Named rjson, appropriately. It’s quite basic just now, but contains methods for interconversion between R objects and JSON. Something like this:

> library(rjson)
> data <- list(a=1,b=2,c=3)
> json <- toJSON(data)
> json
[1] "{\"a\":1,\"b\":2,\"c\":3}"
> cat(json, file="data.json")

Use cases? I wonder if RApache could be used to build an API that serves R data in JSON format?

Recording my Ubuntu upgrade experience has become something of a tradition, so here goes.

• Machine 1 (laptop, LG T1 Express Duo). Largely trouble-free. On-board soundcard not seen on reboot. Fixed (temporarily, lost on reboot) using “sudo alsa force-reload”.
• Machine 2 (various generic hardware, cobbled together over many years). Upgrade smooth until final restart, when machine froze. Rebooted to a blank screen. Fixed by swapping out ATI video card for old NVidia FX5200. Discovered that rsyslog is running riot due to a hot CPU and is trying to fill up /var/log.
• Machine 3 (Dell Optiplex GX550). Install froze at “stopping winbind server”. Rebooted with rescue CD, mounted and chroot-ed into Linux partition, tried “dpkg –force-all –configure -a”, but to no avail. Reinstalled in same partition, all is working well.

All in all, not very impressed. Canonical, could do better.

Not work-related but fun: How to embed a flickr slideshow in WordPress.com using Vodpod. Vodpod is new to me and quite amazing; I imagine this method will allow you to post any Flash with embed code as a video.

As an example, a slideshow of Lane Cove images from my Flickr account.

Ubuntu/jaunty rocks, on the whole. However, to prove that newer does not always equal better, they threw in a couple of shockers. The first is the upgrade from the excellent Amarok 1.4 to the completely-broken 2.0. Head here to repair the damage.

Kile is by far my favourite LaTeX editor and suffered, though not as badly, in the upgrade from 2.0 to 2.1. Reports of various problems litter the web; in my case I see broken toolbar buttons that do nothing when clicked. This fix is much simpler. Just “sudo apt-get remove kile”, scroll to the bottom of the intrepid package page, choose your architecture, download the deb file and “sudo dpkg -i kile_2.0.1-1ubuntu1_i386.deb”. I had no dependency problems and it works just fine.

Finally – open up Synaptic, find kile and from the Package menu item select “lock package”, to prevent future upgrades.

Update: a few days after writing this, an apt-get dist-upgrade upgraded my Kile to 2.1 (despite the “lock package”). However, the broken buttons issue is now fixed for me. Go figure…

Since I haven’t posted for 14 days, what better (and lazier) way to post something than to surf over to a 14-day summary from the Life Scientists Group and link to the top ten items!

1. Review process files in the EMBO Journal – but why only for “the majority of papers”?
2. How XML threatens Big Data. Or not. How JSON might be an alternative – or not.
3. Solve any computer problem – with this classic XKCD flowchart.
4. Science reviews the revolution in ’strategic scientific reading’ – are they way behind the curve, or providing a useful summary for the uninitiated?
5. Best practice in microbial genome annotation – spirited discussion on the nature of best bioinformatics practice.
6. FriendFeed Life Scientists user survey – no further word on whether this will happen.
7. 50 Years of Structure – link to a JMB review on the early days of structural biology.
8. Reflections on Science Online London 2009
9. Workflow tools that speak SOAP?
10. Advice on cleaning up a protein sample – a nice example of useful discussion from the group.

Who knows, this could become a semi-regular feature.

I use Google Reader to subscribe to the RSS feeds from journals that interest me (see my public page). I’m also a big fan of CiteULike as a reference management system.

For a long time I’ve thought: it would be great if GReader handled journal articles more efficiently. Rather than going from link in GReader -> article at journal -> CiteULike bookmark -> back to GReader, how about “post directly from GReader?”

With Google Reader’s new send-to feature, you can do just that. See this forum post for the details. Also, take a look at this how-to for a quick way to post to CiteULike by entering a PubMed PMID, DOI or ISBN identifier in the address bar.

R is terrific, of course, for all your statistical needs. But those data structures! “Everything is a list.” Leading to such wondrous ways to access variables as “p <- Meta(gds)$platform", or "last <- mylist[[1]][length(mylist[[1]])]".

Sometimes, you want something more familiar. An array, a hash, a hash of arrays. Or, you may need to access R data in the language of your choice – e.g. as part of a Rails project.

In Ruby, IRB is your friend. On the right, an IRB session in which we invoke RSRuby, load the GEOquery library from Bioconductor, fetch a dataset from the GEO database and examine the metadata that describes the experiment. Result: a ruby hash of arrays, where the keys are covariate types (“sample, disease.state, description”) and the values covariate names for each column (sample) in the dataset. Now easy to access using:

columns.each_pair do |key,val|
  # do something with keys
  val.each do |element|
    # do something with values
  end
end

I work with people who generate a lot of microarray data. One question that they often ask is: can we find those genes with a two-fold or more change in median expression under two or more different conditions?

For example, let’s say that we have 3 conditions: “normal”, “adenoma” and “cancer”. That gives us 3 pairwise comparisons: normal-adenoma, normal-cancer and adenoma-cancer. Here’s a Ruby solution to the problem.

First, I installed StatArray, a Ruby gem that provides statistical methods for array objects. It’s not been updated for 3 years, but seems to work.

require 'rubygems'
require 'statarray'

Next, credit to David Burger for posting this code solution for combinations in Ruby. You give it an array of elements and the number of elements (r) that you want to see in each combination; it returns arrays of length (r) with each combination. I’ve just wrapped it in a class called Combination:

class Combination
  def generate_combinations(array, r)
    n = array.length
    indices = (0...r).to_a
    final = (n - r...n).to_a
    while indices != final
      yield indices.map {|k| array[k]}
      i = r - 1
      while indices[i] == n - r + i
        i -= 1
      end
      indices[i] += 1
      (i + 1...r).each do |j|
        indices[j] = indices[i] + j - i
      end
    end
    yield indices.map {|k| array[k]}
  end
end

So, let’s represent our covariates (normal, adenoma, cancer) as hash keys and their expression values, from several samples, as arrays which are the hash values. Then, we create a new hash with the same keys where the values are medians, calculated using methods from StatArray. Here, I’m using very silly dummy values for expression, which are supposed to be log base 2 values:

values = {"normal" => [1,2,3,4], "adenoma" => [5,6,7,8], "cancer" => [9,10,11,12]}
median = Hash.new

values.each_pair {|k,v|
  median[k] = v.to_statarray.median
}

I’m sure that there’s a more elegant way to map the values hash to the medians hash, but that’s for another day.

Finally, we generate all possible combinations of two covariates, subtract their median values and convert to an absolute value. If we’re using log base 2, then a change of one unit = a two-fold change in expression:

c = Combination.new
c.generate_combinations(median.keys,2) do |x|
  puts "abs(#{x.join(" - ")}) = #{(median[x[0]] - median[x[1]]).abs}"
end

Result:

abs(normal - adenoma) = 4.0
abs(normal - cancer) = 8.0
abs(adenoma - cancer) = 4.0

Fin.

A couple of months ago, I wrote a brief guide to displaying R plots in a Rails application using RSRuby.

It worked at the time – really it did – but since then, I’ve encountered problems. One is that despite sending the plot output to a PNG file, R X11 windows started to pop up on loading the web page containing the plot. Another is the appearance of long error messages related to “stack smashing” on terminating Mongrel server.

Fortunately, I read an excellent guide by Ana Nelson, R on Rails with RSRuby, which convinced me that I was doing things all wrong. So here is an amended version of my R + RSRuby + Rails “Hello World” application.

We start as usual with “rails Plotter”. Next, I like to delete the public/index.html file and generate a controller named “home” with a single action, “index”.

rails Plotter
cd Plotter
rm public/index.html
./script generate controller Home index

Edit config/routes.rb so as root points to this controller and action:

ActionController::Routing::Routes.draw do |map|
  map.root  :controller => "home"
end

Right, let’s get on with the R integration. Create two new directories, lib/R and public/images/R. Edit a file in lib/R named “plot.R”:

myplot = function() {
  png(filename = "public/images/R/plot.png")
  boxplot(c(1,2,3,4),c(5,6,7,8))
  dev.off()
}

This defines a function, myplot(), which generates a toy boxplot from two sets of values and writes it to the file public/images/R/plot.png.

The file helpers/application_helper.rb, as before, defines the environment variable R_HOME, required for RSRuby to work:

module ApplicationHelper
# set R_HOME if not set
  if ENV['R_HOME'].nil?
    ENV['R_HOME'] = "/usr/lib/R"
  end
end

The file controllers/application_controller.rb defines a method named “get_r”. This method initiates an instance of RSRuby, reads the R script lib/R/plot.R and returns the instance variable:

class ApplicationController < ActionController::Base
  def get_r
    require 'rsruby'
    @@r = RSRuby.instance
    @@r.eval_R("source('lib/R/plot.R')")
    return @@r
  end
end

We’re all set. In the Home controller, controllers/home_controller.rb, we can call the get_r method and then, call the R function, myplot(), on the RSRuby instance:

class HomeController < ApplicationController
  before_filter :get_r

  def index
    @@r.myplot.call
  end
end

This writes the R boxplot to a PNG file. The view, views/home/index.html.erb, is simple:

<h1>Home#index</h1>
<p>Find me in app/views/home/index.html.erb</p>

<%= image_tag("R/plot.png") %>

Fire up Rails with “./script/server”, open “http://localhost:3000″ in a browser and you should see a view like the one shown to the right.

R boxplot in Rails via RSRuby

Next steps: get some variables from Rails into the R plotting function and do something real.

A brief and unscientific survey of social bookmarking buttons provided by journal websites. Take with pinch of salt and/or tongue in cheek.

PLoS get it. As you'd expect.	BMC get it too.	Even ACS kind of get it.
OUP get it to a degree.	Someone doesn't get it.

I know, all bookmarking services provide their own tools, such as bookmarklets. However, I suspect that a button is more convenient for many people: it’s a courtesy and an advertisement for social bookmarking and online reference management tools.

So NPG: is it “we promote sharing” or “we promote sharing so long as you use our service?”