probe_Sequence
# as featureNames, we will get error message.
# Hence, we need to merge duplicates first, before
# set Probe_Sequence as featureNames.
# (2) fixed a bug in myCombine.R: after merging, the order of
# Probe Sequences of data might be different from that of fData
# (3) added check if any subject Ids are duplicated in a SampleProbeProfile
# or QCProbProfile immediately after calling 'lumiR'
# (4) added check if any subject Ids are duplicated in a SampleProbeProfile
# or QCProbProfile in function 'myCombineUnion'
# (5) added function 'LumiBatch2Table' to output exprs, pDat, fDat of
# an LumiBatch object to 3 text files
# (6) added function 'lmFitWrapper' to do gene differential analysis
# (7) added function 'plotSamplep95p05' to draw the trajectories
# of the ratio of the 95-th percentile to the 5-th percentile
# of expression levels across arrays.
# (8) fixed a bug: The error message: "No matched manifest file!" is
# not correct. It should be "All samples in expression sets are not in
# pDat"
# (9) fixed a bug: After the warning message:
# "The following arrays which are in QC probe profile file
# are not in sample probe profile file!"
# we should call 'print(snQC.diff)' instead of 'print(sn.diff)'
# (10) added function 'genderCheck'
# (11) added functions 'readDataLst', 'readDataLst.default',
# and 'dataCleanLst' to read data given a list of
# project locations
#
# Version 0.1.4 (August 25, 2010)
# (1) fixed a bug in lumiB2:
#quantile.ctrl <- apply(control, 2, quantile, probs = probs,
# ...)
#should be
#quantile.ctrl <- apply(control, 2, quantile, probs = probs, na.rm=TRUE,
# ...)
# (2) remove calling 'lumiQ', which will take a lot of times
# (3) set 'QC=FALSE' when calling lumiR
# (4) fixed a bug when sorting Batch_Run_Date, Chip_Barcode,
# and Chip_Address. Bath_Run_Date is a vector of characters.
# so "7/2/2010" will be greater "7/19/2010". To fix this,
# we can use function 'strptime' to convert the character string
# to time string. Then "7/2/2010" will be smaller than "7/19/2010".
# (5) keep the probes in chromosomes with labels
# different than 1, 2, ..., 22, X, or Y, but not empty
# string or not containing string "random".
##
# Version 0.1.3 (August 9, 2010)
# (1) fixed a bug: when outObjFileDir = NULL, the intermediate output
# will print NULL. It should output the current directory
# (2) added an option 'verboseQC' to mask intermediate output of QC
# (3) added an option 'verbose' to mask all intermediate output
# (4) added function 'readGCAll', 'addfDat', and myCombineUnion
# to read GC samples
# which might be from different Illumina chip versions.
# (5) added the argument 'use="complete.obs"' when calling the function
# 'cor' in the function 'R2PlotFunc'.
# (6) added function 'preProcess' to conduct data pre-processing
# including remove replicates (Although there is a function
# called 'lumiExpresso' in lumi package, we do more data pre-processing,
# such as excluding replicated arrays, GC samples, failed samples,
# probes with empty gene symbol and/or in chromosomes other than
# 1,...,22, X, and Y)
#
# (7) added function 'compareManifest' to count the overlap of probes
# in each Illumina chip version
# (8) checking the warning message: no bgAdjust is needed
# (9) in the function R2PlotFunc, get the replicates by checking
# if a subject has more than one array, instead of using
# the column 'Replicate_Sample' of pDat since some time this
# column gives wrong info
# (10) revise read functions so that 'manifestDir' instead of 'mDat' object
# will be the input and the functions will detect which manifest file
# to read based on 'manifestDir' and the column 'Chip_Manifest'
# of pDat. Also sample probe file and QC probe file will be
# read in using one function. controlData slot will be added.
# (11)
# added an argument 'combineRule' to indicate what strategy will be
# used to combine two LumiBatch objects: union or intersection
# (12) in files 'readData.default' and 'readData1File.default',
# added an argument 'sampleFlag'. 'sampleFlag=TRUE' means
# we will read sample probe file; 'sampleFlag=FALSE' means
# we will read QC probe file. The manifest data are different
# for these two types of files.
# (13) There are some minor bugs in lumi's data pre-processing functions
# such as 'min(x)' should be 'min(x, na.rm=TRUE)'. We revised
# these functions so that we can do data pre-processing for
# LumiBatch object obtained by combining severa LumiBatch objects
# using "union" strategy instead of "intersection" strategy.
# (14) deleted functions 'readDataAll', 'readQC', 'readGC'.
# (15) add an argument 'saveFlag' indicating if LumiBatch object
# to be saved.
# (16) delete some arguments, eg. 'QC', 'verboseQC'.
# Version 0.1.2 (August 3, 2010)
# (1) added output number of samples in pDat
#
# Version 0.1.1 (August 3, 2010)
# (1) fixed a bug when handling GConly=TRUE. If there is no GC samples in
# file and GConly=TRUE, warning message will given.
# If there is no GC samples in all files and GConly=TRUE,
# program will exit abnormally with error message.
# (2) in function 'extractIDs', if no GC sample, then return NULL
# instead of exit abnormally
#
# Version 0.1.0 (July 30, 2010)
# (1) added argument 'GCid' for function 'R2PlotFunc'
# (2) added function 'boxPlotFunc.default' to draw boxplot
# for sample probe only, or QC probe only
# (3) added function 'extractSamples' to subset an ExpressionSet
# (4) added function 'extractIDs' to subset a character string vector
# (5) allow GCid be a vector of GC ids
#
# Version 0.0.9 (July 29, 2010)
# (1) added functions 'readDataAll' and 'readQCAll' to read
# all sampleProbeProfile files and QCProbeProfile files in a directory
# and in the subdirectories of this directory
# (2) fixed a bug in subset pDat when 'GConly=TRUE'
#
# Version 0.0.8 (July 28, 2010)
# (1) added an argument 'filter' to functions
# 'readData', 'readData.default', 'readData1File'
# if 'filter=TRUE' (by default), then the "bad" probes
# will be excluded from the output
#
# Version 0.0.7 (July 28, 2010)
# (1) fixed a bug when checking failed samples
# (2) added functions 'readData1File' and 'readQC1File'
#
# Version 0.0.6 (July 27, 2010)
# (1) replaced 'protocalData' by 'Biobase::protocalData' in the file
# myCombine.R and myStack.R
#
# (2) added 'combine' and 'protocalData' in the 'importForm' of 'Biobase'
# in NAMESPACE file
#
# (3) fixed a bug in readData.R. pDat used after rm(pDat)
# (4) change the argument 'outObjFile' to 'outObjFilePrefix'
# (5) added argument 'outObjFileDir'
# Version 0.0.5 (July 21, 2010)
# (1) added 'read.bgx' function which is derived from
# the function 'readBGX' in library 'beadarray'.
# 'readBGX' can handle .bgx file with 3 blocks
# [Heading], [Probes], [Controls], and
# will not handle HumanHT-12 chips which
# have 4 blocks [Heading], [Probes], [Controls], [Columns].
# (2) replace the argument 'mFileName' in read data functions
# by 'mDat'
# (3) add an argument 'labelVariable' to function 'plotQCCurves'
# (4) add an argument 'excludeFails' to read functions
# (5) set the default value of the argument 'bgAdjustMethod' to 'bgAdjust'
#
# Version 0.0.4 (July 20, 2010)
# (1) fixed a bug in assigning featureNames in readData.default
# and readQC.default when adding
# manifest data to combined LumiBatch object
# The bug caused the featureNames became 1, 2, ...
# (2) fixed a bug in calling lumiQ function in readData.default,
# readQC.default, mySummary, and myStack. By default,
# lumiQ will log transform data if data are not log transformed.
# To fix this, set the argument 'logMode=FALSE'.
#
# Version 0.0.3 (July 19, 2010)
# (1) read functions output 'lumiBatch' objects instead of
# 'ExpressionSet' object so that we can use
# lumi's functions 'lumiQ' etc.
# (2) do background correction and data transformation
# after combining data from different chips
# (3) remove some objects when they are no longer used to save memory
#
# Version 0.0.2 (June 28, 2010)
# (1) added output of un-normalized data
# (2) get a complete version
#
# Version 0.0.1 (June 25, 2010)
# QC pipeline data tracking tools for Channing Laboratory Respirotary Group's high-dimensional Illumina mRNA expression data.
#