Tool to generate proportional Venn Diagrams?

I saw this Biostar post on Venn diagrams for GO term overlap. And this tool could be modified to do something like I wish, but was wondering if there's a tool that can make nicely proportional Venn diagrams by entering just the number of elements in each set and intersection.

For instance circle A has 10 items, B has 10 items, A+B has 6 items. Drawing them proportionally would be nice, but not essential. Also nice, but not critical, would be something open source which could be integrated with a suite of Perl scripts.

P.S. not using any standard GO terms, gene IDs, etc. This is for CNVs. So elements would take the form of: chr1:100000-200000.

visualization • 46k views

ADD COMMENT • link •

modified 5 weeks ago by Renesh • 1.8k • written 8.8 years ago by Ryan D • 3.3k

8.8 years ago by

Larry_Parnell ♦ 16k

Boston, MA USA

Larry_Parnell ♦ 16k wrote:

Something quite interesting that I learned from the VizBi conference (workshop on visualizing biological data) last month concerns depicting relationships linearly vs by area. Humans can proportionally relate objects of different lengths. That is if A = 10 and B = 5, A looks exactly twice as long, and if A = 10 and B = 9.8, B looks shorter. If A and B are depicted by area, things get difficult to discern because perceived differences in area relate to one another not by the power of 1.0 but by 0.6-0.7. Thus, one loses the degree of difference between objects when using area.

Although Venn diagrams are popular and convenient, this information is something to keep in mind when trying to depict the relationships.

Added 27 Apr 2011: Try Hohli at charts.hohli.com. With Hohli, you can create a Venn diagram and customize its look and feel. One can specify a title, choose data points, and alter the background. Hohli also allows you to create other charts, including scatter plots and other line charts.

ADD COMMENT • link modified 8.7 years ago • written 8.8 years ago by Larry_Parnell ♦ 16k

Good point! Another option to consider is UpSetR: https://github.com/hms-dbmi/UpSetR

It's a very versatile tool with helpful vignettes.

ADD REPLY • link written 13 days ago by ablazier • 10

A good thought, Larry. I realize Venns are not as sexy as some. Do you have any example links from the conference (other otherwise) that would be good for showing the overlap of A,B,C or even more sets?

ADD REPLY • link written 8.8 years ago by Ryan D • 3.3k

I'll offer you this link and the links therein: http://iphylo.blogspot.com/2011/03/some-vizbi-2011-links.html. No, I have nothing specific because I don't know how many data groupings you have and how many different overlaps. Hive plots (accessed from above link) might be a possibility. Or you could try overlapping lines, bending them as necessary to cover all overlaps.

ADD REPLY • link modified 3 months ago by RamRS ♦ 25k • written 8.8 years ago by Larry_Parnell ♦ 16k

8.8 years ago by

Chris Miller ♦ 21k

Washington University in St. Louis, MO

Chris Miller ♦ 21k wrote:

The Google Chart API will do this.

Update - the Chart API has now been deprecated (bummer...)

ADD COMMENT • link modified 5.9 years ago • written 8.8 years ago by Chris Miller ♦ 21k

That page doesn't show how I can get it to show the numbers of the overlaps too. Do you know if that is possible with Google Chart?

ADD REPLY • link written 8.8 years ago by Egon Willighagen ♦ 5.2k

Chris, that is great. It scales the sizes of the Venns and even lets me specify the size overlap. The resulting HTML is great. Still need to play with the options, but I think this will do the job.

ADD REPLY • link written 8.8 years ago by Ryan D • 3.3k

you could try Lucidchart - that's what I use for diagrams.

ADD REPLY • link modified 5.0 years ago • written 5.0 years ago by Waterlily_32 • 0

8.8 years ago by

Laurent • 1.6k

Cambridge, UK

Laurent • 1.6k wrote:

I can recommend Vennerable. It is quite flexible and will allow you to plot proportional Venn/Euler diagrams and define the set names (foo and bar below) and weights directly (bar only has 7 elements, their intersection has 9, and there are 12 unique elements in foo):

Venn(SetNames = c("foo", "bar"), Weight = c('00' = 0, '01' = 7, '11' = 8, '10' = 12))

If you want to stick with Perl, there is Venn::Chart, which I have never used though.

ADD COMMENT • link modified 8.8 years ago • written 8.8 years ago by Laurent • 1.6k

Thanks Laurent, I'm less familiar with R. The link above doesn't take me to Venn-Chart, but I gather this is it: http://search.cpan.org/~djibel/Venn-Chart-1.00/lib/Venn/Chart.pm

ADD REPLY • link modified 3 months ago by RamRS ♦ 25k • written 8.8 years ago by Ryan D • 3.3k

Sorry for the wrong url, I just updated it. Yes, your link is correct.

ADD REPLY • link written 8.8 years ago by Laurent • 1.6k

@Laurent I think in your example, it should be back ticks instead of single quotes.

ADD REPLY • link written 7.9 years ago by Sameet • 290

@Sameet You can indeed use back ticks, but single quotes do work.

ADD REPLY • link written 7.9 years ago by Laurent • 1.6k

If set A totally include set B, seems the weighted venn diagram not working?

ADD REPLY • link written 6.8 years ago by dli • 220

8.7 years ago by

Madelaine Gogol ♦ 5.1k

Kansas City

Madelaine Gogol ♦ 5.1k wrote:

Like others, I have started using The R VennDiagram package. I just presented on this at our local R user's group:

Powerpoint: Venn and Euler Diagrams in R

R code

For 3-way proportional venns, you could also try this Euler Venn Applet or Vennmaster:

ADD COMMENT • link modified 7.1 years ago • written 8.7 years ago by Madelaine Gogol ♦ 5.1k

I tried Vennerable and it's great too.

ADD REPLY • link written 8.3 years ago by Madelaine Gogol ♦ 5.1k

both links to Dropbox are broken!

ADD REPLY • link written 7.1 years ago by user56 • 290

Okay, I fixed the links. Sorry, switched to google drive. By the way, I recently used vennerable some more, and it will let you do 3-way area proportional venns (approx).

ADD REPLY • link written 7.1 years ago by Madelaine Gogol ♦ 5.1k

8.8 years ago by

Andreas • 2.4k

Singapore

Andreas • 2.4k wrote:

Not exactly sure if this is possible but it might be worth checking the following two references taken from Chen and Boutros BMC Bioinformatics 2011, 12:35 VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R:

Hulsen T, de Vlieg J, Alkema W: BioVenn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC Genomics 2008, 9:488.

Kestler HA, Muller A, Kraus JM, Buchholz M, Gress TM, Liu H, Kane DW, Zeeberg BR, Weinstein JN: VennMaster: area-proportional Euler diagrams for functional GO analysis of microarrays. BMC Bioinfo 2008, 9:67.

Andreas

ADD COMMENT • link written 8.8 years ago by Andreas • 2.4k

Interesting. In the discussion, the authors note: "During development of the VennDiagram package, it was discovered that it was impossible to draw accurate, scaled Venn diagrams with three sets using circles." Apparently this is only possible using elipses or polygons.

ADD REPLY • link written 8.8 years ago by Ryan D • 3.3k

Yeah, I ran into that same realization a few years back when I was trying to draw a three-fold venn. In some cases, it's possible, but if you get out a pencil and paper, it's pretty easy to come up with cases that don't work. For example, A,B,C,ABC=5, AB,AC,BC=0

ADD REPLY • link written 8.8 years ago by Chris Miller ♦ 21k

If A,B,C,ABC=5, then wouldn't AB,AC,BC also be 5?

ADD REPLY • link written 8.4 years ago by Farhat • 2.9k

8.8 years ago by

Ian ♦ 5.6k

University of Manchester, UK

Ian ♦ 5.6k wrote:

My current favourite (for two sets) is pyvenn (a python script).

I have a version of the script that i altered, which does not display any text. This is good for papers where you want to add your own text.

You may also be interested in this link and the associated comments.

ADD COMMENT • link written 8.8 years ago by Ian ♦ 5.6k

Ian, thanks. I read the OpenHelix post. It's where I got the links above from. But again it's within Cytoscape and not proportional.

ADD REPLY • link written 8.8 years ago by Ryan D • 3.3k

8.7 years ago by

Pascal • 40

Pascal • 40 wrote:

BioVenn is another tool for generating rea-proportional Venn diagrams. It supports only 2 or 3 comparisons, but might be handy anyway.

ADD COMMENT • link written 8.7 years ago by Pascal • 40

8.8 years ago by

Khader Shameer ♦ 18k

Manhattan, NY

Khader Shameer ♦ 18k wrote:

Have you tried Venny ?

ADD COMMENT • link written 8.8 years ago by Khader Shameer ♦ 18k

Thanks, Khader. Venny seems to be able to handle my data very well. While not proportional in size, like Chris's solution, it does give me numbers in each group by just pasting my CNV lists.

ADD REPLY • link written 8.8 years ago by Ryan D • 3.3k

Venny was down today... In any event, I would highly recommend I would highly recommend VENDIS:

http://kislingerlab.uhnres.utoronto.ca/projects/VennDIS_v1.0.zip

It just came out and it's really quite good!

ADD REPLY • link written 4.9 years ago by InterestedScientist • 30

8.8 years ago by

Zazo0o • 40

France

Zazo0o • 40 wrote:

Hi!

I don't know if it's possible to integrate R code into a suite of Perl suite, but to generate proportional and colorfull Venn diagrams I use a R package called VennDiagram. You can find it there : VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R

ADD COMMENT • link written 8.8 years ago by Zazo0o • 40

8.8 years ago by

Puthier • 250

Puthier • 250 wrote:

I think that the venneuler R package could do the trick.

library("venneuler")
vd <- venneuler(c(A=0.3, B=0.3, C=1.1, "A&B"=0.1, "A&C"=0.2, "B&C"=0.1 ,"A&B&C"=0.1))
plot(vd)

Depending on your needs, you can also have a look at the venn function in gplots package (not proportional but offers venn diagrams with up to 5 sets). You can get colors by slightly modifying the function.

#construct some fake gene names..
oneName <- function() paste(sample(LETTERS,5,replace=TRUE),collapse="")
geneNames <- replicate(1000, oneName())
GroupA <- sample(geneNames, 400, replace=FALSE)
GroupB <- sample(geneNames, 750, replace=FALSE)
GroupC <- sample(geneNames, 250, replace=FALSE)
GroupD <- sample(geneNames, 300, replace=FALSE)
input  <-list(GroupA,GroupB,GroupC,GroupD)
venn(input)

You may also look at the ClueGO cytoscape plugin (which is intended to summarize the results obtained from any functional enrichment analysis).It offers very nice and meaningful outputs.

Similar posts • Search »