4 min read

Week33 - RBioFabric

Week33 - RBioFabric

This Week’s Widget - RBioFabric

William Longabaugh (@wjrl) developed a novel hairball-less way to visualize networks called BioFabric in his 2012 paper.

Longabaugh, W.J.R. Combing the hairball with BioFabric: a new approach for visualization of large networks. BMC Bioinformatics, 13:275, 2012.

Since then, he has done an incredible amount to speed adoption by implementing BioFabric in Java, JavaScript (d3.js), and R. His work, numerous examples, and discussion convinced me of the power of BioFabric, and I have been waiting for a week that I could commit enough time to attempt an implementation as an htmlwidget. When I saw that the genius prolific node.js programmer Max Ogden @maxogden had taken a first step toward extending and generalizing @wjrl’s d3.js code, I decided to see if I could combine the R and JavaScript into an htmlwidget.

As I plunged through the code, I decided to approach this by using the much fuller featured bioFabric plotting function in R with all the power of igraph to do most of the data conversion and preparation in R. I stripped this from the JavaScript, and instead use d3.js just to render the graph and layout. By far, the Java version is the fullest implementation of BioFabric with lots of helpful interactivity and analysis tools. Using it for inspiration, I added crude pan/zoom and then some mouseover interaction in the JavaScript. There is still a whole lot to do, and I would love help, ideas, comments, and suggestions from anyone willing or interested.

I want to make sure I at least spend another sentence commending and thanking William Longabaugh (@wjrl) for not only his brilliant idea but his incredible commitment to it. I can only hope that this little htmlwidget will help bring a little bit more attention and help spread the word.

Quick Installation

RBioFabric is not yet on CRAN, so for now please install with devtools::install_github.



Let’s start simple with some igraph data. It is beyond the scope of this post to explain BioFabric. I highly encourage enjoying all of William’s incredible resources on BioFabric to get a better sense of how to interpret these special diagrams.

Bridges of Koenigsberg

# devtools::install_github("timelyportfolio/RBioFabric")


#  strange but we actually use the bioFabric plot function
#   to give us the data

bioFabric_htmlwidget( bioFabric( Koenigsberg) )

Ordering Nodes

By default, nodes are sorted by degree. However, bioFabric allows custom sort order through a function with orderFun or as a vector with userOrder.

# devtools::install_github("timelyportfolio/RBioFabric")

# ?fastgreedy.community

fc <- fastgreedy.community(karate)

# let's sort by community and then degree
    , userOrder = 
      data.frame(membership = membership(fc), degree = degree(karate)) %>%
        mutate( id = rownames(.) ) %>%
        arrange( membership, desc( degree ) ) %>%
        select( id ) %>%

Les Mis

No network visualization is complete unless it now includes Les Mis. It’s a little small in the blog, but this will give us a good opportunity to try out the pan and zoom.

# devtools::install_github("timelyportfolio/RBioFabric")

# d3 example from BioFabric
#  source:   https://github.com/wjrl/D3BioFabric
#  example:  http://rawgit.com/wjrl/D3BioFabric/master/src/JustBioFabric.html

miserables <- jsonlite::fromJSON(

mis_igraph <- graph.data.frame(
  d = miserables$links
  ,vertices = data.frame(
    id = as.character( 0:(nrow(miserables$nodes)-1) ) 
    ,name = miserables$nodes
    ,stringsAsFactors = F

bioFabric_htmlwidget( bioFabric( mis_igraph ))


Thanks so much to William Longabaugh (@wjrl) for a brilliant idea and his commitment to it.

As always, thanks to

  • Ramnath Vaidyanathan and RStudio for htmlwidgets
  • all the contributors to R and JavaScript