>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
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 <path>/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=<path>/root
csh family
setenv ROOTSYS <path>/root
3. now type the build commands:
all shells:
./configure --help [optional, shows e.g. <arch> for supported architectures]
./configure [<arch>] [e.g. <arch>= 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 $ROOTSYS/lib to LD_LIBRARY_PATH:
sh family:
export ROOTSYS=<path>/root
export PATH=$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 <path>/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 <iostream>", then start ROOT:
root [0] .L helloROOT.C+
Info in <TUnixSystem::ACLiC>: 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.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>