The installation of gMCP via install.packages("gMCP") should be
unproblematic. However three of the packages gMCP depends on or 
suggests perhaps need more attention:

graph:
======

Since R 2.13 the package graph is not found any more on CRAN,
but is still available from bioconductor:

http://www.bioconductor.org/packages/release/bioc/html/graph.html

Install it via

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

rJava:
======

Follow the instructions of rJava.

For Windows:

If you run into a "MSVCR71.dll is missing" please try updating your
rJava installation (install.packages("rJava")) or add your JAVA_HOME
directory (something like "C:\Program Files\Java\jre6\bin\") to your
PATH.

For Unix:

Most likely you will have to run the command "R CMD javareconf" with
root permissions to let R find your java compiler and headers.

Rgraphviz:
==========

For visualization the method renderGraph from package Rgraphviz can be
used, which uses the graphviz library.  Therefore you need to install
the libraries and headers necessary to link graphviz library routines
into the Rgraphviz package. For example on Debian/Ubuntu you need to
install the package libgraphviz-dev beforehand:

sudo apt-get install libgraphviz-dev

To install Rgraphviz, start R and enter:

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

Compatibility for R 2.8.0
=========================

gMCP is still compatible with R 2.8.0 but not extensively tested.

To get a Windows build go to
http://algorithm-forge.com/gMCPforR28/ 
or (if it's not up-to-date on the site) build the current gMCP package from source for your system.  

If you run into any problems, please write us an email at:

help@small-projects.de
