1

I need to plot area proportional Venn Diagrams with at least 5 variables.

I tried to install Vennerable package but its just not available anymore. Link to windows build doesn't work (page not found).

Is there an alternative package?

Marek
  • 49,472
  • 15
  • 99
  • 121
user670186
  • 2,588
  • 6
  • 37
  • 55
  • 1
    Maybe you should throw in some capitals and exclamation marks to make us all more eager to help you? I wonder why your previous question got some negative response. – Nick Sabbe May 15 '11 at 17:55
  • @user670186 In case you didn't know this: Using capital letters and exclamation marks is considered and the same as shouting, and is very impolite. There is no need to shout on this forum. If you post a clear question, and if someone knows the answer, we will try to help. – Andrie May 15 '11 at 18:19
  • possible duplicate of [R install Vennerable package on windows](http://stackoverflow.com/questions/6007883/r-install-vennerable-package-on-windows) – Joris Meys May 16 '11 at 08:01
  • @Joris Not duplicate. Previous question is about installing Vennerable, this one about alternative of it. – Marek May 16 '11 at 08:11
  • Vennerable is available again. I add answer about it to your [previous question](http://stackoverflow.com/q/6007883/168747). – Marek Jun 06 '11 at 11:30
  • There are some solutions in [old question about Venn diagrams in R](http://stackoverflow.com/q/1428946/168747). – Marek May 17 '11 at 12:30

5 Answers5

9

The source files, last updated in 2007, are downloadable from Sourceforge as a tar.gz file:

http://sourceforge.net/projects/vennerable/files/R%20Source%20package/1.1.1.1/Vennerable_1.1.1.1.tar.gz/download

Better yet, what appears to be more updated (Edit: and broken) source files, last updated in 2009, can be checked out from R-Forge using SVN:

svn checkout svn://svn.r-forge.r-project.org/svnroot/vennerable

Can you build it yourself with Rtools? The Linux x86_64 build log suggests a problem with the vignette causing the build to fail, and the source tar.gz and binaries to be unavailable. Perhaps building it yourself without the vignette will work. The package was last updated 19 months ago, so it does not appear to be currently maintained.

Edit

It seems the most recent revisions from R-Forge break the package. There is a problem with the compute.Venn() function, and a number of tests fail. I built the Venerable_1.1.1.1 revision from Sourceforge for you for x86_64 architecture. You can download it at:

http://commondatastorage.googleapis.com/jthetzel-public/Vennerable_1.1.1.1.zip

I haven't actually tested the package, but it successfully installs and loads into R 2.13.0 x86_64 on Windows 7. If you have further questions, you can email the author of the package, Jonathan Swinton, whose email you will find in the DESCRIPTION file in the zip archive.

Note the rare correct use of all caps to refer to the DESCRIPTION file, which was so named for presumably historic reasons (though I'd be interested if anyone has a definitive explanation). Most other uses of all caps are generally frowned upon, unless you specifically intend to raise the systolic pressure of the SO community.

jthetzel
  • 3,603
  • 3
  • 25
  • 38
6

Follow this:

source("http://bioconductor.org/biocLite.R")       
biocLite("graph")       
biocLite("RBGL")       
biocLite("RColorBrewer")       
biocLite("gtools")       
biocLite("reshape" 
install.packages("Vennerable", repos="http://R-Forge.R-project.org") 
athi
  • 61
  • 1
  • 1
4

A look at CRAN (the place to be for R packages) revealed at least two promising packages: VennDiagram and venneuler.

Nick Sabbe
  • 11,684
  • 1
  • 43
  • 57
  • VennDiagramm: does only circles and area representative Diagrams with up to 3 variables. I have 6. veneuler can do more than 3, but its verz inaccurate ( I tested). I need a package that can display squares or other shapes for proper representation. The only package I found that can do this is Vennerable. But the link is DOWN. Does someone have the Windows Vennerable_2.0 binary? – user670186 May 15 '11 at 16:54
2

Vennerable 3.0 is available. It cand be downloaded eg from within R by
install.packages("Vennerable", repos="http://R-Forge.R-project.org")

  • I think it would be good if you could add information how to get dependent packages `RBGL` and `graph`. – Dieter Menne Jul 29 '13 at 13:24
  • 1
    The above comment suggests providing additional information about getting dependent packages, but the solution had just worked for me seamlessly. It's true that I might have had dependancies already installed from bioconductor, but that should be easy to do by following instructions on [bioconductor install page](http://www.bioconductor.org/install/). Anyways, thanks @Jonathan Swinton – Sergei Jun 07 '15 at 17:08
1

Thorough inspection leads to this bioinformatics engine, of higher performance than available R packages for Venn diagrams out there so far. It is written in Perl by Belgium author Lieven Sterck (lieven.sterck@psb.vib-ugent.be) from Ghent University.

The webtool generates as output not only the requested diagram (provided the input consists of 5 sets or less, already surpassing capabilities of most R packages available for Venn diagrams) but also produces a text file listing the elements contained in each intersection group for detailed inspection. Lists, although not geometric diagrams as such, can be generated even for comparisons exceeding 5 sets. Furthermore, this webtool also outputs the total number of elements per dataset in a tabulated manner for confirmation of unique entries.

The only missing feature: a weighted option equivalent to the "Chow-Ruskey" in the Vennerable R package, which however seems to suffer from some issues (besides flaky installation as mentioned above) adding an extra element to the last group list provided in the data frame for overlap. Vennerable (and other packages of the same ilk) only generates a figure with group labels and numbers the elements contained per intersection, not listing them. Venn diagrams and intersection data in general, although simple in principle are of great utility to illustrate coverage of "multiomics" biological data just to provide with an example.