>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> The package handles the pre-processing and normalization of Affymetrix GeneChip expression arrays, including Exon arrays. It is an R wrapper to XPS (eXpression Profiling System), which is based on ROOT, an object- oriented framework developed at CERN. Thus, the prior installation of ROOT is a prerequisite for the usage of this package. ROOT can be downloaded from http://root.cern.ch. Copyright (C) 2000-2012 Christian Stratowa, Vienna, Austria This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, a copy of the GNU General Public License is available at http://www.gnu.org/copyleft/gpl.html. You can also obtain it by writing to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Content: I. Install ROOT: Ia. Install ROOT from source Ib. Install ROOT for Windows XP Ic. Install ROOT for Mac OS X Id. Notes on ROOT II. Install package 'xps': IIa. Install 'xps' from source IIb. Install 'xps' as binary IIc. Install 'xps' for Windows XP IId. Install 'xps' for Windows 7 (Vista) IIe. Tested architectures III. Learning how to use package 'xps' >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> I. Install ROOT: Ia. Install ROOT from source: Installation from source is explained extensively on the ROOT website, see: http://root.cern.ch/drupal/content/installing-root-source 1. get the sources of the latest ROOT from: http://root.cern.ch/drupal/content/downloading-root or from: ftp://root.cern.ch/root/ Note 1: you can unzip root locally in any directory, resulting in /root Note 2: If you want to install root in a fixed location please read: http://root.cern.ch/drupal/content/installing-root-source 2. set ROOTSYS to the `root' directory: sh family: export ROOTSYS=/root csh family setenv ROOTSYS /root 3. now type the build commands: all shells: ./configure --help [optional, shows e.g. for supported architectures] ./configure [] [e.g. = linux or macosx (32bit) or macosx64 (64bit)] make [or, make -j2 for dual CPU machines] Note: ./configure should automatically detect the architecture, if not use e.g. ./configure linux (32bit) ot linuxx8664gcc (64bit) 4. add bin/ to PATH and lib/ to LD_LIBRARY_PATH. sh family: . bin/thisroot.sh csh family source bin/thisroot.csh 5. add $ROOTSYS/bin to PATH and $ROOTSYS/lib to LD_LIBRARY_PATH: sh family: export ROOTSYS=/root export PATH=$ROOTSYS/bin:$PATH export LD_LIBRARY_PATH=$ROOTSYS/lib:$LD_LIBRARY_PATH export DYLD_LIBRARY_PATH=$ROOTSYS/lib:$DYLD_LIBRARY_PATH (MacOS X only) csh family: setenv ROOTSYS /root setenv PATH ${ROOTSYS}/bin:${PATH} setenv LD_LIBRARY_PATH ${ROOTSYS}/lib:${LD_LIBRARY_PATH} setenv DYLD_LIBRARY_PATH ${ROOTSYS}/lib:${DYLD_LIBRARY_PATH} (MacOS X only) rehash Note: it is best to include above lines in e.g. ".bashrc" or ".bash_profile" 6. try running root: all shells: root Ib. Install ROOT for Windows XP For Windows XP users it is important to install the recommended binary version of ROOT, currently: root_v5.32.01.win32.vc90.msi WindowsXP/NT/w2000 with VC++ 9.0, version 5.32/01 from: http://root.cern.ch/drupal/content/production-version-532 It is recommended to install ROOT in the default directory C:\root. Furthermore, it is necessary to set the environment variable ROOTSYS: under Windows 2000/XP/2003 you can use `System' in the control panel or the properties of `My Computer' (under the `Advanced' tab). Simply add: ROOTSYS=C:\root\bin (or: ROOTSYS=C:\root if you add %ROOTSYS%\bin to PATH) Ic. Install ROOT for Mac OS X Mac users can either install ROOT from source or use the recommended binary version of ROOT. For Leopard you need to compile the source: root_v5.32.01.source.tar.gz For Snow Leopard there exist binaries for 32 bit and 64bit, however, since the xps binary for BioC 2.10 is built on Leopard for 32 bit you need to install: root_v5.32.01.macosx106-i386-gcc-4.2.tar.gz Personally, I am a Mac user and do all my development on my MacBook Pro, where I have installed WinXP and openSUSE 11.4 using VMware Fusion in order to test xps for these architectures. I prefer to install root from source. On Leopard I install the 32 bit version: ./configure macosx make On Snow Leopard I install the 64 bit version: ./configure macosx64 make On Lion I install the 64 bit version: ./configure macosx64 --with-f77="/usr/bin/gfortran-4.2" make Id. Notes on ROOT: Note 1: Alternative ROOT installation If you want to install ROOT in a fixed location, not depending on ROOTSYS, please see: http://root.cern.ch/drupal/content/installing-root-source Note 2: ROOT demos Like R, ROOT contains many examples demonstrating the abilities of ROOT. These examples can be run to test if ROOT has been installed correctly. There are two kind of examples, located in subdirectories tutorials and test, respectively: 1. tutorials: To test root and try some of its tutorials, do: cd root/tutorials root root [0] .x demos.C For other tutorials, see the README file in the tutorials directory. 2. tests: Examples in the test subdirectory need to be compiled first: cd root/test make root root [0] gSystem.Load("Tetris") root [1] Tetris t For other tests, see the README file in the test directory. Note 3: Source vs binary In principle, it is also possible to download and install binaries for some architectures. However, this is not recommended, since the sources for package 'xps' depend on the ROOT framework, and need to be compiled with the same compiler version as ROOT itself. Furthermore, daily versions from ROOT can be downloaded via svn and should work fine with 'xps', however, the current version 5.32/01 and version 5.30/00 have been used to test 'xps' for both Linux openSUSE 11.4 and MacOS X and WinXP. Note 4: ROOT fails to compile Like R, the ROOT framework can be installed on many architectures, see the list of supported architectures, when running "./configure --help". However, ROOT may also compile on a number of other architectures. In case, the current official source file "root_v5.32.01.source.tar.gz" fails to compile, you can download the current development version from svn, see: http://root.cern.ch/drupal/content/subversion-howto If this still fails and you are still interested to compile ROOT, please contact the RootTalk forum at: http://root.cern.ch/phpBB3/ This will help the ROOT developers to eliminate potential problems. Note 5: Can I use ROOT for other purposes? ROOT is a powerful professional C++ framework with many classes,see: http://root.cern.ch/drupal/content/users-guide http://root.cern.ch/drupal/content/reference-guide However, even if you never intend to use any of these classes or not even to use C++, but want to develop your own C functions, as many R developers do, ROOT may be of interest to you, since by typing "root" you start a C/C++ interpreter, which allows you to test your functions without the need to compile them first. Simply create a text file, called a macro, containing your C functions. As an example create a text file "helloROOT.C" (the '.C' extension is important), containing the simple function: void hello_root() { cout << "Hello R, welcome to ROOT!" << endl; } After starting ROOT, type: ('.L' will load your text file) root [0] .L helloROOT.C root [1] hello_root() Hello R, welcome to ROOT! root [2] .q When your function works, then you can compile it directly within a ROOT session by typing ".L helloROOT.C+". Adding the '+' will automatically start the compiler and create a shared library. However, in this case you need to include the necessary header files first, in this case "#include ", then start ROOT: root [0] .L helloROOT.C+ Info in : creating shared library /Volumes/CoreData/CRAN/./helloROOT_C.so root [1] hello_root() Hello R, welcome to ROOT! root [2] .q For more information, please contact the ROOT User Guide: http://root.cern.ch/drupal/content/users-guide >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> II. Install package 'xps': IIa. Install 'xps' from source Currently, xps can be downloaded as gezipped source file "xps_1.16.0.tar.gz" for MacOS X and for Linux. It can be installed with: R CMD INSTALL xps_1.16.0.tar.gz or e.g. on a Mac in the User library: R CMD INSTALL -l ~/Library/R/library xps_1.16.0.tar.gz However, the recommended way to install packages from source is: source("http://bioconductor.org/biocLite.R") biocLite("xps", type="source") IIb. Install 'xps' as binary Package xps can also be installed as binary for MacOS X from the Bioconductor website using: source("http://bioconductor.org/biocLite.R") biocLite("xps") Important note: Please note that Bioconductor has built the xps binaries for release BioC 2.10 using ROOT version root_v5.32.01. Thus it is absolutely essential to install this ROOT version first, otherwise calling xps will result in an error message. You can always download any version of root from: ftp://root.cern.ch/root IIc. Install 'xps' for Windows XP The Windows binary of xps could be downloaded from the Bioconductor web site, and installed using: biocLite("xps") If you run xps from Rterm check if `System' in the control panel does contain the following environment variables: PATH: C:\Programs\R\R-2.13.0\bin;%ROOTSYS%\bin; ROOTSYS:C:\root Note: Starting with BioC 2.9 the main build machine is a Windows Server 2008 R2 Enterprise (64-bit), which is not able to create Windows XP binaries of xps. Thus a downloadable binary for Windows XP does no longer exist. The only option is to install xps for Windows XP from source. Alternatively, I can always send you the xps binary upon request. If you want to install xps from source, then you need to do the following: You can still install the binaries for ROOT and R but in addition you need to install the following: - Rtools.exe from: http://www.murdoch-sutherland.com/Rtools/ - Visual C++: you can download the image "VS2008ExpressENUX1397868.iso" for the current version "Visual Studio 2008 Express" from the Microsoft web site. Before building xps from source you need to start the VC++ batch file provided by Microsoft: C:\Program Files (x86)\Microsoft Visual Studio 9\VC\vcvarsall.bat (or a similar path). Then you can install xps from source as usual: R CMD INSTALL xps_1.16.0.tar.gz IId. Install 'xps' for Windows 7 (Vista) I could install ROOT on Windows 7 (32bit) and could compile xps w/o problems. Furthermore, I could load library(xps). However, using any xps method will result in a bus error. Thus, I am sorry to say that xps does not work with Windows 7. I would like to support Windows 7 but it seems not to be possible, because it seems that R does not support DLLs written in VC++ which depend on other DLLs written in VC++, see: https://www.stat.math.ethz.ch/pipermail/bioconductor/2009-December/030783.html The only possibility would be to compile ROOT with MinGW, the compiler used by R. Many people have tried to compile ROOT with MinGW, but nobody has succeeded yet. IIe. Tested architectures Currently, package 'xps' has been built (with R CMD INSTALL) and passed R CMD check with no errors and no warnings on the following computers: 1. MacBook Pro Core 2 Duo 2GHz with 2 GB RAM running MacOS X 10.6.3 and 10.4.11 R versions tested: R-2.7.x, R-2.8.x, R-2.9.x, R-2.10.x, R-2.11.x, R-2.12.x, R-2.13.x, R-2.14.x, R-2.16.x 2. PowerBook Titanium G4 550MHz with 1 GB RAM running MacOS X 10.4.9 and 10.4.11 R versions tested: R-2.5.0, R-2.6.0, R-2.8.0 3. HP Compaq nx6110 Intel Centrino 1.4 GHz with 512 MB RAM with Fedora Core 4 R versions tested: R-2.4.1, R-2.6.0, R-2.7.0, R-2.7.1 4. HP Compaq nx6110 Intel Centrino 1.4 GHz with 512 MB RAM with Windows XP R versions tested: R-2.6.2, R-2.7.0, R-2.7.1 5. VMware Fusion on MacBook Pro running Windows XP or openSUSE 11.3 R versions tested: R-2.9.x, R-2.10.x, R-2.11.x, R-2.12.x, R-2.13.x, R-2.14.x >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> III. Learning how to use package 'xps' The best way to learn how to use package 'xps' is to follow the examples in the enclosed vignette "An Introduction to the xps Package". As next step it is recommeded to follow the step-by-step examples in the script "script4xps.R", which is available in the "examples" subdirectory of 'xps'. Directory xps/raw contains four CEL-files for chip Test3.CDF, which are used in most of the examples. However, real world examples in this script include a subset of the Affymetrix Exon Array Dataset human Tissues for HG-U133_Plus_2 arrays, HuEx-1_0-st-v2 arrays and HuGene-1_0-st-v1 arrays. The corresponding CEL-files must be downloaded from the Affymetrix web site. The use of Affymetrix human Tissues dataset is also explained in more detail in the example script "script4exon.R". Comments regarding the tested architectures: 1. MacBook Pro Intel Core 2 Duo 2GHz with 2 GB RAM running MacOS X 10.6.6 All examples were tested, including the Affymetrix human mixture dataset for HG-U133_Plus_2, HuGene-1_0-st-v1 AND HuEx-1_0-st-v2.r2 2. PowerBook Titanium PowerPC G4 550MHz with 1 GB RAM running MacOS X 10.4.9 Most examples were tested, including the Affymetrix human mixture dataset for HG-U133_Plus_2 (HuEx-1_0-st-v2.r2 should also run, as I have tested them earlier, but they are very slow on G4 550 MHz) 3. HP Compaq nx6110 Intel Centrino 1.4 GHz with 512 MB RAM with Fedora Core 4 All examples were tested, including the Affymetrix human mixture dataset for HG-U133_Plus_2 but NOT HuEx-1_0-st-v2.r2. 4. HP Compaq nx6110 Intel Centrino 1.4 GHz with 512 MB RAM with Windows XP Most examples were tested, including the Affymetrix human mixture dataset for HG-U133_Plus_2, HuGene-1_0-st-v1 AND HuEx-1_0-st-v2.r2. Note: I could indeed run RMA using 6 exon arrays of the Affmetrix dataset with only 512 MB of RAM. However, it may not be possible to run MAS5, since the MAS5 background correction uses lots of RAM. Furthermore, it was not possible to create the scheme for the exon arrays with "import.exon.scheme()", since the Affymetrix files are too large; you need at least 1GB RAM. >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>