Digital Science

Annotate All The Things: iAnnotate 2013 Round Up

Wed, 29 May 2013 16:29:00 +0000

Last month, the Fort Mason Center in San Francisco was host to a gathering of publishers, editors, technologists, and geeks, all with an interest in the latest developments in Web-based annotation. Over the course of three days attendees of the iAnnotate conference learned about the latest applications of annotation technologies, new standards in the field, and other ideas all related to annotation of content on the World Wide Web.

I was there representing Digital Science, and gave a presentation all about automated annotation - you can find a video recording of it here .

In case you're wondering, "Why should I care about annotation?", it's worth noting that the creators of the World Wide Web - researchers at the CERN lab in Geneva - always envisioned a read-write web, where documents could be retrieved, but also edited. The success of Wikipedia shows why this idea is important: instead of just a one-way flow of information from publisher to consumer, Wikipedia provides a rich ecosystem of community participation and collaboration around human knowledge. The promise of web-based annotation is to bring that same rich interactivity, but to ALL content on the web.

Below, I'll pick out my stand out highlights of the conference. These are entirely subjective, so I encourage you to check out the other presentations published on the conference video channel too!

RapGenius

First up: the presentation on RapGenius , a wonderful example of community adoption of web-based annotation. RapGenius is a hybrid of a fan-site for collecting lyrics for rap songs, and a way for fans to collaboratively define the meaning behind those lyrics.

It’s been adopted by the fan community with great vigour and attracts about 10m unique visitors a month. Just pick a song and click on a lyric and you'll find a rich insight into the meaning of the song, and the history it draws from. While perhaps not the most technically advanced or feature rich application (at the time of writing), RapGenius demonstrates what’s possible with the right level of community commitment to annotating content.

Hilbert Problems of Annotation

My next pick is the presentation on the four Hilbert problems of annotation , given by the lead developer of the Annotator JavaScript library . While the presenter, Nick Stenning, did a convincing job of selling the benefits of using Annotator, the main thrust of his presentation was to provide a framework for understanding the challenges facing the adoption of annotation, and to make a call to action for the community.

Hilbert problems are particularly difficult, unsolved problems from the world of mathematics, described by German Mathematician David Hilbert in 1900. In the world of web-based annotation the problems relate to content rather than mathematical theories. They are:

Discovery - i.e. what annotates any given piece of content?
Documents, not formats - i.e. we need canonical document representations
Changing content - i.e. what exactly am I annotating?
Persistence of reference - i.e. can we put the internet in git, and make append only changes?

Some progress, you could argue, has already been made on the last two of these problems, in the form of the Authorea, and the W3C Open Annotation standard, my final two picks.

Authorea

Authorea is one of several examples of general purpose online annotation / authoring systems, many of which were represented at the conference. They each have their particular focus, whether that's a seamless mobile experience ( DotDotDot ), a minimalist approach for collaborating around prose ( Poetica ), or an understanding of scientific authoring ( Authorea ).

As an employee of Digital Science, I'm naturally drawn to Authorea ... but in addition to science-focused features such as LaTeX authoring , automatic reference resolution from HTTP DOIs , and collaborative authoring, Authorea provides version management using GitHub . This final feature is an example of the kind of foundational shift hinted at in the final Hilbert problem - persistence of reference.

If you haven't encountered it before, Git is the open source code management system developed by Linus Torvalds to handle community contributions to Linux kernel . It provides a de-centralized way for developers to share code, and to collaboratively edit that code. Crucially, all changes are additive, so a complete history is available for resources held in a git repository. The GitHub web site provides hosting for git repositories, along with a large array of helpful features that make collaborating around Git repositories easier.

Hopefully in time, more and more annotation systems and tools will find a way to harness the power of git.

Shared Canvas / W3C Open Annotation

Rob Sanderson’s presentation on the Shared Canvas and W3C Open Annotation is my final pick from iAnnotate 2013. The presentation only gives a taster of what this new standard is about, but at the hackday on Friday afternoon Rob provided a detailed run through of the standard's rationale, practicalities, and applications.

In summary, the standard provides a flexible data model for describing web annotations, along with simple examples (provided in JSON format ) that are both Web and linked-data friendly.

The model includes things like motivation (e.g. description, classification), options for how to identify and locate what is annotated (e.g. text positions, fuzzy text matching, image coordinates ...), as well as precise indicators for what resource is being annotated (think request headers) and when it was annotated (think timestamps).

Apart from providing a rich data model to inspire and guide the annotation community, the standard has a clear, precise, but non-prescriptive foundation in RDF (the Resource Description Framework , a collection of linked data standards). The standard has been created with the web in mind, so annotations can be expressed in a simple, efficient, intuitive JSON form, while still being mappable to the more formal RDF for use in further linked data applications, analyses, and data stores.

It's still early days for this standard, but it shows the signs of being a solid basis for supporting interoperable annotation on the web.

Other Highlights

A few other interesting picks include:

Ian Mulvaney's presentation on Threads , an innovative way of linking and combining journal content. Or, as Ian puts it “A Storify for Science”.
A workshop all about annotating new government bills, as a way of gathering elector feedback before approaching the legislature.
A wonderful example of real world annotation in the form of Book Sprints, where readers collectively read and digest interesting books

All in all, iAnnotate was a fascinating conference and attracted a large number of interesting speakers. I encourage you to attend if you can next year.

Introducing Projects – Digital Science’s first home-grown tool

Wed, 22 May 2013 07:58:00 +0000

Digital Science is proud to announce the launch of its very own home-grown software tool – Projects.

Projects is a safe, simple and structured way for scientific researchers to organise their research outputs. It helps to keep track of experiments, save time and get ahead with research.

The academic research community has repeatedly told us that their information tools are inadequate, making it hard to organise their data files effectively. This unmet need has been the driving force behind the development of Projects.

“A lot of the current problems in research data management stem back to the way in which academics organise the outputs of their research. There is rarely consistency within research groups, let alone between groups, meaning other researchers cannot find files when trying to look through the research of previous researchers”, says Mark Hahnel, Product Owner of Projects and Founder of figshare.

We try to be agile with our tools at Digital Science and we want people to start benefiting from using Projects as soon as possible, so we’ve released it on Mac only to begin with. We’ll be releasing Windows and Linux versions as fast as we can, along with other features including sharing, collaboration and integration with other research tools.

Using Projects you can:

View a unique chronological visual timeline of your research activity and more easily find important files
Manually or automatically take “snapshots” of your work to seamlessly restore previous versions of files
Store your data files in a simple organised way within a handy desktop drive, which can be easily shared or synchronised with remote backup services
Flag, star and take notes on your data files to distinguish important files and to optimise search results

This is just the start. We have plans to:

Link to open access repositories, such as figshare, to ensure every scientific researcher gains credit for 'all' their research
Make remote backups easier for an even greater peace of mind
Make it easier to share and collaborate with others on projects and experiments

So get involved - download Projects today and get a free 30-day trial. As always, we greatly appreciate your feedback so please get in touch via projects.ac

Riemann On Rails (Part Two)

Thu, 02 May 2013 11:46:00 +0000

In a previous post we looked at how to generate metrics in a Ruby on Rails application and how to pass this data into Riemann. In this post we will cover some basics of configuring a Riemann server as well as how we can go about processing our event streams.

Why Riemann

Riemann works on streams of metrics. Think of Riemann more as an event or metric processor than just a simple aggregator / collector. The Riemann concepts page gives a pretty good 'top level' overview of some the core concepts in Riemann. The idea of a stream in Riemann is summed up pretty nicely in the Riemann introduction:

Riemann streams are just functions which accept an event. Events are just structs with some common fields like :host and :service You can use dozens of built-in streams for filtering, altering, and combining events, or write your own. Riemann.io

As our metrics are exposed as functions, we can tie functions together or pass functions into other functions to turn them into actionable events or triggers.

Clojure

Riemann's stream handling logic is written in the Clojure programming language. Clojure is a functional based programming language written on top of the JVM. If you're new to Clojure I recommend the excellent resources available on the Clojure documentation site to help you get started.

Riemann configuration

Riemann's configuration file is written in Clojure, the Riemann example configuration file is well commented and good place to start. This file contains configuration for the Riemann server as well as your stream processing logic. Depending on how you installed Riemann and what platform you're running the Riemann server on, the location on your configuration file may vary. If you installed Riemann server on OSX via homebrew it's likely to be in the following location:

/usr/local/etc/riemann.config

Open this file in your editor of choice and take a look around, it may differ slightly to the example linked to above. When making changes to this configuration file you will need to tell Riemann to reload once you're finished:

kill -HUP <riemann server process id>

This will reload the configuration without having to reboot the entire Riemann server.

Riemann Hello World

In your Riemann configuration file, you should see a call to 'streams', this is the starting point for our custom processing logic. Functions should be defined above the call to 'streams'. We'll start by printing every metric received to the log file with a hello message:

(streams
	(fn [event] (println "hello from: " (get event :host))))

For each event received, we're calling println with the argument 'event'. We are then fetching the :host attribute from the event and printing this to the screen.

Email delivery failures

In part one we defined some metrics that will be sent to Riemann for processing. One of them was tagged with 'email_send_runtime' and was collected from the "deliver.action_mailer" event.

In our Riemann configuration file, first we define a function called 'mail-support' which when called, will send an email to our support team:

(def mail-support (email "support@your.company.com"))

'email' is a function built into Riemann, see the riemann.email api docs for more information. You can also use the 'rollup' to combine multiple events alerts into a single email.

The mailer can accept lists of events, too. To avoid getting slammed with too many emails we can use the rollup function--it will combine multiple events into a single message. To send at most 5 emails every hour:
(def tell-ops (rollup 5 3600 (email "ops@vonbraun.mil")))
Riemann example config

Next, we call our 'mail-support' function for all events from the 'rails-app' service which are tagged 'email_send_runtime' and have the state 'critical'.

(where 
  (and (service "rails-app") (tagged "email_send_runtime") (state "critical") 
    (mail-support (str "rails-app on " host " email delivery critical: " description) ))))

'description' and 'host' are attributes provided by the metric that is passed in from the where function.

Slow database queries

When we're defining functions in Riemann, it's likely we may want to use a similar function for multiple different metrics. Here we define a new generic function called 'metric-threshold-alert' which will email our support team if a given metric exceeds a set threshold:

(defn metric-threshold-alert [service-name tag-name threshold_ms]
  (where 
    (and (service service-name) (tagged tag-name) (>= metric threshold_ms) 
      (mail-support (str service-name " on " host " " tag-name " is " metric "ms") ))))

'[service-name tag-name threshold_ms]' defines the arguments for this function. We then filter events using 'where' to limit events based on the arguments provided.

With this function defined, we can then call this for multiple services, tags or threshold values.

Email our support team if database query execution time exceeds 200ms:

(metric-threshold-alert "rails-app" "db_runtime" 200)

Email our support team if request execution time exceeds 1000ms:

(metric-threshold-alert "rails-app" "request_runtime" 1000)

Alerts based on state of metrics

We might want to alert our support team if we get a certain percentage of warning events within a given time frame. This might be useful for alerting your support team if your application starts experiencing a high percentage of failed sign up or sign in requests. Riemann provides the 'fixed-time-window' function which we can use to accomplish this. First, we'll define a reusable generic function called 'state-percentage-alert':

(defn state-percentage-alert [service-name tag-name]
  (where (and (service service-name) (tagged tag-name)
    (fixed-time-window 60
      (smap (fn [events]
        (let [percent (/ (count (filter #(= (:state %) "warning") events)) (count events))]
          (cond
            (> percent 0.7)
              (mail-support 
                (format (str service-name " " tag-name " is CRITICAL %f") (float percent) ))
            (> percent 0.5)
              (mail-support 
                (format (str service-name " " tag-name " is BAD %f") (float percent) ))))))))))

'[service-name tag-name]' defines the arguments for this function. We then filter events based on the service and tag provided as arguments. 'fixed-time-window' collects events passed in from the 'where' function and passes them into 'smap' every 60 seconds. We then define the variable 'percent' as the percentage of metrics with state :warning. Finally 'cond' sets up some conditions to take action based on this percentage. If the percentage is greater then 50%, send a warning email. If the percentage is greater than 70%, send a 'critical' email.

Then we can configure alerts for events tagged 'sign_up' as follows:

(state-percentage-alert "rails-app" "sign_up")

Wrapping Up

Riemann provides a variety of stream processing functions that can be used to your advantage. It's worth exploring the Riemann API docs for more ideas of what's possible. Riemann supports more than just notifications via email out of the box, it can be configured to push metrics to graphite or send alerts via pagerduty. As the configuration file is Clojure you could add your own outputs too!

Google Glass for researchers

Wed, 24 Apr 2013 09:14:00 +0000

The engineering group at Digital Science have recently become proud members of the Google Glass Explorers programme thanks to our small pitch on Google+.

In anticipation of receiving our Google Glass unit in the next couple of months (hopefully), we want to reach out to the wider scientific community for ideas for Glasses applications within research. We already have a few ideas bubbling away, and are preparing the engineering team for a Glasses themed hack break this summer, but would love to see what other ideas are out there that we might prototype/build.

We'll keep this blog updated on when we'll be receiving our Glasses, but in the meantime if you have any ideas let us know by commenting on this post, or posting to twitter with the hashtag #dsciglasses.

We're already exploring ideas using computer vision and laboratory assistants, but the more the merrier.

Riemann On Rails (Part One)

Tue, 23 Apr 2013 09:05:00 +0000

Application metrics can provide a useful insight into the inner workings of your application. A metric might be the time taken to run a database query, how long a user stayed on one of your pages or generated when an exception is raised. We could have sales based metrics, advertisement interaction metrics and so forth. Using these metrics, we can build tools for alerting or graphing to help keep us informed on how our application is performing. To quote the excellent Etsy engineering blog, Measure Anything, Measure Everything. As it turns out, Ruby on Rails 3.0+ ships with some useful metrics out of the box that we can take advantage of. In this post I describe how to combine Rails notifications with Riemann to build a kick-ass application monitoring solution.

Enter ActiveSupport::Notifications

ActiveSupport::Notifications provides a publisher / subscriber system inside your Rails application. Publishers can push information onto a given 'channel' which subscribers then subscribe to. Many existing Rails features have been written to publish information about their operation to a number of channels. A comprehensive list of available notifications available in Rails can be found here.

If you have ever looked at the log file of a Rails application you may have noticed something like the following:

Completed 200 OK in 1337ms (Views: 482.3ms | ActiveRecord: 28.7ms)

The timings displayed in the above snippet are the result of the application logger subscribing to a channel containing events from your controller.

Subscribing to Events

In order to make use of the events generated by ActiveSupport::Notifications you will need to subscribe to one or more channels. The following code example shows how you would subscribe to a channel called 'channel':

ActiveSupport::Notifications.subscribe "channel" do |name, start, finish, id, payload|
  # your event handling code here
end

Every event contains a 'payload', which is a hash of information populated by the publisher. Based on the contents of this payload your event handling code could email your support team if certain conditions are met, send the remote IP to a geckoboard dashboard or pass the event onto another service such as StatsD. ActiveSupport::Notifications ships with a 'fan out' strategy for event propagation so you can have multiple subscribers to the same channel.

Two of the more interesting channels provided by Rails are "process_action.action_controller" and "deliver.action_mailer" which we'll run through briefly here.

process_action.action_controller event payload

{
  controller: "PostsController",
  action: "index",
  params: {"action" => "index", "controller" => "posts"},
  format: :html,
  method: "GET",
  path: "/posts",
  status: 200,
  view_runtime: 46.848,
  db_runtime: 0.157
}

The above event is generated when a user visits the PostsController's index action and requests a HTML response. Similar events will be generated for any action in a Rails controller, regardless of format. Using this event alone, we've already got information on database query execution time, HTTP status codes and view rendering time. We can also extract total request time from this event by using the supplied start and finish times:

c = "process_action.action_controller"
ActiveSupport::Notifications.subscribe c do |name, start, finish, id, payload|
  request_time = (finish - start) * 1000 # full request time in ms
end

deliver.action_mailer event payload

{
  mailer: "Notification",
  message_id: "1234@mba.local.mail",
  subject: "New Post Notification",
  to: ["list@my-app.com"],
  from: ["no-reply@my-app.com"],
  date: Fri, 19 Apr 2013 16:00:00 +0100,
  mail: "..." # omitted for brevity
}

ActionMailer is the library that ships with the Rails framework for sending emails to your users. The above event is triggered when ActionMailer attempts to deliver an email. Using the start and finish times provided, we can calculate the delivery time for this email:

c = "deliver.action_mailer payload"
ActiveSupport::Notifications.subscribe c do |name, start, finish, id, payload|
  delivery_time = (finish - start) * 1000 # email delivery time in ms
end

Exception notifications

ActiveSupport::Notifications for events generated by Rails also provide information about any exceptions that are raised during a request or when delivering an email. While they do not provide any backtrace information we can still use this event to take action if required. The following example shows an event generated by 'process_action.action_controller' when an exception is raised inside the Devise::PasswordsController.

{
  :controller=>"Devise::PasswordsController", 
  :action=>"create", 
  :params=>{..}, 
  :format=>:html, 
  :method=>"POST",
  :path=>"/users/password", 
  :exception=>["Errno::ECONNREFUSED", "Connection refused - connect(2)"]
}

Defining your own notifications

You can use ActiveSupport::Notifications to generate your own event stream with a few lines of code. The following is an example of how to generate a custom event indicating authentication failure:

# your sign in logic
if authentication.failed?
  event_params = {
    :remote_ip => request.remote_ip, 
    :state => "warning", 
    :description => "invalid password"
  }
  ActiveSupport::Notifications.instrument("auth.sign_in", event_params)
end

ActiveSupport::Notifications can also be used for generating timing events. This is be achieved by wrapping the code you wish to time in a block:

event_params = {
  # what ever you want to include!
}
ActiveSupport::Notifications.instrument("slow_API", event_params) do
  # the code inside this block will be timed
end

When using the instrument method's block syntax, the 'start' and 'finish' arguments will be populated in your subscription handling code as shown earlier.

Now that we are generating and subscribing to all these events, we may wish to process them further in order to trigger alerts or generate graphs in graphite. This is especially useful if you have multiple application / tiers and want to aggregate all of your metrics in a single place. Several options are available, a popular choice being StatsD. At Digital Science, we've been experimenting with using a different event aggregator called Riemann that provides powerful event processing logic out of the box.

Riemann

Riemann aggregates events from your servers and applications with a powerful stream processing language. Send an email for every exception raised by your code. Track the latency distribution of your web app. See the top processes on any host, by memory and CPU. Combine statistics from every Riak node in your cluster and forward to Graphite. Send alerts when a key process fails to check in. Know how many users signed up right this second. Riemann.io

Riemann differs from other 'event aggregators' such as StatsD by offering inbuilt functions for processing your event streams as opposed to being a simple metrics aggregator.

Sending events to Riemann

Riemann clients are provided in a variety of languages and require very little configuration to start pushing events. Riemann events can be transmitted by UDP or TCP, depending on your requirements. The example given uses UDP transmission which is effectively 'fire and forget'.

client = Riemann::Client.new(host: 'localhost', port: 5555)
event = {
  host: "mba.local",  # or `hostname`.strip
  state: "ok",        # ok, warning or critical
  metric: 123,        # the value of your event / metric e.g time in ms
  ttl: 10,            # how long Riemann should keep the event in it's index for
  tags: ['sign_in'],  # any tags you want to associate with the request
  service: 'my-app'   # the service (or app name) that is generating the metrics
}
client << event

You might want to consider a simple wrapper class that handles populating default tags, such as the environment the metric comes from or the hostname.

class RiemannClient
  TTL = 10

  def initialize config
    @host = config['host']
    @port = config['port']
  end

  def client 
    @riemann_client ||= Riemann::Client.new(host: @host, port: @port)
  end

  def gauge tags, state, metric, description=nil, service='my-app-name'
    event = {
      host: `hostname`.strip,
      state: state,
      metric: metric,
      ttl: TTL,
      tags: (tags.dup << Rails.env), # auto tag events with Rails env
      service: service
    } 
    event[:description] = description if description
    client << event if @host && @port
 end
end

This makes the code in our ActiveSupport::Notifications.subscribe blocks considerably less verbose.

deliver.action_mailer

client = RiemannClient.new({"host" => "localhost", "port" => 5555})
c = "deliver.action_mailer"
ActiveSupport::Notifications.subscribe c do |name, start, finish, id, payload|
  tags = [payload[:mailer],'email_send_runtime']
  state = !payload[:exception].nil? ? 'critical' : 'ok'
  delivery_time = (finish - start) * 1000

  client.gauge tags, state, delivery_time, payload[:exception]
end

By checking for the presence of :exception in the payload, we can flip the state we pass onto Riemann making our alerting process easier to implement. This is because we can take direct action based on metric state being 'critical' rather than trying to figure this out in our Riemann stream processing.

process_action.action_controller

c = "process_action.action_controller"
ActiveSupport::Notifications.subscribe c do |name, start, finish, id, payload|
  tags = [payload[:controller], payload[:action]]
  state = !payload[:exception].nil? ? 'critical' : 'ok'
  client.gauge (tags.dup << 'http_status'), state, payload[:status]
  client.gauge (tags.dup << 'view_runtime'), state, payload[:view_runtime]
  client.gauge (tags.dup << 'request_runtime'), state, (finish - start) * 1000
  client.gauge (tags.dup << 'db_runtime'), state, payload[:db_runtime]
end

In the above example we are generating four events to send to Riemann, all provided in a single "process_action.action_controller" event. So with only a few lines of code, we're ready to alert on email delivery times, email exceptions, request times, view times, full request times, slow queries, http statuses and pretty much anything else you can think of. Neat.

In part two, we'll write some Clojure to process these event streams in Riemann!

Pi-day: When Justin Bieber met Newton

Wed, 03 Apr 2013 07:27:00 +0000

OR: What happens when you don't define the shape of your pie

Last month heralded Pi Day, in celebration of the mathematical constant and irrational number Pi.

To mark the occasion at here at Digital Science, we held an informal math challenge. Don't be scared to keep reading if, like me, mathematics isn't your first language. The only equations here involve Justin Bieber, pie filling, and tenuous but amusing proofs.

The challenge was as follows:

"You have exactly 400 x Pi cubic centimetres of the finest pie mix [about 1,250 cm cubed], but you don't have much pie crust mix and you want to use the smallest amount possible. What are the best dimensions to use for your pie, to use the smallest amount of crust? Hint: check out the SureChem logo. Double hint: look at both the shape and purpose of the logo."

...and the prize at stake was various tasty pies from a local bakery.

For the more math-savvy, the problem isn't too hard to work out - it's an optimization problem, and it can be solved using some calculus.

Unfortunately, maths problems require enough information so they can be solved correctly, and despite a few hints pointing to the shape of the SureChem logo (a Digital Science product), it quickly became apparent that there wasn't enough information about the shape of the pie. The resulting interpretations of the problem from our colleagues were far more interesting than the actual answer.

Answers included:

A fairly logical conclusion: "At the moment the answer is that the pie is spherical with radius 6.69433cm (or 2.13Pi cm, if you like)."

A serendipitous and delicious invention: "Assuming a [company] canteen-style pie, where you basically get some pie mix, add mystery meat, and then top it off with a crust. So, the problem becomes that of minimising the area of the crust on top. Clearly then, the best pie shape is an infinitely long cylinder with an infinitesimal speck of crust on the 'top' - although I guess it's a philosophical question as to whether an infinitely long cylinder has a top. This will not only be delicious, if we spin the cylinder it will also act as a convenient time machine."

Some helpful housing advice: "Alternatively, create a spherical shell of pie filling, with the hollow centre as small as possible. Define yourself to be inside the pie, making the hollow centre the outside and coat that in pie crust. Depending on how dextrous you are, the amount of pie crust required can be infinitesimally small. This method shows how you can reduce the costs of living by careful use of pure mathematics."

A practical refinement: "Your infinite pie crust solution doesn't work, because to be pie crust, there necessarily has to be things that make it pie crust, i.e., flour, water, butter, or a.) it will have to be made of some exotic matter, and b.) it will be too soggy. In order to minimise pie crust area while still getting a delicious golden brown finish, we will stack these ingredients in the correct ratios in a pile one atom across. Therefore the pie lid will be one chlorine diameter across as I think chlorine in the pinch of salt is the biggest atom in pie. My periodic table app keeps force closing."

Also an idea for a Stephen Hawking-inspired episode of a home improvement show: "Alternatively, you could set up a kitchen inside of a black hole, in which your pie mix will be compressed to an absolutely minuscule volume. You then have to spread the pie crust mix as thinly as possible to cover everything. If it still isn't enough, you just move deeper into the black hole. There's one caveat though - you have to consume your entire pie inside the black hole as well - if you wanted to bring it back out of it (which is no mean feat I tell you) - the entire pie will explode..."

A few observations about the precise definition of a pie cropped up too. One colleague was keen to point to the Wikipedia page for pie, which indicates that pies are “defined by their crust”, thus rendering the entire competition null and void because critical information was omitted from the description. It was also suggested that the shape of the pie in the correct answer was actually more or a "deep-pan tart". Finally, one helpful insight was that "if they can get away with making "beef" lasagne with 100% horse, I'm sure we can make a pie (of any shape) without using any actual crust."

There were also cartoon and picture submissions including The Wife of Pi, The Life of Pi, and something of a wierd coincidence.

The pie-winning entry, for submitting the correct answer before anyone else and for the funniest submission came from a member of the Digital Science finance team. The correct answer is a pie with both radius and height of 7.368cm, which was skillfully written up with the help of Justin Bieber and writelatex in a groundbreaking mathematical proof.

If you're wondering where the correct answer comes from, check out this video by Jim Fowler of Ohio State University. It's for the "soup can" problem, which the pie problem is closely based on.

Thanks to everyone for submitting their answers, and to Jim Fowler of The Ohio State University and Coursera, both for running an amazing Calculus course and for providing the inspiration for the Pi-day challenge.

Lesson learned: Always define the shape of your pie.

Attribution: Pi-pie photograph by Paul Smith.

Digital Science and SciBite form strategic partnership for life sciences text mining

Tue, 26 Mar 2013 13:51:00 +0000

Some exciting news for Digital Science on the text mining front. We recently signed a strategic partnership agreement with SciBite, a UK startup that has put together an impressive scientific news and alerts service focused on drug discovery. SciBite is highly complementary to SureChem, our patent chemistry search offering, and we will be looking at ways of quickly combining them, particularly the direct data API services they both offer.

But this partnership is also significant in that it marks the expansion of Digital Science text mining beyond patents and chemistry to other data sources and biology. This agreement gives SciBite and Digital Science access to each others annotation and data integration technologies and content, enabling us to create a world class life sciences text mining pipeline.

In addition to SureChem and SciBite, there are multiple Digital Science products that would benefit from this capability and as you might imagine, there are plenty of conversations under way already.

You can read more about the partnership on our press release page here. For more, follow us at @digitalsci, and do stay tuned for more in coming months.

A new look (and features) for Labguru

Tue, 05 Mar 2013 12:54:00 +0000

Congratulations to our colleagues at BioData who today pushed a significant upgrade to Labguru, a web-based app that helps scientists better manage their research and more broadly, their lab.

The upgrade features a new, more minimalist, organised interface, providing easier access to the following features:

Simpler menus and navigation, enabling users to keep track of their location deep within projects or documents.
A live feed of the lab’s work to encourage deeper awareness of colleagues’ research and opportunities to collaborate or share knowledge.
A prominent whiteboard for sharing important messages with the entire lab.
An improved project view enhances scientific thinking, not just efficient experiment planning.
A more comprehensive storage module to help everyone quickly locate and track items in the lab.
Streamlined experimental protocol editing, archiving, and syncing with the free bench-side iPad app. This works for both custom protocols and preloaded ones from the journal Nature Protocols and vendors such as New England Biolabs, Sigma Life Science, and Luminex.
Bulk tagging and organization tools for lab inventory, helping the lab buy supplies on-time and avoid duplicate orders. Labguru also has a built-in catalog of more than 800,000 products to simplify tracking inventory digitally.

For more on the release, see the press release or visit Labguru's web site. Already a Labguru user? We'd love to hear what you think.

Wiley makes PDFs interactive with ReadCube Web Reader

Wed, 06 Feb 2013 16:29:00 +0000

It's been a busy few weeks for our colleagues at Labtiva, creators of ReadCube, who today launched the ReadCube Web Reader on content from John Wiley & Sons, a global provider of content in scientific, technical, medical and scholarly research. The web reader will be available on their online library, and will provide users hyperlinked in-line citations, annotations, direct access to supplemental material, and the means to not only mark up the content in the browser but sync to their ReadCube desktop app. This is ReadCube's second integration with a major publisher this year, following on their recent announcement on pairing up with Open Access publisher, Frontiers (third integration in total, together with Nature.com).

From their press release:

“Our goal is to deliver smarter, more connected and more useful journal content to researchers, in formats that integrate with the way researchers work” said Jay Flynn, Vice President and Director, Web Strategy and Business Development, Wiley. “ReadCube brings long-needed functionality to the PDF, retaining its benefits of portability, familiarity, and print-like layout while transforming it into an interactive gateway to related articles and cited literature.”

The ReadCube Web Reader is available now on 109 journals in the Wiley Online Library and will gradually become available for the whole collection during the first half of 2013. Wiley Online Library is a multidisciplinary collection of online resources covering life, health and physical sciences, social science, and the humanities, and contains approximately six million scholarly articles from over 1,500 journals and reference works.

“We’re thrilled to have Wiley as a partner,” said Robert McGrath, co-founder and CEO of Labtiva. “Together, we’re making it easier for scientists to critically analyze their research and accelerate the dissemination of scientific knowledge. One of our goals is to create a multi-publisher ecosystem with millions of interconnected articles to help publishers maintain engagement with their readers. Our partnership with Wiley is a critical part of this endeavor.”

For more on the announcement, view their press release, or follow ReadCube online for more at @ReadCube.

figshare announces new partnership with the Public Library of Science

Wed, 30 Jan 2013 09:31:00 +0000

figshare, a community-driven open data project, today announced a new partnership with the Public Library of Science (PLOS) to aid the visualisation of different types of data across the PLOS journals and make the data more widely accessible. figshare will host the supplemental data for all seven PLOS journals, as well as provide a widget that will enable PLOS users to view data in the articles in the browser alongside the content. figshare was founded by Mark Hahnel, and is part of the Digital Science portfolio.

“PLOS believes in making data as visible and useful as possible,” said Kristen Ratan, Chief Publishing and Product Officer at PLOS. “Partnering with figshare is an important step in increasing the accessibility of the data associated with our research articles.”

The growing diversity of research outputs (ie., data, figures, videos, code) has led to exploration by service providers and, more importantly, publishers into ways of incorporating different file types into research articles, making for a more complete record. An early example of this can be seen in Elsevier’s article of the future project. figshare is working with publishers to help them improve the submission process for data and other research products, at a time when there is increased pressure to include other components necessary to reproduce an experiment in the journal article.

This partnership illustrates PLOS’ continued innovation in the academic publishing space, with a focus on catering to the needs of their authors and evolving publishing paradigms. The new functionality on PLOS also allows the files to be discovered through a portal on figshare which will go live in the coming weeks, enhancing the discoverability of the data and providing each submission with its own Digital Object Identifier (DOI) so it can be cited. The partnership will also help PLOS fulfil the growing requests from funders for academics to make all of their research outputs available, as all data on figshare is openly licensed under CC-BY (if copyright applies) or made available under CC0.

“We are delighted to be partnering with PLOS on this project, a publisher with a similar ethos,” said Mark Hahnel, founder of figshare. “We hope this will better enable researchers to do more with the data behind the papers, improving the transparency of the research and aid reproducibility.”

It has been a busy week for figshare, who also recently announced their new figshare Advisors programme. The advisor scheme rewards figshare users who encourage their colleagues to use figshare for both their private cloud based management of research outputs and the ability to get credit for all of their research.

The press release announcing the partnership with PLOS can be found here.

As always, feedback, comments, suggestions and ideas are welcomed by figshare. You can get in touch with the team at info@figshare.com or via twitter, facebook or google+.

ReadCube web reader now available on Frontiers content

Wed, 16 Jan 2013 10:16:00 +0000

We're delighted to announce that Frontiers, an Open Access publisher of scientific literature, has partnered with Labtiva to integrate the ReadCube web reader for all of their peer-reviewed online content. This will allow users to access Frontiers content online in an editable, interactive form that can sync to their ReadCube desktop application.

This builds on Labtiva's flagship product ReadCube - a free, cross-platform desktop application that provides researchers with a better means of organising, accessing, and discovering literature in their field.

Frontiers readers will now have access to additional functionality in the browser through the web reader including annotation tools, hyperlinked in-line citations, and clickable author names that initiate searches to quickly find related articles. In an effort to best serve the Frontiers community, Labtiva have customized the web reader so that hyperlinked author names also link to the author’s profile page in the Frontiers Research Network. You can see an example of what this looks like here: http://www.readcube.com/articles/10.3389/fonc.2012.00190

From yesterday's announcement:

“Our goal is to catapult scholarly publishing into the digital era,” says Kamila Markram, CEO of Frontiers and a researcher at EPFL. “The ReadCube Web Reader will help us to achieve our mission by providing quick and simple features for our growing community of researchers to organize, collaboratively annotate and share scientific papers online”.

“In the digital age, scientists should not have to waste time chasing information through complicated and confusing systems,” says Robert McGrath, Labtiva’s co-founder and CEO. “ReadCube’s purpose is to simplify many common time-consuming steps and thus free up time and energy that can be better spent engaging with literature and generating new data.”

The ReadCube web reader is also available on a number of Nature journals, which you can read more about here. For the full Frontiers announcement, visit Labtiva's press page.

Labguru pairs up with Altmetric, adds real-time data

Tue, 15 Jan 2013 08:25:00 +0000

Labguru users will now be able to see related news articles, blog posts and publicly visible shares on social media sites for papers they upload, thanks to a new collaboration announced yesterday with Altmetric. Labguru, a lightweight research management app, is the flagship project of Biodata. Altmetric enables users to track and measure online activity around academic research. Both companies are supported by Digital Science.

From the release:

“Labguru is constantly striving to add features and tools which enable researchers to concentrate on science and simplify their access to relevant and focused information,” said Adam Sartiel, BioData’s CEO. “We believe that the integration of Altmetric’s attractive ‘donut’ tool into Labguru helps scientists evaluate the impact of scientific papers much better than any traditional scoring method.”

"It's becoming increasingly important for researchers to follow the online conversations around the papers they care about," said Euan Adie, founding partner of Altmetric. "By collaborating with BioData we've made that much easier for users of Labguru."

Are you a Labguru user? Let us know what you think of the new functionality. We'd love to hear your thoughts.

Happening this week: IDCC '13

Mon, 14 Jan 2013 09:07:00 +0000

Today marks the start of the Digital Curation Centre's 8th international conference, known as IDCC '13. The program is a diverse mix of some of the top thinkers when it comes to issues of digital curation, data sharing, standards and information management. I'm delighted to be joining such an all-star group, speaking Tuesday in the Innovation/Applications track on our work at Digital Science, and in general, making research more efficient. Mark Hahnel of figshare will also be joining for the event.

At the tail end of last year, the organisers asked if I'd be interested in engaging in an email interview leading up to the event. Below is an excerpted version of the interview. For the full post, visit their website. You can also find the program here. You can also follow the conference online on Twitter by searching for the hashtag #idcc13.

-----

Your presentation will focus on Infrastructure. Are there any specific messages would you like people to take away from your talk?

It's easy to think that we've worked out most of the kinks in research when we look at some of the latest advances in astronomy, genomics, and high-energy physics in the news, from the work at the LHC to the ENCODE project. But there are still a number of baseline assumptions in research that need rethinking - and in many cases, fixing. That's what Digital Science was created to address, some of the oft overlooked roadblocks in things like search in the sciences, information management, and the dated incentive system which is keeping us from fully updating our practices in the lab.

We address three areas in our call this year - Infrastructure, Intelligence and Innovation. What do you see as the most pressing challenges across these?

Having worked on infrastructure issues in research for the last six years, I'd say one of the main challenges remains making the right design decisions. Whether that's an open platform that operates on the back bone of the web or a lightweight software application for use in a research setting, design decisions are key, and in my experience, are often not thought through to the extent warranted.

There's a reason why inefficiency still exists in modern research labs, and it's not a shortage of tools. Part of that still lies in how the systems are crafted for the individual user, but also how it speaks to other systems.

Also, the age old incentive problem is still keeping us from reaching our full potential, as we continue to largely measure impact as papers produced. Not only does that skew researchers' incentives to better manage and make available say, for instance, the data accompanying their research or the code needed to execute the experiment, but it only presents issues for other specialists whose main output may be software, not scholarly papers.

We need to rethink how we measure and reward research so that it better reflects a researcher's contribution on his/her community and give the system a hard refresh.

And in terms of opportunities, do you see potential in data science as a new discipline?

Absolutely ... though it's not a "new" discipline, necessarily. There is an increasing understanding about the power in bringing together skillsets such as mathematics, machine learning, statistics, computer science and domain expertise (though not always necessary), which is helping us redefine how we think of hypothesis-driven research, becoming more data-driven.

What I find particularly fascinating is the spotlight it's putting on how we teach science undergraduates - making sure they not only have the practical skills for working in a lab or conducting an experiment, but also the statistical literacy and analytical reasoning to understand the information they're producing and collecting.

The conference theme recognises that the term ‘data’ can be applied to all manner of content. Do you also apply such a broad definition or are you less convinced that all data are equal?

I'm an equal opportunity data fan (and open purist, carried over from my time at Creative Commons). Too often, I feel, we get caught up in debates about the "worthiness" or "value" of particular data sets, a legacy from the publication world where only the most polished, interesting data counts. It's pervasive and keeping us from doing more robust, reproducible work. I am a strong proponent of not cutting oneself off from yet unknown opportunities, and unfortunately classifications such as "junk data" are not only increasingly silly in the digital age, but borderline harmful.

The lab safety problem - a personal story

Tue, 08 Jan 2013 17:05:00 +0000

The following is a guest post written by Nathan Watson, founder of BioRAFT, a Digital Science-supported company.

My left index finger should probably glow green under 488nm light for the rest of my life. As a junior researcher, I was producing recombinant viruses to overexpress genes including Green Florescent Protein (GFP) and a number of oncogenes. While isolating the virus carrying GFP, I was supporting the test tube with my left hand and slowly pressing a syringe needle through the tube’s soft plastic wall. I pressed too hard and the needle went completely through the tube and into my left index finger. I washed the finger with iodine and soap, put on a new pair of gloves, and finished the virus isolation.

I told no one.

I checked the cells and my finger daily over the next two weeks, and to my relief, the cells glowed and my finger did not. I corrected my methods when working with the oncogenes and consider myself lucky I had my incident with just the GFP.

This is the reality of research. If you had asked me if safety was important to me, I would have said, “sure it is.” If you had asked me if my lab was safe and if my PI cared about safety, I would have said yes. Yet, did I get training on the specific experiments I was doing or the hazards I was working with? Not really. Safety training, beyond the general lab safety lecture during new employee orientation, often entailed a post doc giving you a detailed protocol and encouraging you to ask questions. Yet, as an enthusiastic researcher who believed he already knew his way around a lab, I (and others like me) was more concerned with the scientific objectives, not the rigid tedium of safe experimentation.

That was how it was, and that is how it is today, but that isn’t how it has to be tomorrow. In the wake of survey results on lab safety culture published by BioRAFT (the company I founded to address this problem), Nature Publishing Group, and the UC Center for Laboratory Safety, we see that similar stories are playing out in research facilities across the world. Some researchers will be lucky as I was, but others may be working with oncogenes in their first accident, or not wearing a lab coat when a highly hazardous chemical spill occurs. The results of the survey point strongly to the need for research institutions and companies to take up the cause of safe working conditions in labs and to go beyond checklists and protocols to develop a new paradigm for safe experimentation.

The reality at your institution may surprise you. Go on, take a look for yourself …

For more information on the survey, visit BioRAFT’s website. You can also download the report here.

Season's Greetings From Digital Science

Thu, 20 Dec 2012 11:28:00 +0000

Tis the season to be... sciencey.

So we've put together a very serious video on seasonal science in action in the Digital Science office starring Digital Science staff.

Mention the video on Twitter and you'll be entered in to the running for a Digital Science swag pack!

We've got an exploding Christmas pudding and a cameo from Brian Cox.

It's very high brow stuff, so enjoy:

Season's Greetings and best wishes for 2013!

From all of us at Digital Science.

figshare links your desktop to the cloud

Wed, 12 Dec 2012 09:07:00 +0000

Yesterday saw the launch of the first figshare desktop uploader. The first release of the desktop uploader is a beta version and is only Windows supported, but versions for other operating systems will be coming soon.

It’s a very simple app that allows resumable uploads to your private ‘My Data’ section of figshare, where you can add some metadata and make your research outputs sharable, citable and discoverable, or keep it private. You can read more about the types of research outputs you can upload to figshare here.

The private section of figshare gives you 1GB of storage space at no cost, so that you can access your research outputs in a secure manner from any computer in the world.

The desktop uploader is a lightweight app that sits on your desktop and allows you to drag in your research outputs, where you can then track their progress as they are uploaded to the cloud. You can upload many files at once and the uploader supports resumable uploads. This means if your internet connection drops, you don’t need to start the uploads again.

This adds to the ever-growing collection of tools that you can use to help better manage your research outputs via figshare. You can read more about the desktop uploader features, as well as the figshare API and badges on the tools page. The developments have been coming thick and fast recently. Here is a couple more of the improvements from last week:

Reserve DOIs

A common use case for figshare is for users to upload research outputs that don’t fit into traditional publications, which they then link to in the reference list of the said journal article. However, some users suggested that they often did not want to make the research output on figshare live until the paper was published. For this reason, we have added a new feature which lets you see the complete citation for all of your private objects, including the DOI. This DOI is not active, just reserved.

This means that you can include this reference in articles in the reference list, without having to make the research output publicly available on figshare until the paper goes live. The ‘Reserve citation’ link can be found at the bottom of each expanded article in 'My Data'.

Presentations section

As time has gone on, we have seen researchers wanting to get credit for an ever growing diverse amount of research outputs. The most recent of these is presentations. So today we are adding a presentations type to the repertoire.

Complementing this nicely is the new custom built document viewer to be rolled out across all relevant articles very soon. This handles both papers viewing and presentation slides. There are lots of subtle improvements in the design that you will also notice, including batch edit and a more simple way to create filesets. Authors and tags can now be dragged and dropped into the order of your choice.

You can find out more about the different research output formats on the ‘types’ page. We are also looking into a code/software type. The problem we are trying to get our head around is licensing as there seems to be a lot of discrepancies in the licenses by which people share their code. Of course any code or software that you upload to figshare at the moment we are happy to make citable for you and sharable under a CC-BY license.

As always, feedback, comments, suggestions and ideas are welcomed. Please get in touch with the figshare team at info@figshare.com or via twitter, facebook or google+.

SureChem Deposition to PubChem

Mon, 10 Dec 2012 15:08:00 +0000

In a move that will boost open drug discovery, SureChem has deposited more than eight million chemical compounds into PubChem, the first time that any patent chemistry database has been made publicly available in its entirety. More than half of these are novel to PubChem - the world's primary public chemistry database - providing a rich new source of medicinally relevant compounds for researchers worldwide. Users can view patents related to SureChem data in PubChem through links provided to SureChemOpen, a free patent chemistry search resource.

This data deposition is key part of SureChem's mission to integrate patent chemistry into the online research community, enabling access to wider range of users. SureChem is also linked to the free ChemSpider database as well as chemical structure data in Royal Society of Chemistry journals. More links to other public and proprietary resources will follow in 2013.

For more information on the PubChem deposition, visit the SureChem blog or read today's press release.

The SureChem patent chemistry line of products is an internal offering from Digital Science. SureChemOpen offers basic free patent chemistry search and document viewing. SureChemDirect provides batch access to its complete chemistry and full text data collection, enabling integration of patent chemistry with internal workflows and databases. Professional and frequent searchers will soon have access to SureChemPro, and end user tool that provides powerful search, export and analysis tools. For more information, visit surechem.com.

Happy Birthday To Us!

Fri, 07 Dec 2012 16:59:00 +0000

Today marks Digital Science's second birthday, and as we round out another year, what better time to reflect on some of the progress made by members of the Digital Science family in 2012. It's been an extraordinary year (and an exceptionally busy one, to boot), and we wanted to pull together some fun facts and numbers as we look back on 2012. So here's a glimpse of not only our activity this year, but also a sign of what we hope to see more of as we look towards the new year.

Check out the amazing cake made by @Laura Thomson

Altmetric, a project that tracks and measures the the attention paid to research online, captured 2,282,826 mentions of 873,687 distinct scholarly articles in the past year.
The research article with the highest Altmetric score (1454) this year is ... drumroll please ... "Association of Coffee Drinking with Total and Cause-Specific Mortality" by Freedman et al. in the New England Journal of Medicine. You might want to have that cup of joe *before* reading that one ...
Labguru, a lab information management tool from our colleagues at BioData, grew in users from 3,500 users last year to over 40,000. They were also recently ranked #4 on the Top 10 Innovation for 2012 list from The Scientist.
1DegreeBio, an online marketplace for research materials and services, now has 1 million products and services listed, with over 4 million pages in product-related content accessible on the site. They also had over 145,000 unique visitors this year, almost triple the traffic they had last year. Need an antibody? Well, look no further.
Submissions for figshare, our open data project, have tripled since the service added DOIs (Digital Object Identifier ... an important step forward for research objects like data) and opened their API. Well done. This is a move we hope to see plenty more of in the new year.
Here's one to noodle on: Did you know that the average laboratory researcher is supposed to take five safety and compliance training courses each year. For the average lab, that's 60 training courses. For an institution with 500 laboratories, that's 30,000 training courses that need to be completed per year in order to be compliant. 80 percent of institutions have no method whatsoever in place to keep track of who is in each laboratory nor a way to track the hazards to which individuals are exposed. That's some serious risk. One of our most recent investments this year, BioRAFT, is looking at ways to lower that percentage and make the process more intuitive. (Also, stay tuned for more from BioRAFT in the new year regarding lab safety ... )
SureChem, our patent chemistry product line, has made available 4 million novel patent chemistry compounds in PubChem the world's largest free public chemistry database. More news on that to come before the year is out. Watch this space.
Symplectic, a research information management company for institutions, added this year integration with CrossRef and RePeC. With the addition of those two information sources, their flagship product Elements now automatically aggregates bibliographic data from over 30,000 journals across nearly 2000 disciplines. That's no small order.

Congratulations to all of our colleagues around the world on their work this year and for their continuing commitment to working tirelessly on problems affecting the research community. We're not nearly there yet, but it's reassuring to know that we are still working towards our vision of a more efficient and reproducible means of doing research that utilises the tools and technology we have before us. I, for one, cannot wait to see where 2013 takes us.

On a separate note, and for a bit of fun, a look back to our public launch two years ago, and to the video we had commissioned for it. It's well worth a watch (and no, we can't sell you the Lego kit to build this yourself :) ). Enjoy.

Digital Science at Online

Tue, 20 Nov 2012 15:29:00 +0000

Today marks the first day of Online Information, an annual event in London for information professionals. Digital Science is a gold sponsor of this year's event, and both Mark Hahnel of figshare and Euan Adie of Altmetric will be joining panels lending their experiences to sessions on working in the cloud and metrics. (More at the end about where you can see them.)

Our Managing Director Timo Hannay recently sat down for an interview to share his insights on how those in the research community are grappling with today's myriad of information challenges. The video helps set the scene for this week's event as well as explain our work (and that of our portfolio) in this space.

Interview with Timo Hannay - Online Information from Digital Science on Vimeo.

The two-day event will focus on the following areas:

• Adapting to a multi-platform, multi device world
• The power of social: creating value through collaboration
• Making sense of big data and working in the cloud
• Search, research and discovery
• New frontiers in information management

Mark Hahnel of figshare will be on the "Working In The Cloud" panel Tuesday at 12pm. Euan Adie of Altmetric will be on the "Metrics and Measurement" panel on Wednesday at 12pm. You can also come visit us at the Digital Science stand to find out more about what we do. We hope to see you there.

Video, links from Monday's App Slam at SpotOn London

Wed, 14 Nov 2012 15:52:00 +0000

At Digital Science, we work to develop software internally as well as support other scientific startups working to craft tools to enable more efficient research through better use of technology. We have a keen interest in scientific software, and wanted to open up the floor to the SpotOn London attendees this past Monday to hear about tools they were developing to help researchers or ones they find especially useful in circumnavigating a problem.

Following an open call, we selected nine projects and tools for the App Slam, and invited a representative from each on stage to tell us more for 5 minutes. To help frame the session, we set four categories for submissions based on the tool's target area for application: literature review, experimentation, analysis and measurement and finally publication and dissemination.

Video for the session is now up, which you can access here. Working on an app or use one in your day-to-day work that helps with your research that you don't see here? We'd love to hear about it. Leave a note in the comments or drop us a line at info@digital-science.com.

And if you'd like more information about the tools shown and/or presenters, check out the links below.

Literature Review: F1000 Prime - Rebecca Lawrence (@F1000); Colwiz - Dina Fejes (@dinafejes, @colwiz); Grokk - Chris Rohde </p>

Experimentation: Scrazzl - David Kavanagh (@dkavanagh, @scrazzl); Wikipathways - Chris Evelo (@Chris_Evelo)

Anaylsis and Measurement: Emotion Sense - Neal Lathia (@neal_lathia); BRISSKit - Jonathan Tedds (@jtedds)

Publication and Dissemination: figshare - Mark Hahnel (@figshare); Altmetric - Euan Adie (@stew, @altmetric)

Labtiva and Nature Publishing Group expand user access to articles

Thu, 08 Nov 2012 13:19:00 +0000

Announced today in line with the Charleston Conference, the public can now buy and rent a selection of nature.com content for a lower price point through a new pilot developed by Labtiva, one of Digital Science's portfolio companies. The pilot is an extension of the ReadCube Access program which we wrote about back in September, being trialled with the University of Utah. As of today, the general public - irregardless of institutional affiliation - will have access to the same feature through ReadCube, Labtiva's flagship product that helps users better organise, annotate and discover content.

Articles read by clicking the ReadCube Access icon on nature.com pages cost between $5 and $11, compared to the $32 pay-per-view fee. A 48-hour rental costs between $3 and $5. Content can also be purchased or rented through the ReadCube application, as well.

“We are committed to continued innovation and experimentation with new channels and pricing models. Our goal as a scientific publisher is to sustainably bring affordable access to high quality scientific literature to the widest possible audience,” said Steven Inchcoombe, Managing Director of NPG, in today's joint release. “ReadCube Access coexists alongside our existing HTML full-text, PDF and app access options. We will be interested to see the uptake of this convenient format and price: whether by scientists on the go or checking for relevance; or by those in industry, patients and science-interested citizens.”

For more information about the pilot, see today's release.

What apps help you do your research?

Wed, 24 Oct 2012 15:54:00 +0000

Do you use any apps to do your research better? What's your favorite? We're looking for a few willing folks to give a short 5-minute talk as part of our App Slam at SpotOn London on 12 November. It's like a show and tell for software - informal, quick presentations with a strict no slides policy (though, we're happy to pull up the website).

We're framing this around the research cycle, highlighting tools and software that can aid researchers in their work. There are four different categories: tools that help you in your literature review, in running your experiment, in analysis and measurement, and finally, apps that help you publish and disseminate your work. You don't need to be the founder to participate - we want to hear about your experience as a user, too.

Care to join us? Fill out the short form on the SpotOn website. We'll be accepting submissions through 6 November. We hope to see you there.

figshare and symplectic - ORCID launch partners

Tue, 16 Oct 2012 14:27:00 +0000

Cross posted from the figshare blog.

Today see’s the release of ORCID (Open Researcher and Contributor ID), where researchers can distinguish themselves by creating a unique personal identifier. Two Digital Science companies are launch partners on the project, Symplectic and figshare .

Accurate identification of researchers and their work is growing in importance as researchers look to get credit for all of their research and track their ever-growing volume and diversity of academic outputs.

Other researcher identifier systems are currently in use serving specific populations or types of research work. ORCID does not compete with these systems, but rather provides a switchboard for crosslinking. Elsevier’s Scopus system and Thomson Reuters’ ResearcherID® will both link to ORCID and allow users to synchronize their publication information between the two systems.

“ORCID addresses a problem shared by individuals and organizations across the research community: reliably connecting research with researchers,” said Laure Haak, Executive Director of the non-profit ORCID organization. “But ORCID is more than a Registry, it is a community effort to embed these identifiers in research workflows.”

figshare and Symplectic are in good company with the following launch partners all implementing the service as of today:

• American Physical Society
• Aries Systems
• AVEDAS
• Boston University
• California Institute of Technology
• CrossRef
• Elsevier
• Faculty of 1000
• figshare
• Hindawi Publishing Corporation
• KNODE, Inc.
• Nature Publishing Group
• SafetyLit
• Symplectic
• Thomson Reuters
• ImpactStory
• Wellcome Trust

figshare will be using ORCID in several ways. As of today, you can link your figshare account to your ORCID id. This enables us to populate your profile with your research publications and push your figshare publically available research outputs to ORCID, should you wish. This functionality will be rolling out in the coming weeks.

You can sign up for an ORCID account when you sign up to figshare here, or through your profile here, when logged in.

As always, feedback, comments, suggestions and ideas are welcomed. Please get in touch with the figshare team at info@figshare.com or via twitter, facebook or google+.

New ReadCube pilot introduces cost savers for research community

Thu, 27 Sep 2012 16:44:00 +0000

Labtiva, one of our portfolio companies, today launched a pilot program that will provide librarians and institutions a means of renting or buying articles outside of their institution's catalog at a discounted price. The pilot program, called "ReadCube Access", will enable researchers instant access to certain non-subscription content through their flagship product, ReadCube, a free application that helps researchers better manage their literature collections. The pilot is being run initially with Nature Publishing Group and the University of Utah.

“We want to help scientists access any journal article they need, without having to pay $30 per article or jump through other hoops,” says Siniša Hrvatin, co-founder of Labtiva, in their press release.

From their announcement today:

Researchers at the university no longer need to pay to read articles in any journal on Nature.com, including more than 50 titles for which the institution does not hold full site licenses. Instead, the library picks up the tab as researchers download articles in a single click through the free ReadCube literature management application.

The library will pay only $5 to $10 for each download. Researchers can also choose to rent an article for 48 hours, which will cost the library only $3 to $6. Articles downloaded through ReadCube Access appear as normal PDFs, but cost much less because they may not be printed or shared.

For more information on the pilot, see their announcement, or let them know what you think online at @ReadCube. You can also download ReadCube for free on their website. We'd love to know what you think.

Strata Conference is coming to London ...

Tue, 25 Sep 2012 17:39:00 +0000

Next week marks the start of the inaugural Strata Conference in London, an O'Reilly event on all things data. This is the first time the conference series, started back in 2011, is being held outside of the US and it's a thrill to be co-chairing it alongside O'Reilly's Edd Dumbill, one of the industry's leading thinkers on data and the technologies necessary for analysis, synthesis, and much more.

Don't have your ticket just yet? Don't fret. Plug in the discount code at the bottom of the post and save 25% off the main ticket price.

Now back to next week's event and what you can look forward to. The event is dedicated to exploring the opportunities made possible by the latest technologies to better understand the world we live in. Over the course of the two days, we'll hear from practitioners, policy makers, researchers and more, each sharing their perspectives and expertise on how they're dealing with data. We have a diverse mix of sectors and topic areas represented - different lenses, if you will - from government and business, to media and scientific research. The program includes a mix of nuts-and-bolts sessions for building a data-driven business, case studies and narratives highlighting , and nitty gritty technical sessions led by practitioners pushing the envelope from the US, UK and Europe.

But I won't ruin the surprise. Check out the program for more on our keynotes and sessions.

And remember, it's not too late to still join us. Plug in the code DIGSCI25 when you register for a 25% discount. We hope to see you there.

Digital Science Hackbreaks: The SureChem Mobile App

Wed, 19 Sep 2012 09:17:00 +0000

Scientific Literature Survey using Machine Vision

Earlier this year, a group of Digital Science engineers decamped to a holiday home in Norfolk for three days of intensive hacking. This blog posting is about one of the applications developed at the Hackbreak: SureChem Mobile, a smartphone application to help chemists learn more about chemical compounds in printed matter. To recognize the compounds in photographs taken on the smartphone, we used a combination of Keymodule's CLiDE machine vision tool plus several custom image pre-processing steps.

Here’s a video of the prototype SureChem mobile application in action. This first version links out to widely used, general purpose chemistry databases. In a future version we'll be adding support for the more specialized data available in SureChem patent documents:

SureChemMobile Prototype Demo from Digital Science on Vimeo.

The aim of the application is to provide a new workflow for scientific survey:

Use a mobile phone camera to take a photograph of a chemical structure
Upload the picture to the SureChem server for image processing and analysis
After successful recognition, retrieve resulting data from the server, including the Molfile and SMILES representation of the chemical structure, and chemical meta-data such as a generated name.
Query external chemistry resources to learn more, such as detailed chemical information from ChemSpider and Journal articles from the Royal Society of Chemistry that refer to the compound

The idea is to make it easier for a researcher to learn more about a compound after seeing it in a publication.

Implementation

The prototype mobile application was created over three days (and nights) in the holiday home shown below, by team J - that is, Jan Wedekind (expert in all things image processing), Jim Siddle (SureChem back end guru), and Jose Airosa (SureChem API master).

We used the PhoneGap toolkit to develop the mobile app, and a Sinatra server application was used to receive the queries and return the results via a RESTful API. Incoming images were preprocessed using the HornetsEye Ruby library and conversion of the chemical image to a searchable chemical structure was performed using Keymodule’s CLiDE. Chemical meta-data was generated using the ChemAxon JChemBase toolkit.

Below, we’ve included a series of images that together describe the preprocessing algorithm that identifies and isolates a chemical compound in an image. This algorithm is applied to images before they are passed to CLiDE for chemical recognition.

	The initial image. An enlarged version of the area marked by the red rectangle is shown in the top-right corner of the image.
	A dilated version of the input image. This is used as an estimate of the local background brightness.
	The difference of the (greyscale) input image and the dilated image serves as a basis for thresholding.
	Otsu's method is used to reduce the data to a binary image.
	The input image is dilated and the connected components are identified.
	A weighted histogram is used to find a large component close to the centre of the image.
	The selected component of the dilated binary image and the initial binary image are used to extract the graphical representation of the chemical structure.
	The result is ready to be processed with optical structure recognition software.

Here’s a larger version of the input photograph, side-by-side with the image we pass for chemical recognition.

That's all for now, but keep an eye out for future articles where we'll describe other experiments and prototypes from Digital Science hackbreaks.

figshare launches public API

Wed, 29 Aug 2012 16:46:00 +0000

After some very successful private beta testing, open data platform figshare is happy to announce the public beta release of our API. A big part of figshare ’s goal is to allow the simple distribution of research outputs and the ability to build on top of the research in an efficient manner.

This is what the API is for. You can push data to figshare , or pull data out. This first version is a basic implementation that allows you to manage your figshare account or build applications on top of the figshare platform and public research. We will continue to develop the API and add new cool stuff like resumable uploads and more advanced search filters.

To start creating applications on top of the figshare data, log in to your account and click ‘Applications’. You can check out the API documentation available at api.figshare.com/docs.

Now that the API is ticked, we can continue to develop new and existing features to make things even better. As usual, feedback is always welcomed as researchers help to shape what we become. You can get in touch with us via at info@figshare.com or via twitter, facebook or google+.

SureChem launches its first enterprise product, offering direct API access

Mon, 20 Aug 2012 13:00:00 +0000

SureChem, our patent chemistry product line, today launches a new enterprise product called 'SureChemDirect'. The tool will provide API access to the content that powers SureChem: patent full text and metadata currently comprising 12 million structures, 90 million patent abstracts and full text records from US, European and WIPO patent authorities. SureChemDirect enables customers to incorporate sophisticated chemical patent search into their workflows, perform batch services on their own platforms, and enhance their internal databases. The announcement was made in line with this week's American Chemical Society (ACS) meeting in Philadelphia, where the team is exhibiting and presenting.

SureChemDirect is the second product released in the SureChem product line, following the launch of SureChemOpen last March, their free portal for researchers. Both aid in making patent chemistry search (think: early stage drug development) more streamlined and approachable for researchers and businesses alike, putting high-powered tools and databases into the hands of those who use them.

The underlying technology works to make patent chemistry less painful and less reliant on brittle string matching and text string searches by allowing chemists to search how they think - in chemical structures - the interface allowing users to actually *draw* chemical structures and sift through the database that way. This functionality, paired with highly curated underlying content, makes for an extremely powerful tool - with that premium content now available for use via SureChemDirects's API access.

For more information, see the official press release, or visit www.surechem.com.

figshare teams up with Science Exchange and PLOS on Reproducibility Initiative

Tue, 14 Aug 2012 11:33:00 +0000

Open science platform figshare is happy to announce our involvement with the Reproducibility Initiative in collaboration with Science Exchange and PLOS. The Reproducibility Initiative is a new program to help scientists, institutions and funding agencies validate their critical research findings.

“In the last year, problems in reproducing academic research have drawn a lot of public attention, particularly in the context of translating research into medical advances. Recent studies indicate that up to 70% of research from academic labs cannot be reproduced, representing an enormous waste of money and effort," said Dr. Elizabeth Iorns, Science Exchange’s co-founder and CEO. “In my experience as a researcher, I found that the problem lay primarily in the lack of incentives and opportunities for validation - the Reproducibility Initiative directly tackles these missing pieces."

The Reproducibility Initiative provides both a mechanism for scientists to independently replicate their findings and a reward for doing so. Scientists who apply to have their studies replicated are matched with experimental service providers based on the expertise required. The data from these studies will be hosted on figshare . The initiative leverages Science Exchange’s existing marketplace for scientific services. Scientists will receive the results of their validation studies and have the opportunity to publish them in the journal PLOS ONE as part of a Special Collection highlighting the importance of reproducibility in scientific research.

The Reproducibility Initiative will provide a mechanism for industry to identify robust new drug targets for developing effective new therapies. "It is critically important to independently validate preclinical data before moving to the clinic," said Dr. Lee Ellis of MD Anderson, an Advisor to the Reproducibility Initiative who co-authored a widely read article in Nature earlier this year on the need to improve the reliability of preclinical cancer studies. “Patients are not just subjects in a study, and we owe it to them to validate preclinical studies prior to moving a drug or a regimen into clinical trials."

Also expressing his support for the Initiative was Dr John Ioannidis of Stanford University, "The Reproducibility Initiative is a very important pilot effort, offering valuable insights on how reproducibility checks can work in real life." Dr. Ioannidis, also an Advisor to the Reproducibility Initiative, published a study in PLOS Medicine titled “Why Most Published Research Findings Are False".

The Reproducibility Initiative is initially accepting 40-50 studies for validation. Studies will be selected on the basis of potential clinical impact and the scope of the experiments required and, in aggregate, may eventually serve as a proof-of-concept for this mechanism of validation to funding agencies and patient groups.

As always, feedback, comments, suggestions and ideas are welcomed. If you represent a publisher and would like to hear more about figshare's offerings for publishers, please get in touch at info@figshare.com or via twitter, facebook or google+.

Mining Patents in the Cloud (part 3): Design Issues

Thu, 02 Aug 2012 15:20:00 +0000

Introduction

In parts one and two of this series of blog postings, we’ve introduced SureChemOpen, and described how patent documents are processed to extract interesting chemical compounds for searching. We’ve also described how cloud technologies were employed to implement the pipeline.

The final part of this article digs into the design of the SureChemOpen pipeline, discussing what worked well, what didn’t work quite so well, and describing a few pitfalls you might encounter if you’re thinking of building a cloud-based data mining pipeline of your own. The discussion also touches on common operational and developmental requirements such as scalability, reliability, complexity, data integrity, and performance.

Scaling

Scaling up and out was one of our top priorities, and it’s one of the key benefits of a cloud-based pipeline. The ability to (relatively) easily add new compute nodes means that additional processing capacity can be added as needed to deal with higher workload, and the use of a queuing system to coordinate work means that new task instances can be added almost without limit. Practically speaking, other bottlenecks will emerge in a pipeline with very large numbers of tasks - for example in accessing shared resources such as a database.

It’s also possible to utilize AWS “spot instances” to expand your processing capacity further - a spot instance is a lower-price compute node made available at off-peak times via a bidding system. While we’re not quite there yet with SureChem, it’s also possible to automate the use of spot instances to add tasks and increase capacity.

The main tradeoff in scaling is complexity.

A single, serialized, data processing pipeline instance is much easier to manage and maintain than a multi-node, multi-process, multi-threaded, highly concurrent system. In the process of scaling out we’ve had to think hard about the consequences of any two units of work being done at the same time, key challenges being database integrity (more on this below) and multi-threading. We've found that Java processes aggressively acquire memory, meaning multiple JVM instances on a single system is undesirable. As such, we’ve had to make sure our tasks can run alongside each other in the same JVM process and to partition those tasks that don’t play well together into different processes.

Scaling out also can’t be rushed - if you’re processing a very large volume of data, make sure you leave plenty of time to build up to full capacity, don’t just expect to switch on another twenty nodes and be done. While this is possible, it’s a good idea to take time to build the trust in your system - to work out the bugs, build essential monitoring and configuration infrastructure, etc. Once you’re confident your system will work seamlessly at scale, you can add more capacity in the knowledge that the risk of expensive mistakes has been minimized.

It’s also important to consider the cost of scaling. While small-scale processing is cheap, costs can quickly grow as you add more capacity (i.e. more EC2 nodes) and as you process more data. You’re also at the mercy of Amazon’s pricing policy, changes in which may mean fluctuating or unexpected costs.

Optimization

Closely related to scaling is optimization - how to get the optimal performance from the resources you have. While it may be tempting to just add more capacity, make sure you’ve performed the easiest and most obvious optimizations first, at least.

The main workflow in data pipeline optimization is to find the bottleneck, remove it, then repeat. CloudWatch is your friend here, because it will show you where the bottleneck is in your system for any particular build or configuration. Look for one particular task which is preventing work from reaching other downstream tasks, then investigate why it’s taking so long. Check if your task is particularly CPU intensive, requires more memory (e.g. for caching), or is blocked on I/O.

For CPU intensive tasks, you may be lucky and find that the cause is unnecessarily complex code - in which case you can rework it to improve performance. More likely though, the complexity may be necessary or very expensive to speed up in which case you can look to scale out. Consider adding more instances of the bottleneck task; run these as new process instances, or as additional threads alongside existing tasks if you’re running JVM instances. In our initial pipeline configuration, we found that even though we had CPU-bound tasks, CPU utilization on our system was very low while memory use was high. By rearranging where different tasks were running, and by adding more instances of tasks running as threads alongside other tasks, we were able to increase throughput and system utilization massively. If CPU and memory resources are fully utilized, add new nodes.

You may find that your tasks are blocked on I/O - i.e. taking a long time to read or write to the filesystem, or similarly blocked on network accesses such as to a remote database or service. Again you may simply need to review your code to see if there are any obvious optimizations - unnecessary or poorly implemented I/O operations. If not, review the granularity of work units your tasks are performing; very fine grained tasks mean more I/O is required for coordination and for sharing of intermediate resources.

If it looks like the bottleneck is the network, make sure you batch as many remote requests as possible - particularly SQS messages and database requests. SQS provides batch reading and writing of messages (limited to 10 per request, unfortunately) which speeds up message handling and reduces costs. Similarly if you’re interacting with an SQL database, one large query or write operation will typically perform better than lots of small operations. In the SureChemOpen pipeline we increased annotation throughput by an order of magnitude, simply by performing a compound INSERT statement - interestingly this bottleneck only appeared once we were running in the cloud, because the pipeline was no longer interacting with the database locally.

Message filtering, Idempotence, and Transactions

Once you move to a distributed pipeline, you’ll need to consider what happens if your pipeline receives duplicate messages, or different messages with equivalent or interacting side effects.

Duplicate messages can occur for several reasons. They may appear because of bugs in your code, or due to bugs in your chosen messaging system - prevention of duplicate messages is a knotty problem and cannot be guaranteed, just made very very unlikely. You may also find duplicate messages passing through your system because a job has been re-submitted.

Whatever the reason, how duplicates are handled is an important consideration. One option is to filter them - for example by performing a check against a database and throwing away messages that have clearly been processed already. This helps ensure data integrity and prevent unnecessary rework; that said you may wish to make filtering conditional based on a flag included in the message. In SureChemOpen we reject messages that would cause documents to be re-processed, unless they are marked as requiring reannotation within the message.

Even if you choose to filter duplicates, it’s a good idea to try to make the processing of each task idempotent. Those familiar with RESTful principles may know that idempotence is a characteristic of the PUT verb in the HTTP protocol. Simply put, it means that state of the system is consistent no matter how many times a particular message is processed. Practically speaking, idempotence may mean resetting stored data prior to message processing, or using insert/update operations rather than purely inserting new records. SQL, for example, provides the INSERT IGNORE operator which can be used to ensure database consistency. In SureChemOpen, we clear out all annotations for a document prior to reannotation - failing to do so would simply mean duplicate records in the database. Similarly when storing new chemical names, we use an INSERT IGNORE operation to ensure we only record each name once.

Another integrity consideration is transactionality. Ideally, the processing of each message will be atomic - meaning that it’s either fully processed (and thus deleted from the input queue) or nothing is processed and there are no side effects. Failing to ensure atomic operations can mean inconsistent or incomplete data, or in some cases downstream messages that cannot be processed because required resources are missing. You may be able to partition your pipeline so that tasks are naturally atomic - e.g. in SureChemOpen, the result of processing an OCR correction message is a single, downstream message. The only side effect of processing failure is that the message hangs around on the queue and the logs are a bit noisy.

If your tasks aren’t naturally atomic, transactionality can be expensive - both in terms of development and processing time. If you’re using a Relational database such as MySQL then you’ll get a certain amount of transactionality for free (except for processing overhead), but make sure you consider message transmission too. One option is to retain a buffer of messages that will only be transmitted once the input message has been successfully processed.

Keeping your development team sane(ish)

While the above topics have focussed on data processing performance and integrity, it’s also important to consider developmental requirements. While your pipeline may be fast and scalable, it should also be easy to develop and maintain. Here are a few suggestions for how to make life easier:

Logs everywhere. We try to log every significant event, often as informational messages but sometimes as debug messages for low level events. The combination of thorough logging with a log collection service such as greylog2 makes tracking and problem determination easier.

Use a task management framework. The AWS SDK for Java makes it easy to interact with AWS services, but it doesn’t provide a framework for task management or message processing. We developed a simple framework for pipeline tasks which handles all SQS related activity such as reading, writing, and deleting from queues. Having a single, unified abstraction for task processing makes it easy to apply wide-reaching, fundamental changes to your core aspects of your pipeline. Some languages (such as Erlang) provide this for free, though choosing a less mainstream language often means integration with third party software will be harder.

Start with a zero-tolerance approach to exceptions - only add more lenient exception handling when your system is running reliably. Most exceptions in SureChemOpen are re-thrown as a runtime exceptions, causing message processing failure. This increases the visibility of issues during development.

Choose third party dependencies carefully. SureChemOpen relies heavily on third party software for name to structure conversion, structure processing, and image analysis. Because our pipeline is fully automated and we process massive numbers of documents, tasks must be highly resilient. We’ve managed to expose numerous bugs in third party tools by hitting edge cases. So either avoid third party software if you can, or limit the number of dependencies.

In Conclusion

This article introduced SureChemOpen, and described the text mining pipeline and cloud-based implementation which provide the interesting chemical data at the heart of SureChem. It also discussed several design considerations to bear in mind when developing your own pipeline.

Overall we’ve found the combination of EC2, SQS, S3, and CloudWatch to be a good foundation for data mining in the cloud. By combining these technologies, we were able to build an easily scalable pipeline for generating SureChem data. The main tradeoffs for improved scalability are greater management and development complexity, along with less predictable costs.

Exploring handsfree technology in the lab

Wed, 25 Jul 2012 09:45:00 +0000

At Digital Science, we work to make research more efficient through better use of technology. It's a message we firmly believe in at our core and a reality many of us have experienced personally as scientists ourselves. The world of research has long been riddled with inefficiencies, be they technical or even social, each serving as a roadblock keeping us from doing better, more reproducible science.

Many a time, it's not because the technology doesn't exist - rather that it's not been implemented or tailored to a researcher's needs. To further explore that, we have produced a concept video looking at the possible benefit of repurposing a well-known piece of kit (in this case, a Microsoft Kinect) for use in a real laboratory setting. This is the first in an occasional series of concept videos, technologies and products that we're not necessarily developing here at Digital Science, but rather using as a conversation starting to show the potential of everyday tools in modern day research.

The full video can be viewed here.

Many thanks to my colleagues Laura Thomson, Steve Scott and Timo Hannay for their work on this video. Do stay tuned for future instalments, and let us know what you think either here in the comments or by e-mail. We'd love to hear your thoughts.

BioData secure follow-on funding after exceptional year

Tue, 17 Jul 2012 23:47:00 +0000

We're delighted to announce follow-on funding for our portfolio company BioData, producers of the lab research management software Labguru.

The company has had a tremendous year to date, launching a free iPad app designed to play off the strengths of their flagship web app, as well as securing Technion's Rappaport Institute as a customer and integrating some of the best peer-reviewed protocols into their system.

From their press release, issued today:

"BioData and Digital Science exist at a time when scientists are changing long-established habits and increasingly adopting digital productivity tools out of necessity to remain competitive. This is an exciting time for all of the BioData team. We enjoy support from Digital Science along with an ongoing dialog with top researchers worldwide - ensuring we develop the lab management tools a research group needs in order to do great science," said Jonathan Gross, Founder and Co-CEO of BioData. "We have invested a lot this year in rebranding and revamping the product, building a winning team, and developing tools to better serve our growing user base."

The news from BioData follows a similar announcement from another one of our portfolio companies 1DegreeBio, a testament to the work being done by both of these startups and our belief in the promise of their products.

"BioData has had a tremendous year. Labguru is proving to be a powerful and distinctive product, and customer feedback has been excellent. It really is changing the way that researchers work," said Timo Hannay, Managing Director, Digital Science. "We are delighted to have BioData as part of our portfolio, and look forward to even greater success in the year to come."

Congratulations again to our colleagues at BioData. Stay tuned for more in the coming months.

figshare and F1000 research shake up traditional academic publishing format

Mon, 16 Jul 2012 12:31:00 +0000

Last week we announced that research data platform figshare would be partnering with F1000 research to bring integrated research articles with the appropriate associated data files. The first articles have gone live on the F1000 Research site today and illustrate how the future of scholarly publishing is changing and how figshare is offering the tools to help publishers move forward to accommodate these changes.

One of the articles published today contains over 10GB of raw data files within the article, freely available to anybody to peruse in the browser and download should they wish. One of the widgets contains 1474 files. This is what science looks like now, it is not 6 static images in a pdf. As research outputs grow in size and number, academic publishers need to adapt in order to transform the research outputs so that they can be disseminated in this traditional manner.

Of note, the author of this paper, Manuel Corpas, has a figshare account to which these objects can be linked. This means that all the metrics associated with this research data is aggregated on his account and he gets the associated credit for all of his research outputs.

These embeddable widgets allow publishers to disseminate research outputs in a much more granular form, with all objects citable with a DOI. All objects are available on figshare with enhanced linking back to the original full article on F1000 Research. This aids discoverability of both the individual research objects as well as driving traffic to the full story on F1000 Research.

figshare will be using this model to help other publishers move forward in accommodating the needs of researchers and funders. The embeddable widgets will be available to individual figshare users soon.

If you represent a publisher and would like to hear more about our offerings for publishers, please get in touch at info@figshare.com or via twitter, facebook or google+.

Altmetric.com - making article level metrics easy

Thu, 21 Jun 2012 00:06:00 +0000

The following post comes from Euan Adie, founder of Altmetric, the newest addition to the Digital Science family, announced today. Euan also doubles as a product manager at Digital Science working on our data tools. You can follow him at @stew or the project at @altmetric.

-----

We've got a simple sounding mission at Altmetric - to track and analyse the online activity around scholarly literature.

If I'm honest the main reason I started working on the system is because I think aggregating scholarly metadata is cool (what? No, I don't get out much). You should care about it too though:

- If you're a scientist you should have a quick and easy way to see which papers your peers are talking about and where.

- If you're an author you should be able to identify and respond to comments about your work.

- If you're a patient you should be able to put any relevant research you find in context.

- If you're an editor you should be able to track the trends and hot topics in your field of interest.

- If you're a publisher you should be able to curate and host the conversation around the content you publish and make it available to your readers.

Altmetric is designed to help with all of above.

It began with some prototypes (one of which won Elsevier's Apps for Science competition, which paid for hosting and licensing costs) and with help from some awesome advisors took shape over about a year of evenings, weekends and holidays. Digital Science have always been supportive of Altmetric and I'm really pleased that they're now supporting it a bit more formally.

Here's how it works. Altmetric tracks three types of data sources:

- Social media like Twitter, Facebook, Google+, Pinterest and blogs

- Traditional media - both mainstream (The Guardian, New York Times) and science specific (New Scientist, Scientific American). Many non-english language titles are covered.

- Online reference managers like Mendeley and CiteULike

We pull out mentions of scholarly articles or datasets from these data streams, aggregate them, clean them up and then normalise all the resulting data - we track twenty thousand unique papers a week, but some people link to the PubMed abstract, some to the publisher site, some to dx.doi.org, etc. - before making it available to researchers and publishers though an API, embeddable badges and an analytics tool called the Explorer.

You check out our badges on sites like Biomedcentral, Frontiers or Scopus and in apps like Utopia Docs.

The free bookmarklet fetches and displays data from the Altmetric database for whatever article you're currently viewing. It usually needs a DOI to work, so if it doesn't find out automatically try selecting one on the page before hitting the button.

The Explorer is a paid product and aimed primary at editors, press officers and bibliometrics researchers. It allows you answer questions like "what neurology papers have researchers been sharing this week?"

(on a tangent, that first article - Altered processing of sweet taste in the brain of diet soda drinkers - is pretty interesting).

"which news outlets have been linking to articles in Nature journal this week?"

"which journals published by the Royal Society of Chemistry got the most attention online in the past six months?"

If you're interested in exploring any of these further I'd love to hear from you - just drop me a line at euan@altmetric.com and I'll be happy to get you set up with a trial account and any of the open data.

Labtiva introduces new "enhanced PDF" functionality for ReadCube

Wed, 20 Jun 2012 17:37:00 +0000

ReadCube users can now interact with their PDF collections in new ways, thanks to the newly introduced "enhanced PDF" functionality in the product.

New features include clickable in-line references and integrated author information, allowing users to click on a name to search for other publications or contact details for corresponding authors. Users can now view bibliographic references next to the article in an interactive side panel, providing quick access to related literature. There is also a new toolbar at the top of the window, providing better navigation to additional resources important to the researcher - supplementary information, notes, editor's summaries - all in one easy-to-find location.

You can learn more about the features in the video below, or on their website.

Mining Patents in the Cloud (part 2): Amazon Web Services

Fri, 15 Jun 2012 15:24:00 +0000

Introduction

In part one of this series of articles on SureChem's text mining process, we described the different processing steps required to extract searchable chemistry from patent documents. In part two below, we describe how the pipeline is implemented using different cloud services provided by Amazon.

To implement the text mining process described previously, we use Java-based processes integrated using Amazon Web Services (AWS) technologies. AWS provides compute power via Elastic Compute Cloud (EC2), message queueing via the Simple Queueing Service (SQS), data storage via the Simple Storage Service (S3), and metrics via CloudWatch. We also use greylog2 for low level logging. SureChemOpen code is Java-based primarily because many of our third-party dependencies are only available as Java libraries. One consequence of this is that interaction with AWS services is easily achieved using the AWS Java SDK.

The pipeline itself consists of many independent tasks running on several EC2 nodes that communicate via SQS and S3, with monitoring provided by CloudWatch and greylog2. Below, we’ll describe each of the major cloud technologies, and the role they play in the pipeline.

Elastic Compute Cloud (EC2) Nodes

EC2 nodes provide the computing power for the pipeline. Each node hosts one or more Java Virtual Machine instances, each of which runs one or more tasks such as text annotation, name to structure conversion, etc. The number of JVM instances and the tasks they host is controlled via configuration files, which can be modified depending on workload. We use a combination of foreman and an in-house task management framework written in Java to manage which tasks are running in which processes.

Currently we have just one standard configuration, which has been fine-tuned for performance on a single type of processing node to make optimal use of a particular quantity of memory and CPU resource. We’re working on an automated configuration mechanism which will add or remove task instances or EC2 nodes depending on workload, the chef automation framework helps with this.

Simple Queuing Service

SQS provides the queuing mechanism for units of work. Each task in our pipeline is associated with a particular queue, from which it reads messages that describe units of work to perform. For example a message may indicate that a particular document should be annotated, a name be converted into structures, or a chemical structure standardized and stored in a central database.

The result of performing a particular unit of work is some combination of database writes, S3 writes, and new messages on “downstream” queues. Once a task has read a message, SQS locks it for a certain amount of time to allow the task to perform the unit of work - this ensures that other tasks reading from the same queue won’t read the same message, and perform the same work.

Once the unit of work is complete, the task deletes the message from the input queue. If, however, there is some problem in processing the message - perhaps a timeout caused by hanging code, or an exception or segmentation fault causing the process to restart - then the message won’t be deleted, and after a preset time period will be made available again. This can have some unfortunate consequences such as churning or blocked processes, so it’s important to add safeguards such as task restart delays, or error queues for messages that repeatedly fail to be processed successfully - we employ both.

The combination of EC2 nodes, SQS (or another queueing system), with self-contained tasks that operate on individual messages provides the primary means of scaling the pipeline. To increase throughput of a particular part of the pipeline, we just add more instances of a particular task, potentially on new EC2 nodes depending on resource utilization. In our current pipeline, we have four annotators running, four name to structure converters, and ten structure handlers - among other tasks.

Our message formatting needs are fairly simple, requiring the transmission of easily encoded strings and numbers. So for message encoding and decoding, we use Google Gson. This provides a simple, easy to use encoding / decoding mechanism allowing messages to be defined using POJOs (Plain Ordinary Java objects).

Simple Storage Service

While EC2 and SQS can be combined to produce a working pipeline, one important missing ingredient is temporary storage space. SQS messages are (as of the current date) limited in size to 64k, which we’ve found is not enough for every type of intermediate processing artifact we generate. Chemical structure data, for example can exceed 64k in size, so messages that need to transmit such data (e.g. from Name-To-Structure tasks to Structure Handler tasks) cannot rely on messages to hold the required data.

As such, we use S3 for temporary storage. Essentially S3 plays the role of distributed file system for our pipeline. Any files required for processing are written into S3 buckets, then can easily be read by downstream tasks for further processing. In addition to providing the means for transmitting more than 64k between tasks, S3 also records these intermediate processing artifacts making it easier to diagnose processing errors, and easier to re-process downstream work.

CloudWatch

CloudWatch is a metrics gathering service, which is automatically updated by many different AWS services. Any use of SQS, for example, is automatically recorded by CloudWatch and numerous free metrics are made available for monitoring purposes. A simple command line service is available, in addition to graphs via the AWS web console.

There are seven pre-defined CloudWatch metrics captured for SQS, all of which are useful for day-to-day monitoring. One particular metric has proved invaluable - number of messages deleted. We’ve found that a simple summation of this metric, on an hourly basis over several days, provides an excellent overview of throughput for each of our tasks. It’s also a great way of seeing how work has progressed down the pipeline, and where bottlenecks are - simply overlay the message deletion counts for your different queues to see what happened when for a particular batch of input messages.

Using this approach, we’ve been able to determine when text annotation for a batch of documents finishes, when new names from those documents are converted to structures, and when the resulting structures are standardized and stored. All of which is useful information for tweaking pipeline configuration and estimating processing times for large datasets.

greylog2

So far we’ve seen how AWS technologies provide infrastructure and monitoring for a distributed, scalable pipeline. The final piece in the puzzle is low-level monitoring, which is essential for any non-trivial pipeline. We use the open source greylog2 to capture and collate low level logs generated by JVM processes.

Like many Java-based software systems we instrument our code using Log4J, which provides a uniform interface for capturing logging messages and transmitting them to various destinations. A number of Log4J “appenders” are available for greylog2, which transmit log messages to predefined greylog2 systems via UDP. While delivery of UDP datagrams is not guaranteed (by design) it is typically sufficient for monitoring purposes; that said we also record log entries in local files providing a system-local copy which can be easily processed using command line tools and also acts as a backup.

Because the SureChemOpen pipeline is complex and has many moving parts, there are numerous different aspects of the system that require monitoring on a daily basis. Separating these aspects out of the raw stream of messages can be difficult, so we use multiple Log4J appender definitions in combination with greylog2 streams to allow different processing aspects to be filtered via the greylog2 user interface. For example we send all error messages to one stream, all document processing messages to another stream, all task startup messages to another … etc. This allows for fast detection of errors, and makes regular status checking easy.

Intermission

That's all for now, but tune in to the final part of this series which will discuss the design of the pipeline. Specifically what worked well, what didn’t work quite so well, and a few pitfalls we encountered. In the meantime, here's an inquisitive husky enjoying a singing wine glass.

Investing in a safer lab workplace

Thu, 14 Jun 2012 10:08:00 +0000

Today, we are delighted to announce our investment in BioRAFT, a US-based software company specialising in tools to better manage health and safety protocols in the lab and compliance processes. Today also marks the launch of a global lab safety survey, done in part by BioRAFT, the University of California's Laboratory Safety Center and Nature Publishing Group, our sister company. You can read more about that here.

University health and safety practices have been under increased scrutiny in recent years following a number of accidents and fatalities that have made the mainstream news, such as the young chemistry researcher from UCLA who died in a lab fire in 2008, charges filed this past December. The reality of the systemic problem plaguing research institutions is hard to ignore, arising from antiquated information management systems, inadequate training regimens and inefficient compliance mechanisms that place too much admin on the researcher - all failing to provide for a safe and efficient workplace for students and faculty alike.

BioRAFT was founded in 2003 by Nathan Watson, a former researcher and lab manager, to address this issue, creating the software to solve a pressing need he saw in his day-to-day work.

“During my time in the lab, I learned that if lab safety and regulatory compliance processes are not as easy as possible to follow and manage, then researchers miss requirements, institutions receive fines, and people get hurt,” said Nathan Watson, CEO of BioRAFT.

The addition of BioRAFT to our portfolio signifies a move forward in providing better tools to make research more efficient and safer, a leading concern for the members of our community, and one that's oft neglected.

"Science is a profession, and we created Digital Science to provide the professional-grade software that researchers deserve," said Timo Hannay, Managing Director of Digital Science. "Nothing in a lab is more important than the safety and training of its members, and BioRAFT allows its users to achieve that with the minimum of distraction from their real aim – to make new discoveries. We are delighted to welcome them to Digital Science's growing network of innovative young companies."

For more information on BioRAFT, visit their website. You can also read more about today's announcement on our press release page. You can also find the lab safety culture survey here.

Funding news, new epigenetic portal from 1DegreeBio

Tue, 12 Jun 2012 12:44:00 +0000

It's been a big month in the news department for 1DegreeBio, a portfolio company of ours that aids researchers in sourcing and reviewing antibodies. The Toronto-based start-up announced yesterday that it had closed its Series B round of funding, led by a follow-on investment by Digital Science. Philip Armstrong, CEO of Jovian Capital Corporation and 1DegreeBio senior management also participated in the round.

The company has experienced significant growth over the past year, growing its user base from 23,000 to over 75,000 users, and building its product catalog from 400,000 listings to more than 800,000.

From their funding press release:

“1DegreeBio has proven a tremendously valuable resource to the research community,” said Timo Hannay, Managing Director, Digital Science. “We are pleased to have them as part of the Digital Science family, and look forward to seeing even more developments in their offering in the year to come.”

The announcement follows the launch of their Epigenetics Portal, launched on June 4 as a place where epigenetics researchers can find targets, tools and product data specific to their research. As of last week, validation data and community reviews for an initial set of 10 renewable recombinant antibodies for epigenetic targets can be found on 1DegreeBio.org. These antibodies are important due to their involvement in a wide variety of chronic diseases, and are the targets of a new field of drug discovery.

A massive congratulations to 1DegreeBio for their recent successes, and do stay tuned for more.

For more information on their recent funding round or the Epigenetics Portal visit their press page at 1DegreeBio.org.

Technion Medical School adopts Labguru for use in all research labs

Thu, 07 Jun 2012 02:21:00 +0000

We're delighted to share news from our portfolio company BioData, who today announced their newest customer - Technion-Israel Institute of Technology's Medical School. The deal will involve making Labguru, a lightweight lab management tool and BioData's flagship product, available to all of the research labs in the Rappaport Faculty of Medicine.

Technion is the oldest university in Israel and is ranked one of the top science and technology universities in Israel and Europe. Their move to adopt Labguru shows the university's interest in increasing the productivity of its researchers, enabling better communication between its research faculty and enable more efficient research.

From the official press release, “We are continuously seeking ways to support our faculty leverage scientific talent and resources to the maximum degree. Based upon the past few years of staff utilizing BioData’s services, the faculty found Labguru delivers a much needed competitive edge to our labs’ management and student mentorship,” says Professor Eliezer Shalev, dean of the Technion Medical School.

Labguru’s web service is a collaborative lab and research system that helps academic labs plan experiments, track progress on projects and milestones, share results, organize their protocols and data, and manage inventories. The tool is currently in use by over 35,000 researchers to simplify their lab experience and better manage their day-to-day research. For more on Labguru, visit their product page or their website. You can also read their press release in full here.

Help us in mapping the future of digital infrastructure in the UK

Wed, 30 May 2012 08:45:00 +0000

In late July 2011, David Willetts, the UK's Minister for Universities and Science, invited academics, technologists, hardware and software suppliers and experts from the Research Councils to discuss the establishment of an e-infrastructure roadmap for the UK. Following on that, a report was written exploring these issues, a significant investment was made for a 10-year roadmap, and a leadership council was assembled (of which I'm a member) to advise the government on how best to spend this investment for the maximum public good.

The funds allocated to this project come from the hard infrastructure budget for the UK (think roads, tolls, etc.), dedicated in this programme to “e-infrastructure” - digital infrastructure to help business, academia, industry, and the research enterprise. This includes funding and maintenance of a plethora of things: networks, computation, software, skills training, technologies. Together with other representatives from industry, academe, funding bodies and officials, we will be advising the government on how best to use these funds over the next 10 years to help encourage data science, technological development, and emerging business.

Digital Science sits as an industry representative on this council, as a voice for the communities we work with and care about: the start-up community, researchers looking to embrace digital technologies or spin-out businesses of their own, and small business. Following the first meeting of the leadership council (ELC) this past March, it became apparent that those communities were not as well represented in the discussion as they could be - their needs not mapped to what is, in essence, public money to foster and further emerging business, innovation, and research in the UK.

I've been tasked with leading a working group to explore the needs of these groups to help inform the broader council on how e-infrastructure affects them, and to gauge their awareness on existing programs and infrastructures that although publicly funded and designed for their use, may be under-utilised.

To do this, we have released as of today a short survey called "The future of UK's digital infrastructure". It will be live for the next two weeks, closing on 13 June, from which we'll distill the results into a report to present to the larger council in July. All of the results will be made publicly available under a Creative Commons license.

Do join me in spreading the word about this survey, accessible at the following address: https://www.surveymonkey.com/s/einfrastructure This is your chance to directly influence not only how this funding is divvied up, but also affect how the UK views data-intensive science, business, and digital technologies for years to come.

(Update: The deadline for the survey has been extended to 22 June 2012.)

Mining patents in the Cloud (part 1): the SureChem data processing pipeline

Mon, 28 May 2012 12:00:00 +0000

Introduction

This is the first part of a three-part blog series on how the SureChem team have rebuilt their data processing pipeline using AWS tools.

Digital Science recently launched SureChemOpen, a free service to help research chemists to find interesting chemistry in patents. This article is about the text mining infrastructure that makes SureChemOpen possible.

To start with, we’ll talk about what SureChemOpen is, what it’s used for, how to use it, and the data required to enable patent chemistry searching. We’ll then describe the text mining process: i.,e. how to go from a textual patent document, through annotation and chemistry detection, to building a database of patent chemistry. The SureChemOpen pipeline is built using Amazon Web Services technologies, and a future article will describe how we implemented the text mining pipeline in the cloud, using EC2, SQS, S3, and other technologies. Also to come is a design discussion about the scalability, reliability, complexity, data integrity, and the performance of our cloud-based data mining implementation.

What is SureChemOpen?

SureChemOpen is a search engine for chemists interested in patent chemistry, such as researchers at institutions working on drug discovery. Typical uses for SureChemOpen are to check if particular compounds have been protected (and thus may or may not be patentable), or to identify new or unexplored types of compound which may be candidates for research projects. SureChemOpen is a free version of the SureChem product portfolio, allowing low-intensity searches for free. Premium offerings (SureChemDirect and SureChemPro) are planned for release later this year.

At the core of SureChemOpen is a text mining pipeline, which we’ll describe in detail in the next section. But fundamentally, we start with a corpus of patent documents, run them all through a cloud-based data processing pipeline, and in the process build up a collection of chemical name annotations, chemicals found in images, and where possible, chemical structure data for names.

Every chemical we find is added to a searchable database, which allows chemists to find “interesting” chemistry. A typical chemistry search might involve entering one compound, and searching for any compounds with a 95%+ similarity level. Or a chemist may enter one part of a chemical compound, and search for all other compounds that contain that fragment as a “substructure”. After finding one or more interesting compounds, the chemist will naturally want to view the documents that contain them. Clicking one more more structures from a SureChemOpen structure results list shows a list of matching documents; these can then be opened, and the matching names or image annotations shown.

The Text Mining Process

So what is really involved in generating data for SureChemOpen? The data mining process can be broken down into the following discrete tasks.

1. Text annotation

The first step in the pipeline is text annotation. Here, we take the raw text of the document (typically HTML, provided by patent offices), and run it through a machine-learning based named-entity recognition tool, referred to as the SureChem Entity Extractor (EE). The tool is used to find “systematic” chemical names, for example:

1-[4-ethoxy-3-(6,7-dihydro-1-methyl-7-oxo-3-propyl-1H-pyrazolo[4,3-d]pyrimidin-5-yl)phenylsulfonyl]-4-methylpiperazine

The SureChem EE identifies chemical names in text by first tokenizing around white space and other significant separators, then calculating a probability for whether each token is chemical.

The probability is calculated based on which “n-grams” the token contains, where the presence of certain n-grams is a strong indicator that the full token is actually a systematic chemical. An n-gram is a sequence of characters of length n; so for example the word “example” has the following 4-grams: exam, xamp, ampl, and mple. We identify the 4-grams of each potential chemical name, then combine the “chemical” probability for each of these 4-grams to get the overall likelihood that the name is chemical. A finely tuned threshold ensures a high “F-measure” score, meaning that we find the vast majority of chemical names, and very few false-positives.

A machine-learning model isn’t all there is, however. We use dictionaries to find well known drug names, as well as heuristics and certain post-processing steps to improve the quality of our annotations.

There are two outputs of the annotation task: annotations and names. Annotations are simply start and end positions for the chemicals in the document; these are stored in a database and can be extracted later for rendering the document with chemistry. The names are sent on to the next “downstream” task for further processing.

2. Convert names to structures

Next, we try to generate chemical structures for every name detected in the previous step. There are several third-party tools (both commercial and open source) that take one or more names, then provide the chemical structure data that the name corresponds to. The most common structure format output by the tools is the .mol file.

We try our best to convert the name by passing it to five different tools. This can result in more than one chemical compound being generated for a given name; we capture everything and ensure that searching and exporting handle these cases appropriately.

Names that can be converted are sent, with all generated chemistry, for standardization and storage (see step 4, below). Non-converting names however are sent for OCR correction...

3. Optical Character Recognition Correction

Unfortunately, not every name can be converted to a structure. Sometimes this is because the name just isn’t known to the tools, and occasionally because we’ve falsely identified some text as a chemical. But often, names don’t convert because they contain errors introduced through Optical Character Recognition (OCR).

Many patents (even newly published ones) are digitized using OCR, which can mean slight mistakes in chemical names because OCR classifiers are typically trained to recognize prose, rather than systematic names. Common OCR errors include spurious spaces being inserted into chemical names (often around commas, as is typical in prose), or certain numbers being changed to similar looking letters (the number 1 changed to the letter l, for example).

The next step in our pipeline is to try to correct these. We use a combination of heuristics, dictionary lookups, and third party tools to create correction candidates. Every correction candidate is sent back to the previous step, and if convertable to a structure it is treated as non-corrected names, and sent for structure standardization and storage.

4. Structure Standardization and Storage

Every structure generated in the SureChemOpen data processing pipeline is ultimately processed by what we typically call our “Structure Handler”. The Structure Handler is responsible for processing every chemical generated by earlier steps. This means standardization, error checking, chemical property calculation and storage.

We use a third party standardizer and error checker provided by ChemAxon, which (using a custom configuration) ensures the output is a valid chemical in a consistent form. Automated chemistry extraction can generate spurious chemicals, so by running a series of careful checks (such as checking size, or atomic makeup) we can ensure the storage of meaningful chemistry. Similarly, standardization steps such as de-aromatization ensure that all chemicals are in a consistent form, making chemists lives easier and reducing duplication in our database (see below)

Chemicals that pass standardization and error checking are added to a searchable database, along with a number of derived properties. After being added to this database, the chemicals will appear in search results on SureChemOpen. Often, different names will generate the same chemical structure (think “water” and “dihydrogen monoxide”), in these cases we detect the duplicate chemistry and only store one searchable chemical.

Each chemical successfully processed by the Structure Handler (whether resulting in a new chemical or recognition as a duplicate) will now have a unique ID, which will be sent on (with the originating name) to the Entity Mapper task.

5. Entity Mapping

The final step in the SureChemOpen text annotation pipeline is entity mapping. So far, we’ve seen that documents have been annotated with chemistry, names have been converted to chemical structures, and chemical structures have been stored in a searchable database. But what’s missing is a link between annotations in documents and the chemicals generated from them. Without this information, it’s impossible to find documents that match results from chemical searches; it also makes it hard to show chemical structures for annotations in documents.

The Entity Mapper, therefore, is passed pairs of names and chemical IDs, and updates the database of annotations to ensure that the relationship from annotation to chemistry is recorded.

Image Analysis

Another aspect to the data processing pipeline not mentioned above is extraction of chemistry from images. In the SureChemOpen pipeline, this is done in a similar way to name extraction. Documents are sent to a task that retrieves clipped images from patents, and processes them using CLiDE (a third party tool for detecting chemical compounds in images).

The resulting image annotations are stored in the database, chemistry is standardized and stored by the Structure Handler, and image annotations are associated with detected chemistry. The only significant difference from text processing is that chemistry from images is aggressively filtered because it’s very difficult to prevent non-chemical images from being processed, and false positives can easily be detected.

Intermission

Part two of this series will focus on how we've utilised AWS technologies to build the data processing pipeline. Stay tuned.

Presenting Baton

Fri, 11 May 2012 21:37:00 +0000

We, at Digital Science, are pleased to announce that we open sourced one of the key pieces of our technical infrastructure, baton, which is a general purpose tool to query and run commands on our servers. It provides a basic framework that is meant to be extended in order to fit specific needs such as deployment tasks, monitoring tasks, etc.

This is the first step Digital Science takes towards contributing to the open source community. We have always been trying new technologies and therefore we rely on many open source projects. Thus, contributing to the open source community is a natural step for our company and we very much hope this will be the first of many other projects that we will share.

Why baton?

We strongly believe that automating daily tasks can dramatically increase productivity. One of our goals was having one click deployment of code to development/production servers or quick service discovery. After trying existing tools, we found nothing that was good enough.

That's when we decided to build baton.

Since then, we created a number of baton extentions that have been powering code deployments, service discoveries, handle alert messages, etc.

Internals

Baton is a ruby gem. It relies on EventMachine and AMQP for message passing. The gem defines a basic set of classes operating on top of RabbitMQ.

The initial configuration will setup an input exchange and an output exchange.

On the input exchange, baton will wait for meaningful messages to perform actions (described by each service) and it will output messages to the output exchange.

For example, one could send a message to discover services in production. Baton would receive that message on the input exchange, and every service that responds to that message would write an output message to the output exchange.

API

This is the entry point for input messages. One should extend the API class and add meaningful methods that ultimately use `publish` to publish messages to the input exchange. One example can be found here.

Executable script

Any baton extension should have an executable script that will start the extension service. Here is an example.

Service

The Service is the starting point of any baton extension. The idea of the service is to setup consumers for the input messages arriving from the API. By implementing `setup_consumers` one will allow the consumers to receive messages. Here is an example.

Consumer Manager

This class is an orchestration class that attaches observers to the consumers (like logger, etc), binds the input queues to the correct exchanges, dispatches the received messages to the consumers and updates the observers on changes. One doesn't need to extend this class unless one wants to change its behaviour.

Consumer

This class must be extended in order to process each received message. One should implement `process_message` at least, in order to give meaning to each received message. One can also override `routing_key` in order to listen to specific messages. Here is an example of an implementation.

Channel

Like the consumer manager, this class doesn't need to be extended. It provides functionality to setup the exchanges and add consumers.

Observer

The observer class provides methods to notify observers. It is by default included in the consumers so that the output exchange (and possibly loggers, etc) receive the output messages.

baton-deploy, an extension

One of the most important extensions created so far was baton-deploy, which is a an extension that allows one to deploy code to our servers.

We have installed baton-deploy on each server (development environments included). Each baton-deploy deamon will read a configuration file which will provide information concerning which applications are expected to be running on that server.

Baton-deploy then creates queues with specific routing keys (e.g. such as application_name.current_environment). When the services start listening on those queues for incoming messages, we then use a web interface to interact with baton's api. One could then deploy a given application to all production servers, all development servers, or ultimately target specific servers.

Your turn

You can easily extend baton to perform your own tasks. baton-ping provides what we consider to be a minimal extension to baton. One should note that there is an extra class on `baton-ping` called monitor. This is a good example of what to do with output messages from a baton extension. Together with baton-ping-monitor, it provides a standard way of consuming output messages and do something relevant with them.

If you would like to create your own extension, simply install baton (`gem install baton`) and run the following command:

batonize gem GEMNAME -b

This will create a basic gem structure with the necessary files to create a minimum viable baton extension.

Welcome to the Digital Science technology blog

Thu, 03 May 2012 10:14:00 +0000

With the launch of the new Digital Science website, it's great to announce that our engineering team will finally have a dedicated place to talk about the projects and technologies they're working on, and the events we're attending. We'll also talk in more depth about positions we're looking to fill, and our hack breaks, which are already the stuff of legend.

The Digital Science engineering team come from a wide range of backgrounds and disciplines, from start-ups to academic research, front-end development to natural language processing, open-source and big business. We're also a growing team, and we hope this blog will give you an insight in what it means to be a Digital Science engineer.

Drug discovery is no longer the province of pharma – Introducing SureChemOpen

Mon, 26 Mar 2012 12:00:00 +0000

Today, thanks to our text mining team, users looking for better patent chemistry search have a free means of doing so. Announced in line with the American Chemical Society’s spring meeting in San Diego, Calif., SureChem (our patent chemistry product line here at Digital Science) has introduced a new open interface for researchers needing a more effective means of searching patent literature.

“SureChemOpen” is a free layer added to SureChem’s product line, building off the technology of their flagship product. To date, sophisticated patent search has been largely walled off through costly enterprise software most commonly associated with pharmaceutical research. However, drug discovery is no longer the province of pharma, as increasingly patents are being integrated more into other pockets of the research ecosystem. SureChemOpen is an important first step in making this sort of search more accessible to other members of the research community.

The underlying technology works to make patent chemistry less painful and less reliant on brittle string matching and text string searches by allowing chemists to search how they think – in chemical structures – the interface allowing users to actually *draw* chemical structures and sift through the database that way (ask a chemist, this is big). As seen in other pockets of life science research, structures or entities can often be classified under a myriad of names, inadequately linked to contextual information (if at all), all further exacerbating the sifting problem, and as a result, slowing down research dramatically.

There are other free chemistry databases, most notably resources like ChemSpider from the Royal Society of Chemistry and PubChem from the US National Library of Medicine. Whilst complimenting these resources, SureChem is unique in that its the only free patent chemistry database searchable by chemical structure, each structure linked directly to the patent literature. This not only maps to how chemists think, but makes for an incredibly powerful resource and way of sifting through the wealth of information tucked away in patents.

SureChemOpen is the first part of a re-launch of the SureChem product line, with additional products planned to roll out over the coming months. For more information on the release, see our press release. And do let us know what you think of SureChemOpen. We’d love to hear your thoughts.

Labguru iPad app pairs up with Nature Protocols for seamless integration

Mon, 13 Feb 2012 12:00:00 +0000

You might remember us recently writing about the launch of Labguru’s iPad app, a new product from BioData, one of our portfolio companies. The application is an extension of the web service, a lightweight research management tool designed to alleviate some of the everyday pains of managing and running an academic lab.

The free iPad app builds off of and syncs with the web service, allowing researchers a means of accessing and conducting multiple experiments at the bench, and as of today, will include 20 high-quality protocols from Nature Protocols, a journal of our sister company Nature Publishing Group. The protocols – some of the most fundamental and widely used in the lab – are now fully integrated into the application and web service, as well as interactive, with Labguru building in preset timers and additional information where necessary to further enrich the user’s experience.

Each of the protocols are peer-reviewed, commissioned by the journal’s editorial team. They contain special introductory content, a list of materials needed for the experiment, a guide to anticipated results, troubleshooting and references – all of the necessary components for recreating an experiment in the lab.

“We are grateful that, with the assistance of Digital Science, Nature Publishing Group seized this opportunity to improve access to protocols and relieve researchers of some of the time-consuming tasks involved in working at the bench. We hope other publishers will follow suit,” says Jonathan Gross, Founder and Co-CEO of BioData. “The Labguru web service and iPad app prepare Nature Protocols for bench use and add interactive features, such as progress tracking and preset timers, that prevent headaches—and heartaches—when juggling multiple experiments at once. Results and notes recorded on the iPad app sync right back to the web app, so you’re always organized to easily analyze data and write papers.”
We’re thrilled to see Nature Protocols make a subset of their procedures available in Labguru, putting them in the hands of working researchers in the lab in a whole new way, and we look forward to future developments.

For more on the release, visit our Latest News section. For more on Labguru or to download the free web service or iPad app, visit www.labguru.com.

UPDATE: As of 22 March, Labguru users can now access 50 protocols from Sigma Life Science. For more information, see their press release in our Latest News section.

Citizen science project Zooniverse make data available on figshare

Tue, 07 Feb 2012 08:33:00 +0000

It has been a couple of weeks now since figshare relaunched and the public uploads have been coming thick and fast. It is truly fantastic to see researchers opening up their unpublished data, be it confocal images of the larval brain hemispheres, or a complete unpublished chapter of a PhD thesis. My personal highlight is the sharing of data from the Zooniverse project. Uploaded by Robert Simpson of the University of Oxford Physics Department, he suggested that there may be a need for group accounts, a feature that is in the figshare pipeline.

For those who dont know, the Zooniverse is home to the internet's largest, most popular and most successful citizen science projects.

“The Zooniverse began with a single project, Galaxy Zoo, which was launched in July 2007. The Galaxy Zoo team had expected a fairly quiet life, but were overwhelmed and overawed by the response to the project. Once they'd recovered from their server buckling under the strain, they set about planning the future!”

For this reason, it was really encouraging to see that the Zooniverse team wanted to share their large datasets with the wider world in a citable manner, through figshare. The context for the datasets is described below:

The original Galaxy Zoo project ran from July 2007 until February 2009, before being replaced by Galaxy Zoo 2 and then, in turn by Galaxy Zoo: Hubble. Volunteers were asked to classify SDSS galaxies as belonging to one of six categories - elliptical, clockwise spiral, anticlockwise spiral, edge-on galaxies, 'star/don't know' or merger. These files contain the resulting classifications of nearly 900,000 galaxies drawn from the Sloan Digital Sky Survey.

Ths sucess of Galaxy Zoo inspired the Zooniverse team to open up the platform and there are currently 11 different projects that are being investigated, or have reached completion. Currently over half a million members are helping with projects ranging from 'helping marine researchers understand what whales are saying', to 'help understand how stars form'. The latter goes by the name, 'The Milky Way Project'.

“The Milky Way Project aims to sort and measure our galaxy using infrared data from the Spitzer Space Telescope. Understanding the cold, dusty material that we see in these images, helps scientists to learn how stars form and how our galaxy changes and evolves with time.”

The results of this project are now in, and as well as posting the pre-print publication in Arxiv, the team realised the benefits of sharing the raw data through figshare .

We've also put our bubble catalogue on @figshare figshare.com/articles/Milky…

— Milky Way Project (@milkywayproj) February 1, 2012

Seeing the diverse fields from which users are sharing their objects illustrates just how widespread the inefficiency of scientific publishing is. In order to continue the good work being done, please continue to let you colleagues know about how they can get credit for all of their research on figshare .

New market report shows antibody usage growing, efficacy still an issue

Wed, 25 Jan 2012 16:06:00 +0000

Our colleagues at 1DegreeBio today released their 2012 Antibody Sourcing and Usage Report, a comprehensive report surveying 400 life scientists on their antibody purchasing habits and practices.

The report showed signs of growth in usage levels, with 94% of the respondents claiming they plan on increasing or sustain current antibody usage. With that said, 48% reported that half of the products they’ve worked on were of insufficient quality, highlighting a systemic issue in antibody sourcing.

1DegreeBio is working to help with this efficacy problem, pairing vendor listings with reviews of the products, providing users with more information regarding the efficacy of the product so they can make an educated decision before hitting “buy”. The report echoed support for similar peer review avenues with approaches such as colleague recommendations, as well as journal and independent reviews ranking highest for use in product sourcing.

“These results show that the culture of peer review that dominates the scientific world also shapes the way this community evaluates products,” says Alex Hodgson, CEO of 1DegreeBio. “As with journals, not all review sources are considered equal.”

The 28-page report is available for purchase off 1DegreeBio’s website. For the full release, visit our Latest News section.

Simplifying research 'app' the bench - Labguru

Tue, 24 Jan 2012 08:25:00 +0000

We wrote about the launch of BioData's flagship product LabGuru last December, a lightweight research management tool for use in the lab. The web application, designed by a researcher frustrated with the inefficiency and mismanagement of lab resources, helps users organise the myriad information sources in their lab, plan their experiments, and share results with peers.

Today, BioData unveil a free iPad app that seamlessly integrates with Labguru, allowing researchers a means of conducting and managing multiple experiments at the bench, as well as syncing their results and progress back to their main account. The application is free in the app store, and is designed to not only build upon the user's experience but enhance it, enabling protocol formatting for bench use, embedded timers, space for reference material and even room for additional notes.

We've seen a push in recent years for a more digital experience at the bench - a naturally analog workspace - but none to date who have moved beyond electronic lab notebook functionality and note capture to a more intuitive interface that allows researchers to manage *all* of their experiments easily with integrated, easy to follow protocols. Labguru's iPad app not only allows for researchers to have a place for their notes and protocols, but also seamlessly syncs the results and additional data to their main account for later reference.

Our colleagues at BioData have put together an informative video to walk you through the additional functionality. We hope you'll give the tool a try. We'd love to hear your thoughts

For the full announcement, visit our Latest News section. You can download the app here, and find out more about Labguru and BioData on their website.

Nature Publishing Group, Digital Science and Palgrave joint statement on proposed SOPA and PIPA legislation in the US

Wed, 18 Jan 2012 08:20:00 +0000

The following is a statement that we have issued today together with our sister businesses, Nature Publishing Group and Palgrave.

Nature Publishing Group, Digital Science and Palgrave joint statement on proposed SOPA and PIPA legislation in the U.S.

Nature Publishing Group (NPG), Digital Science and Palgrave provide a range of publications, information and digital services to a wide variety of customers, including scientists, scholars, teachers and students, as well as members of the public. Our customers are spread across the world, many of them in the U.S.

We are concerned that the approaches proposed by the Stop Online Piracy Act (SOPA) and Protect IP Act (PIPA) currently under consideration by the U.S. federal government are likely to create unintended counterproductive effects in many of the communities we serve, and to the functionality of the internet itself. Therefore we do not support SOPA or PIPA.

Our businesses are based on attracting, creating and distributing valuable content. We therefore support appropriate and proportionate initiatives to protect copyright and other forms of intellectual property for the benefit of its creators and society as a whole. In particular, we support actions that effectively target the systematic illegal sharing of copyrighted works. But we do not support approaches that would undermine and put at risk the legitimate activities of our customers and other responsible users.

As innovative and progressive publishing and technology companies, we seek to enable the open exchange of ideas, especially in scholarly communities, because this is ultimately the main way in which human knowledge is created and disseminated, a process that we exist to support. We believe that the proposed legislation could unintentionally harm this process. In addition, the proposed interference with the infrastructure of the internet risks unintended consequences that could reduce its benefits for everyone.

For these reasons we feel it is important to clearly state our position on the proposed legislation.

Timo Hannay, Managing Director, Digital Science

Steven Inchcoombe, Managing Director, Nature Publishing Group

Dominic Knight, Managing Director, Palgrave

If you have comments or questions, please contact me, Timo Hannay at t.hannay [at] digital-science [dot] com.

Making data citable, searchable and discoverable

Tue, 17 Jan 2012 16:12:00 +0000

The topic of reproducible science has caught the attention of government officials, academics and the mainstream media over recent months, with heightened scrutiny over current access policies, reward mechanisms and the need for transparency. For instance, in the US, the call for information that will inevitably shape the nation’s policies on data has just closed. Across the pond, the science minister David Willetts has echoed a similar sentiment in regards to the Coalition government’s growth agenda, and the role of transparency and access to funded research in the research enterprise. And Professor Colin Blakemore from Oxford recently raised the issue on the Today programme on BBC Radio 4.

The discussion is shifting, shaking up entrenched practices that researchers have relied upon for decades.

At Digital Science, we develop software tools that help bridge this gap in research practice, and explore opportunities in the digital space that can help make research more efficient. We’re thrilled to announce the relaunch of figshare , a community-based open data platform that we announced our support in last September, who today unveil a new look and feel as well as additional functionality. The project was founded by Mark Hahnel, developed while he was finishing his PhD at Imperial College.

figshare was founded on the principle that researchers should get credit for *all* of their research, in a system where “impact” is often solely based on published journal articles. The platform enables researchers to make their figures, datasets, images and videos publicly available, providing the researcher with a citable, searchable and discoverable endpoint – all for free. This is important not only for the supplementary data accompanying one’s experiment, but even negative results.

New functionality includes dramatically quicker upload (reduced down to a matter of seconds), unlimited storage space for data that is made publicly available on the site, and as of today up to 1GB of free storage space for users looking for a secure, private area to store their research. Users of the new site will maintain full control over the management of their research whilst benefiting from global access, version control and secure backups in the cloud.

You can access the full press release here. For more information on figshare and to stay in the loop for upcoming developments, follow them on Twitter at @figshare or visit their website. And do tell your friends. – we’d love to hear what you think.

Data-driven science, and lessons from the research world - Strata

Mon, 16 Jan 2012 08:18:00 +0000

Did you miss last month's Strata online conference "Moving to Big Data"? I joined other industry experts, offering my thoughts on big data from the viewpoint of a researcher in the sciences. In this talk, I posit that we're not only "getting this wrong" in the academic research context, but that these problems are also finding their way in to the broader research enterprise. Tune in to hear more about how financial, social and technical constraints are affecting the way we perceive and handle information, and what the best means forward to fixing the breaks in the system is.

You can browse through the other top-notch presentations from representatives from IBM, Intel, Forbes and more.

I'll also be speaking on the classification "junk data" and data sharing, as well as data privacy at the next (in-person, not online) Strata Conference this March in Santa Clara, California. Interested in joining? Plug in the code KT30 for a 30% discount. I hope to see you there.

Bringing the lab into the 21st century

Mon, 05 Dec 2011 19:32:00 +0000

In today’s life sciences laboratories, researchers and lab managers are faced with a vast array of information sources – from journal articles, lab notebooks and protocols to the specimens, ordering and tracking information and calibrations. The way this information is managed is less than practical given the technology we have at our fingertips – distributed and stored between a mix of poorly annotated Excel spreadsheets, post-it notes, invoices and color-coded stickers to help distinguish which plate belongs to which experiment.

Needless to say, we can do better to get the administrative burden out of the way, and free up researchers to do better science, helping them keep track of the progress of their own work and collaborate with others.

Biodata, one of our portfolio companies, is working on just that. They have unveiled a new research management tool designed to help make researchers more productive, and enable better, more efficient science. The tool is called Labguru, a lightweight research management tool for use in the lab. The announcement came from Denver, where BioData are exhibiting at American Society for Cell Biology’s Annual Meeting. Labguru comes in the form of a web application, designed to help researchers better organise the various bits of information in their lab, plan their experiments, share results with peers and track progress.

The tool builds off their earlier product, BioKM, as well as the LabLife software (assets that BioData acquired earlier this autumn.

Labguru is free for all individuals. Additional functionality and sharing capability (think, expanding the use to other colleagues and/or peers in the lab) can be added for a nominal monthly fee. For a full look at the feature’s visit their web site, or have a look at their press release in our Latest News section.

This is only the first step in making research at the bench more efficient, and we welcome your feedback. Help us spread the word, and do let us know what you think. Stay tuned for more from BioData in the new year.

"Data for Impact" workshop (IDCC)

Mon, 05 Dec 2011 08:17:00 +0000

This Monday, Kaitlin Thaney, Mark Hahnel (FigShare) and Daniel Hook (Symplectic), will be participating in a workshop on data and impact at the 7th Annual International Digital Curation Conference (IDCC) in Bristol. They'll be joined by Brial Hole (Ubiquity Press), Mark Thorley (NERC) and Heather Piwowar (Dryad, Total Impact). For more, visit the workshop home page.

A new way to view literature on nature.com

Wed, 02 Nov 2011 16:43:00 +0000

As of today, you can now view articles on nature.com in a new way. We're proud to announce new functionality, brought to our sister company Nature Publishing Group by Labtiva, a Boston-based startup we recently announced our investment in.

The tool comes in the form of a web reader that allows users to access nature.com content, in an editable, interactive form that you can sync to your ReadCube desktop application. The web reader also provides an integrated view of the supplementary information and associated News & Views, and functionality including direct links out to the literature with in-line references.

This builds on Labtiva's recently launched product ReadCube - a free, cross-platform desktop application that provides researchers with a better means of organising, accessing, and discovering literature in their field.

The ReadCube web reader is currently available on Scientific Reports, Nature and 18 Nature research journals: Nature Biotechnology, Nature Cell Biology, Nature Chemical Biology, Nature Chemistry, Nature Climate Change, Nature Communications, Nature Genetics, Nature Geoscience, Nature Immunology, Nature Materials, Nature Medicine, Nature Methods, Nature Nanotechnology, Nature Neuroscience, Nature Photonics, Nature Physics, Nature Protocols, and Nature Structural & Molecular Biology. ReadCube integration will be added to additional journals on nature.com over time.

ReadCube's integration with nature.com is only the beginning, with plans to extend their reach and availability to other content providers in the sciences in the future. Do stay tuned for future developments.

In the meantime, we hope you'll give the web reader a try.

For more information on Labtiva or their product ReadCube, visit www.ReadCube.com. You can also read our official press release here.

Welcoming Labtiva, and unveiling their product ReadCube

Tue, 11 Oct 2011 16:39:00 +0000

Today we’re happy to announce both our investment in Boston based start-up Labtiva as well as the global release of their first product ‘ReadCube’ - a software application designed to provide researchers a more efficient and intuitive means of managing and discovering relevant literature in their field. The company joins Digital Science’s portfolio of research tools as our first content management offering.

Labtiva was founded by two Harvard alumni, Robert McGrath (a computer scientist) and Sinisa Hrvatin (a biologist). The two paired up during their studies and began work on a tool that allowed for Sinisa in his research to better manage the deluge of content in his day-to-day, as well as keep on top of new literature relevant to his topic area. ‘ReadCube’ was born out of that frustration experienced in the lab, as a lightweight application that allows researchers to create and manage their content collections, as well as discover other relevant literature.

The application also features a personalised recommendation system that updates daily based on the user’s existing library, minimising the chances they’ll miss out on advances in their field.

The tool is free to download and use, and we welcome your feedback. The team will be rolling out additional functionality over the coming months, so do stay tuned for developments. In the meantime, we hope you’ll give it a go.

For more information, you can read the official press release, visit the ReadCube site, or follow them on twitter @readcube.

Update: Labtiva recently held their launch party on November 4 in Cambridge, MA. For photos and video of the event, visit their Facebook page.

Celebrating the women of Digital Science

Fri, 07 Oct 2011 16:35:00 +0000

For this year’s Ada Lovelace Day, I wanted to take a moment to give a nod to the exemplary women I am fortunate enough to work with on a day-to-day basis here at Digital Science. As Tim O’Reilly so nicely pointed out in his Ada Lovelace post from 2009, there are many ways to contribute to the world of technology, beyond strictly doing code. My female colleagues here exemplify that, each offering invaluable skill, expertise and utility to the company and the cause: of truly making research more efficient through better use of technology.

So this one is for the ladies.

It’s perhaps most appropriate to start this post off with a thank you to Annette Thomas, the CEO of Macmillan Publishers, our parent company, and one of Digital Science’s most ardent supporters. Annette’s vision, stewardship and willingness to take a risk has been instrumental to the creation and growth of the company. Annette has a long history with the company, joining Macmillan in 1993 as the cell biology editor for Nature magazine. She has served in a number of editorial and publishing roles since, before being appointed the CEO of Macmillan in 2007, including serving as the Managing Director of Nature. During her time, she’s been a key proponent of innovation in the company, and we’re incredibly grateful to have such a phenomenal woman at the helm and involved in our work.

Moving to our core team, we have Caitlin Trasande as our head of Science Metrics, where she works with a team of developers and computational linguists on tools and technology for measuring scientific contribution in the digital age. Caitlin came to us from Nature Publishing Group, where she used to manage the project "Nature Trends". She also has been known to dabble in neurobiology, receiving her PhD in the subject from the University of Chicago.

Jumping to our text mining division, we have two women - one covering European sales (based here in London), the other covering the North American market (based on the east coast in the US), working to support our text mining company SureChem. Babeth Piveteau, the European Sales Manager, joined us with 15 years of experience in sales and business development at places like MicroPatent, LexisNexis and MatrixWare - an information retrieval start-up where she helped extend the business in areas such as semantic search and machine translation. Janice Stevenson, SureChem’s Sales Manager for the Americas is one of the most recent additions to the Digital Science family, also joining us from Micropatent as well spending time at SAP and American Express.

Complimenting the development work for SureChem we have Lezan Hawizy, one of our computational linguists on staff, who came on board from the Murray-Rust Lab at Cambridge University, where she worked on natural language processing and text mining within chemistry - key skills needed to support our text mining division and the cheminformatics work that SureChem conducts.

Supporting out Research Tools division, we have Laura Thomson, one of our product managers, who has spent time working in product roles at Science Navigation Group, British Standards Institute and Thomson Reuters. She also has a background in the life sciences, holding a PhD in biochemistry, as well as an MBA. Her insight into building out teams and products in the research space is no doubt sharpened by her diverse academic and business experience.

The following women work across the various business divisions, providing invaluable services to the staff internally, to the companies we incubate and partner with, as well as the community (myself falling into the last category there :) ).

Liz Pitt, our account manager, works hand-in-hand with our finance team to manage our books, her fascination lying in maths and statistics. Prior to joining Macmillan (and consequently, Digital Science), Liz worked at Barclays, after retraining herself as an account following years as a primary school teacher.

There’s also our Marketing Manager, Georgina Gurnhill, who is in many ways “coming home” to our Crinan Street offices here in London where she used to work for our sister company, Nature. Georgina has held senior marketing positions at some of the top publishing companies and information providers, from Nature, to Elsevier, Thomson Reuters and the Institute of Physics. Her knowledge of the scientific research market is extremely helpful in expanding the reach and impact of our internal projects as well as for the companies we partner with.

And last but not least is Amarjit Myers, or “Marj” as we call her - our office manager. Amarjit joined us from NESTA FutureLab, and serves as the office linchpin, making sure all of the trains are running on time, and each division, employee and everything in between are adequately supported and welcomed - an oft overlooked, but nevertheless critical position in the company. We are fortunate to have her here to keep us all in line. With that, I’d like to tip my hat to my esteemed female colleagues. It is truly heartwarming as a female technologist myself to be able to say on Ada Lovelace Day that I’m a part of both a technology company and one with a science bend - two main areas that today is designed to celebrate - alongside the colleagues listed above.

I hope you’ll take a moment today to think of the women who inspire you. For more stories like this, visit Finding Ada.

Reinventing Discovery

Mon, 03 Oct 2011 16:34:00 +0000

Every movement needs it articulators and storytellers: open-source software has Eric Raymond; the web has Tim O'Reilly. A strong candidate for that mantle in e-science is Michael Nielsen, who keynoted at this year's Science Online London and whose much anticipated book, Reinventing Discovery , is about to be published by Princeton University Press. I was lucky enough to see a preview copy a few weeks ago and my thoughts are now out in Nature Physics . The review is freely available online until November 1, so access it while you can. More importantly, read the book – it's very good.

Reflecting on this year's Sci Foo

Fri, 23 Sep 2011 16:32:00 +0000

It's been slightly over a month since Science Foo Camp ("Sci Foo" for short), and we've been lucky to have a number of Sci Foo alum visit us in the weeks since the event. We most recently were joined by Scott Snibbe, an interactive designer who is working with Bjork on a stunning app for her latest (science-themed) album Biophilia. There's also been some outstanding post-Sci Foo coverage, including a series of essays and videos in this month's Edge, coordinated by Sci Foo friend John Brockman, culling together some of this year's campers' key take-aways. It features Nobel Laureate Frank Wilczek, video of some of this year's campers and words from others (I won't spoil all of the surprise... )

Our very own Timo Hannay reflects on his experience - quite a daunting task as he'll detail below. Here's an excerpt from his essay on Edge, where he describes what it's like to even attempt to sum up the highlights from the weekend. You can read more about Timo's Sci Foo here:

"Trying to sum up Sci Foo is a bit like trying to describe the Mandelbrot set. You can give the formula, but that doesn't begin to communicate the experience. Alternatively, you can try to summarise the overall shape, perhaps by outlining some of the main events and themes, but to me that somehow misses the essence too. So I'll make do with describing some of the personal delights that I encountered during my wanderings among the intellectual whorls and crenulations that make up this event, even at the risk of missing the big picture (for which it is perhaps best simply to imagine the following experiences multiplied by a few hundred). [...]"

Timo and I are honored to be a part of the organising team, along with our colleagues Tim O'Reilly and Sara Winge of O'Reilly Media, and Chris Dibona and Cat Allman from Google. For more information on the event, visit the conference site where you can read more about the history of Sci Foo and check out the excellent video coverage from year's past, put together by our colleagues at Nature Video.

Many thanks to John again for his coverage of this year's event. For more, visit the full feature at Edge.

Even 'free' has its costs

Mon, 19 Sep 2011 16:29:00 +0000

Last week Biodata, one of the companies in which we've invested, announced that they had acquired a similar service, LabLife. This is a very significant development and I'd like to say a couple of things about it.

First, these are two amazing teams with two outstanding products. It goes without saying that we're big fans of Biodata – after all, we've invested in them and work closely with them every day. But it's worth adding that we've known the LabLife team for even longer and have a huge amount of respect and fondness for them too. The founders are Ken Fan, Melina Fan and Benjie Chen, who also created Addgene, the definitive plasmid repository and sharing service. I've visited them several times and their vision, intelligence and execution are second to none. So I'm delighted that these two teams found a way to work together to keep LabLife going, even once the founders had decided that they couldn't support the service to the degree that their users deserved. In fact, Biodata has committed to maintaining LabLife for at least another two years, and keeping it free for existing users – whilst of course hoping that at least some of them will switch to Biodata's own (paid) service.

This leads to my other point: free versus paid. I'm a big fan of freemium business models – we're already using them at Digital Science and will be announcing more examples in due course. As with so many other things, my thinking here has been heavily influenced by Tim O'Reilly. His talk at the 2008 Tools of Change Conference, 'Free Is More Complicated Than You Think' (a title taken from Scott Adams) is perhaps the best explanation yet of this approach, at least from a publishing point of view. (And I like to think that in my own small way I contributed back: Tim's description of ASOS and related businesses was a result of my introducing him to an old school friend, Quentin Griffiths, who's a founder of those companies.) The point is that 'free' isn't just about charity, or the logical conclusion of self-defeating a price war; neither is it just about slapping ads onto every content page and hoping for the best. But with the right accompanying revenue-generating strategies it can be a sound, even optimal, business strategy.

Yet the complexity of 'free' cuts both ways, and nice as it might seem to users, it's not always optimal even for them. This is nicely summed up in a quote (attributed to Andrew Lewis) from Eli Parser's book, 'The Filter Bubble': "If you’re not paying for something, you’re not the customer; you’re the product being sold."

In other words, if someone else if paying for the service – and someone somewhere, whether an advertiser, an investor, a philanthropist or whomever, is always paying – then whose interests does the service provider have at the top of their mind? Even more profoundly, if someone else is paying then how long will they continue to foot the bill, and what happens to me as a user if they stop? This was effectively the situation confronted by LabLife when they began to doubt that their existing approach would be sustainable, and without the deep pockets of a large organisation they had to make a decision quickly. To their huge credit, they secured a great deal and a wonderful new home for their loyal users.

Sometimes it feels as if the idea that online services have to be free at the point of use has been elevated to the web business equivalent of scriptural doctrine. But to echo Tim O'Reilly's thesis, it's more complicated than that. We're all still finding our way in this exciting new landscape and experimenting with a variety of approaches. 'Free' and 'freemium' definitely have their places, including in our business, but the evidence from this particular episode points to the conclusion that when it comes to lab management tools, modest user charges may be a more reliable approach – and therefore better for users and providers alike – than trying to make everything free, because even 'free' has its costs.

Welcoming FigShare, an open data project, to the Digital Science family

Thu, 08 Sep 2011 16:28:00 +0000

We're thrilled to welcome FigShare, an open data project founded by Mark Hahnel, to the Digital Science team. Mark will join the Research Tools team as a Product Manager for FigShare, where he will continue to develop this community resource with support from Digital Science.

Mark started work on FigShare, while working towards his PhD at Imperial College. He was frustrated by the duplication and waste in scientific research that results from a lack of data sharing. FigShare provides a means to publish your data in an easily searchable, sharable and citable manner. The ultimate intention is to enable and encourage users to share all of their data, whether published, unpublished or containing negative results.

Over the last few months we've announced a number of investments and partnerships with teams that are building software tools for research. This reflects our hybrid approach, which combines internal software development with a startup incubator. However, our relationship with FigShare is slightly different to our other collaborations to date because it will retain autonomy as a non-commercial community project while receiving support from Digital Science.

We believe that there's much to be learned from ambitious projects like FigShare, and much to be gained by science if they can become useful and ubiquitous.

For more information, visit http://FigShare.com. You can also read more in our press release.

We're off to (Science Foo) camp

Fri, 12 Aug 2011 16:24:00 +0000

It's that time of year again - time for Science Foo Camp ("Sci Foo" for short), and we're off to Mountain View, California.

The event is now in its sixth year, an unconference that brings together 200-300 leading scientists, writers, and hackers each summer at Google's headquarters all linked together by an interest in science.

The concept of “Foo” originated in 2003, as an experiment by Tim O'Reilly and his colleagues to bring together 200+ “Friends of O'Reilly” Media (which is where the “Foo” in Science “Foo” Camp comes from). He let them create the agenda from start to finish, putting little if any boundaries to what can be discussed.

The results were astounding, redefining the way we think of cross-disciplinary collaborations today and challenging one's conventional idea of a “conference”. In 2006, O'Reilly paired up with Nature Publishing Group (and now, Digital Science) and Google to hold the first Science Foo Camp (now known as “Sci Foo”), inviting over 200 leading scientists, technologists, writers and other thought-leaders to gather at the Googleplex for a weekend of unbridled discussion, demonstration and debate.

Timo and I are delighted to join the folks at Google and O'Reilly on the organising committee, representing Digital Science and our sister company NPG - where Timo used to run the online division. Sci Foo kicks off tonight, wrapping up this Sunday, and we're off to Mountain View. Timo and I will also be joined by our Head of Text Mining, Nicko Goncharoff, and Jonathan Gross - founder of BioData, one of our portfolio companies.

The event may be invitation-only, but keep an eye out online for chatter about the event. The official hashtag is #scifoo and we'll be doing our best to tweet at @timohannay, @kaythaney, @digitalsci. You can also check out some of the video coverage from our 2009 and 2010 events put together by our Nature video colleagues, Adam Rutherford and Charlotte Stoddard.

And an enormous thank you to our co-organisers Tim O'Reilly and Sara Winge from O'Reilly Media and Cat Allman and Chris Dibona from Google. The event would not be possible without their help. Are you attending this year's event? Do come say hello to Timo and myself. We look forward to seeing you there.

Announcing the Digital Science Catalyst Grant

Tue, 02 Aug 2011 16:22:00 +0000

Since we launched in December 2010, the Digital Science team has been hard at work not only crafting our own software solutions for science, but also engaging with and supporting a range of start-ups and innovators who are working to make research more effective. Today we're launching a new initiative set to push that engagement and investment in technology one step further.

We're thrilled to unveil the Catalyst Grant – a programme designed to unleash the most promising new ideas for software in science. This provides grants up to £15,000 (around $24,000) each to fund exciting innovations, and to take them from concept to prototype. They also come with the opportunity to work with the Digital Science team to help refine, develop and promote the innovation. In this way we hope to lower barriers and foster greater creativity in information-technology solutions for science.

The process is simple and fast. Applicants are asked to submit a short proposal detailing their idea and the impact they envision it having in scientific research. Those who pass an initial screening are then asked to present their idea, preferably in person at a Digital Science office (in London, New York or Tokyo), following which a final decision is made. Applications are open now and are accepted at any time.

For more information, and to apply, visit the Catalyst Grant web page. You can also read our press release. We love new ideas, and look forward to hearing yours.

Welcoming 1DegreeBio

Wed, 06 Jul 2011 16:19:00 +0000

Today, we're happy to announce a new investment in 1DegreeBio, a Toronto-based startup that provides online information about vital research reagents such as antibodies, enabling researchers a more efficient means of sourcing high-quality biomedical products - the physical materials on which almost all biomedical research is based. This serves a pressing need for biologists and other life science researchers, who often find it hard to know what is available, which suppliers can provide it, and how best to attain the materials in a timely manner. Many millions of dollars and man-years of effort are wasted every year in biomedical research as a result of inadequate reagents and roadblocks that prevent efficient acquisition of biological materials, and 1DegreeBio offers a promising, lightweight solution.

1DegreeBio was founded by Alex Hodgson, their CEO, and Ron Yeung, Director of Technology, both alumni of the Structural Genomics Consortium, led by Dr. Aled Edwards. The idea came from frustration in the lab in discovering and sourcing the best biological materials for experimentation - an oft overlooked bottleneck to scientific research. To date, the process is at best les than straightforward and labor-intensive, with few (if any) resources existing that serve the needs of the scientist in letting them navigate listings, easily acquire materials, review products, and so on. Each of these things is almost taken for granted in the age of eBay, Google and Amazon, where products of all shapes and sizes are just one click away.

Our investment in 1DegreeBio builds upon our current offering list in the research tools space, joining other portfolio companies Symplectic Ltd., SureChem and BioData. Stay tuned over the coming months for more on 1DegreeBio. We'll be posting the latest here to our blog and on Twitter at @digitalsci. And for more information on our recent investment in 1DegreeBio, see our formal press release.

What's in a name? A message from Nature's editor

Tue, 21 Jun 2011 16:18:00 +0000

This has almost nothing to do with technology and science, but it seemed too good a story not to share. Here is a message received recently from Nature's editor, Phil Campbell:

Dear all

A number of NPG colleagues (and others) have kindly sent me links to various news stories about the events in the little town of Phil Campbell, Alabama, USA. To save anyone a similar effort, and for the fun of it (intermingled with sadness), here’s the history from my point of view.

Background: see

http://www.nytimes.com/2011/06/18/us/18alabama.html

http://www.bbc.co.uk/news/world-us-canada-13818591

http://www.alabama.travel/events/phil_campbell_hoe_down.html

I first heard about Phil Campbell AL (PCA) in 1995, when I’d just come back to Nature. Joe Palca, of National Public Radio and an ex-Nature colleague, had sent my then-assistant Mary Sheehan a news clipping about a gathering of Phil Campbells in PCA.

Ever since, that clipping has been pinned up on a board in my home and I’ve been hoping to find an opportunity to go there.

Earlier this year, another ex-Nature colleague invited me to visit her and colleagues in her institute in Huntsville AL, and to visit NASA and other centres there. And, she said, I should go to the nearby gathering of Phil Campbells in PCA happening today, 18 June. To my delight, this coincided with my trip to the West Coast for meetings next week. So I made arrangements to stop off in Alabama on the way over. The gathering had been featured in the Wall St Journal and was clearly turning into something of an international event for those of us with that moniker.

Then the tornadoes hit. They devastated PCA and also caused severe disruption and damage in Huntsville. So I decided that I’d best postpone my visit.

Then the PCA organisers decided to go ahead anyway and run the festivities as a fund-raiser. Hence the news in the links above.

It was too late to reverse my arrangements, but I’ve looked at the video on the BBC website and feel a twinge of guilt for abandoning PCA and, of course, sympathy for the displaced families in the town.

So I’ve made a donation, and I hope you’ll forgive me for encouraging you readers of my little saga to look at the video and then maybe do the same.

All the best

Phil

Arfon Smith and David Eagleman - Digital Science Speaker Series

Tue, 31 May 2011 16:15:00 +0000

In the last few months, we've had the pleasure of hosting a number of visitors here in our London office and have them share their expertise as part of the Digital Science speaker series.

Our guests have been kind enough to let us share the videos of their talks here on the blog.

The first is a talk by Arfon Smith, technical lead of Zooniverse and GalaxyZoo. Arfon is an astrophysicist by training as well as a software developer, formerly working at the Sanger Institute before taking the technical reins for Galaxy Zoo - one of the most notable citizen science projects of late.

Galaxy Zoo launched in July 2007 and asked members of the public to classify 1 million galaxies by their shape. Four years later, a community of nearly 400,000 citizen scientist have between them provided more than 250 million data classifications across eight Zooniverse projects. Galaxy Zoo worked because there were *only* 1 million galaxies; in the future datasets are projected to reach volumes that will make purely human-based analysis impractical.

In this talk, Arfon reviews the current state of crowdsourced citizen science and argues that smarter approaches must be developed if citizen science is to continue to be a useful tool for data-rich research

Speaker Series: Arfon Smith, GalaxyZoo from Digital Science on Vimeo.

The second talk is by neuroscientist and author, David Eagleman. David is a researcher at the Baylor College of Medicine, where he directs the Laboratory for Perception and Action and the Initiative on Neuroscience and Law. Among other things, he is well known for this work on time perception, synesthesia, and the legal implications of neuroscience. He is also the bestselling author of 'Sum', a sublime collection of essays providing (presumably) fictional accounts of the afterlife. And in 2010 he published one of the first 'books' to make use of the iPad's potential: 'Why the Net Matters'.

In this talk, David tells us more about his latest work: 'Incognito: The Secret Lives of the Brain'.

Speaker Series: David Eagleman, Neuroscientist / Author from Digital Science on Vimeo.

Stay tuned for more from our speaker series. And do drop me a note if you're passing through town - we're always open to new talk ideas.

Digital Science in the news and on the Web

Tue, 03 May 2011 16:13:00 +0000

These last few months we've been busy getting the word out about our new division and what our vision is for the future of digital research and technology. Here are some places you can hear (and read) more about us:

* "Strata Con and Big Data" - Episode 0 of the InSCIght podcast - a scientific computing podcast put together by the folks at Enthought, featuring Anthony Scopatz (Enthought), Puneet Kishor (open data advocate, Creative Commons, University of Wisconsin-Madison), Peter Wang, and myself. The podcast is a follow on to O'Reilly's Strata Conference - a data science event held this past Feburary in California.

* "Digital Science" - feature article in Nodalities, a magazine put out by our friends at Talis for the semantic web community. This article (on page 4) outlines the thinking behind the start of Digital Science, the ecosystem in which we exist, and our vision. I highly recommend you take a look.

Would you like to learn more about Digital Science? Check out our events page to see where my I'll be next or drop me an email. I'd love to tell you more.

Digital Science compiles ranking index of top research institutions

Thu, 24 Mar 2011 16:05:00 +0000

For the last few months, our metrics team, headed by Caitlin Trasande, have been hard at work on a project with Nature Publishing Group. Yesterday, the fruits of that project went live - the Nature Publishing Index Global Top 50 - a ranked list of the top 50 institutions in the world, according to how many primary research articles in Nature research journals.

Caitlin joins us here on the Digital Science blog to tell us more about the work involved in pulling this index together - not an easy nor straightforward task. Science metrics is an area of the research process that's riddled with inefficiencies and ambiguity, yet another area that can benefit from better use of technology and technical know-how. The Global Top 50 represents the first example of what can be done in this space, in pairing open and meaningful data from the science publishing world with information and metrics from other resources to surface new insights - in this case, the number of Nature articles published by institution.

And with that, I'll hand the mic over to Caitlin:

Technically, this begins as a task of deciphering and mapping of the myriad institutional names (and forms) used by authors in Nature research journals and in many cases, harmonising their affiliate institutions and campuses. Take for example "Max Planck Institute" - a network of nearly 80 research institutes in Germany. In some cases, a judgement call has to be made about when to aggregate data from those institutes, or, in the case of individual campuses in the University of California system, when to let those institutes stand alone. The data is messy, presenting matching issues alongside scalability challenges, which the Metrics team has been diligently working to make better sense of in collaboration with Nature Publishing Group.

The organisational structure for modern scientific research institutions is complex, often poorly documented (if at all), and multi-dimensional - now including more traditional divisions based on college or discipline, specialised affiliate institutions, satellite institutes, affiliated research hospitals, and so forth. Defining an institute's boundary is challenging. Mapping journal articles and authors to that information, as a result, calls for a clean baseline data set to start with. Which was only the starting point for the Metrics team.

The Global Top 50 gave us the opportunity to peek under the hood and see first hand what the challenges were in this sort of task, helping us gain a better understanding and stronger technical grasp on how we can help in this space, providing better, more representative tools and information for the scientific community. Do stay tuned.

Also, we'd love to hear what you think about the index. Feedback can be sent to institutes@digital-science.com. For more on the index, check out Nature Publishing Group's press release. Kudos again to Caitlin, Dan Hagon and Johannes Goller on the Metrics team for their tremendous work on this project. Keep up the good work!

The future of science from Sci Foo

Thu, 03 Mar 2011 16:03:00 +0000

Planning is now in full swing for this summer's Science Foo Camp ("Sci Foo" for short), an unconference we co-organise each year with O'Reilly Media and Google. We're pleased to share four new videos from last summer's event, put together by our colleagues at Nature Video.

You may remember back in December us writing about the first of a series of videos, a short piece capturing the spirit of last year's Sci Foo in the words of some of our participants including Andrew Marr, Aleks Krotoski and Marcus du Sautoy. The remaining videos in that series are now live on the Sci Foo site (and posted below), each posing a specific question to our Sci Foo friends on the future of science and the environment. We encourage you to give them a look. And many thanks again to Adam Rutherford and Charlotte Stoddard at Nature Video for their brilliant work. On to the videos ...

Part 1: What worries you? We asked a skeptic, an internet guru, a philosopher and some scientists about their worries for the future.

Part 2: Imagining the future Which technology will have the greatest impact in the next 10 years: the internet, genomics or geo-engineering? Are you optimistic about the future, or pessimistic? Find out how some of the sharpest thinkers imagine the world in the next decade and in 2050.

Part 3: Food & energy How do we feed 9 billion people? How can we reduce our carbon dioxide emissions? Campers at the 2010 meeting came up with a variety of answers.

Part 4: 1 billion dollars How do we feed 9 billion people? How can we reduce our carbon dioxide emissions? Campers at the 2010 meeting came up with a variety of answers.

Science Online - from North Carolina to the British Library

Wed, 02 Feb 2011 15:57:00 +0000

There has been a wealth of post-conference coverage following last month's Science Online in the Research Triangle, North Carolina. From blog posts, to video, images and tweets, the conversation truly has carried on, even though the event has come to a close. My slides from Saturday's technology discussion are now online, as well. The session evolved organically, my slides just a starting point meant to stimulate the discussion on science and its relationship to technology and tools development. Many thanks to all who participated, both in person and through online channels.

To build on the Science Online buzz, I'm pleased to announce that planning for Science Online London (#solo11, @soloconf) is now underway. We're thrilled to be joining esteemed colleagues Martin Fenner (Hannover Medical School, PLoS Blogs) and Lou Woodley (Nature Network) in planning this year's event as an organiser and sponsor, which will be held at the British Library on September 2-3, 2011.

We'll be building on previous years events to further explore science in the digital age - from the way it's communicated and conducted, to the tools and technology needed, to the social issue of rewards and incentives. Topic areas we hope to add to the mix and/or expand upon include big data, changing paradigms in science communication, open science, infrastructure / cloud computing and even bibliometrics. If you have any ideas or programme suggestions, please feel free to contact either myself or Lou. We'd love to hear from you.

Interested in sponsoring Science Online London? Drop me an email. I'd be happy to send over information on sponsorship opportunities, so we can keep Science Online as accessible as possible.

Don't forget to mark your calendars -- Sept. 2-3, at the British Library. More information to come. We hope you can make it.

On the road: Science Online '11

Thu, 13 Jan 2011 15:55:00 +0000

It's a new year, and Digital Science is hitting the road.

First on the list (for me, at least) is Science Online North Carolina, organised by Bora Zivkovic and Anton Zuiker in the Research Triangle. This is their fifth annual meeting, held from 13-15 January on a wide range of topics on science and the Web. The event has evolved over the years from its roots in scientific blogging and communication, and we're pleased to join the program this year. This will be my first Science Online (NC), having had missed it previously due to last minute scheduling issues. I'm ecstatic I'm able to make it this year and look forward to joining the conversation.

Speaking of conversation, I'll be leading a discussion session first thing Saturday morning at 9 am EST. The session is called "The Digital Toolbox: What's Needed?" and will explore how we can better use technology to our advantage and make research more efficient. Science and technology have been inextricably linked for quite some time, but there still remain gaps in fully utilising the potential technology presents to truly accelerating and enhancing scientific research. My hope in this session is to engage the group in a discussion to help identify the remaining pain points and inefficiencies and explore the underlying infrastructure, tools and technology needed. I hope you'll join me.

The rest of the agenda is chock-full of thought-provoking sessions, with offerings for those interested more in science communication, data issues, and even a few this year on Open Science, citation and alt-metrics that I'm especially looking forward to. Bora and Anton have done a fantastic job at making this an event that caters to a wide breadth of interest, expertise, age, etc -- not an easy task. I encourage you all to have a look.

Can't make the event? Never fear. Most of the sessions will be livestreamed and available here on the "Watch" page. (Note, the session are in EST, so plan accordingly.) You can also follow on twitter: the conference hashtag is #scio11, and I'll be tweeting at @kaythaney and @digitalsci. Keep an eye out for news on the London-based Science Online in the coming days, as well.

More to come from the road in the coming days, from both Science Online NC and the next stop on the trip - "Beyond the PDF" in San Diego, CA. Will you be in attendance? If so, I hope you'll come say hello.

A look at this year's Science Foo Camp

Mon, 20 Dec 2010 15:49:00 +0000

The first video from last summer's Science Foo Camp ("Sci Foo" for short) is now available, produced by our colleagues at Nature Video.

The video captures the essence of the event, co-organised each year by Google, O'Reilly and Nature. Included in the piece are a number of our Sci Foo friends, from Marcus du Sautoy, and Paul Davies to Aleks Krotoski and Andrew Marr.

Sci Foo originated as an extension of an experiment by Tim O'Reilly and his colleagues in 2003. He brought together 200+ "Friends of O'Reilly" Media (where the "Foo" in Sci "Foo" comes from), let the participants create the agenda from start to finish, putting few if any boundaries to what can be discussed.

The results were astounding, challenging the conventional idea of a "conference". In 2006, O'Reilly joined up with Nature and Google to hold the first Science Foo Camp (“Sci Foo”), inviting over 200 leading scientists, technologists, writers and other thought-leaders to gather at the Googleplex for a weekend of unbridled discussion, demonstration and debate. We're now in our sixth year of the event.

For more information on Sci Foo, visit the event page over on nature.com. We'll be posting the rest of the series of videos, as well, so watch this space.

A short movie break...

Fri, 10 Dec 2010 15:43:00 +0000

To mark the launch of Digital Science, we’ve produced a short movie in collaboration with our colleagues at Nature Video. The idea came from a meeting with Andrew Carol, one of the many remarkable guests at this year's Science Foo Camp. It has nothing directly to do with our projects, but it chimes with our philosophy and aims. The video pays homage to the oldest know scientific computer, the Antikythera Mechanism, adding a modern spin with stop-motion animation and Lego building blocks (yes, Lego). It serves as a vivid celebration of human technological and scientific ingenuity through the ages. Beyond that, well ... it’s just wonderfully cool (in a geeky kind of way). We hope you enjoy it. And for those who like DVD extras, here’s some background footage showing how the movie was made. For more on the history of the Antikythera Mechanism, check out the latest from Jo Marchant in Nature. Many thanks to Andrew Carroll, John Pavlus of Small Mammal, and Adam Rutherford at Nature for putting this together. The video is available under a CC-BY-3.0 Unported license.

Introducing ourselves

Tue, 07 Dec 2010 15:38:00 +0000

We're thrilled to announce the launch of Digital Science -- a new business division of Macmillan Publishers and younger sibling of Nature Publishing Group. You can read more about our approach and product offerings in our about page or on our products page. In this post we wanted to take a closer look at a couple of other key questions: why are we creating this new business, and what is our ultimate goal?

Digital technologies are transforming all areas of our lives -- commerce, education, the arts -- and science, too, is changing. The web has already revolutionised the way we produce, publish and disseminate scientific knowledge. It also provides fundamental new opportunities for processing, annotating, curating, querying and sharing information, as well as organising our laboratories and the research process itself.

But we're still much nearer the beginning of this journey than the end. For all the recent progress, we're not even close to fully realising the potential of information technology to accelerate the discovery and and application of scientific insights. At Digital Science, we aim to to contribute to these important developments by providing tools and services that will make researchers more productive through state-of-the-art software.

In that sense we're very different to the content-based businesses normally associated with a publishing company like Macmillan. But we will also be working closely with scientific publishers -- not least our colleagues at Nature Publishing Group -- to make the most of their scientific expertise and carefully curated content, because only by combining expert human judgement with the best technology can we hope to have the impact that we seek.

To fulfil our mission, we're bringing together a team with a wide range of experiences, from software development to business development, and pharma to open science, to research. I encourage you to visit our team page to learn more about those already on board.

We'll be posting regular updates here as we work towards these goal. I hope you'll stay tuned.