Innovators 2011 part one – Jason Swedlow
10 May 2011
In a series of three articles, BBSRC Innovator of the Year 2011 winners reveal the secrets behind their innovations.
In this, the first, Professor Jason Swedlow explains how the Open Microscopy Environment uses open source imaging software to drive innovation and research across the life sciences. In the second, Professor Chris Lowe describes the commercial potential of 'smart holograms' and his history in spin-out companies. In the third, Dr Keith Waldron details how his research and collaboration with industry partners has led to the development of a novel peat replacement from food chain wastes produced by a brand-new composting process.
In 2009 BBSRC established the annual Innovator of the Year competition to celebrate scientists who delivered science with high economic and social impact. Now, as in the past, innovation lies at the heart of the technological treadmill that can solve both local and global problems, drive economic growth, and make our lives longer, easier and happier (see 'The money of all invention').
Professor Jason Swedlow – Innovator of the Year 2011
How does it feel to win?
Overwhelming and incredibly surprising. The interview panel was quite something. They asked great questions about how to make the project and model work and how to go forward. They were asking the same questions that we're struggling with. We're daunted in a sense, and maybe humbled by the scale of the recognition. We have a lot to live up to.
Jason Swedlow, Professor of Quantitative Cell Biology at the University of Dundee, collects his award and £10,000 as Innovator of the Year 2011. Image: Andrew Davis
Describe your innovation in your own words...
We build data specification and software to enable image data access. And access as broadly as possible – from anywhere by anyone, assuming appropriate credentials – using analysis and visualisation tools.
What we've been doing for many years is building specifications and releasing software using open source licenses, so users can access, modify and share the code.
What was your Eureka moment?
I'm not sure if there was a Eureka moment but we were inspired by the experiments we wanted to do, studying the process of cell division and dynamics of the mitotic spindle. We wanted to use time-lapse multi-dimensional imaging as a basis for quantitative measurements mitotic spindle dynamics. We had microscopes and cell lines but what we didn't have were tools to view and manage the scale of the data we would produce, or the means to share it with collaborators.
How were the first steps to getting it off the ground?
We just started! We sat around and talked about what the data types were that we needed to deal with. Then we identified a great postdoc, Ilya Goldberg, who wrote the first version of the OME Software that showed the power of using informatics tools for processing large amounts of image data.
We didn't have any funding. We didn't have anyone saying it was a good idea. It just seemed like something worth doing. Once we had demo version we wrote a white paper (ref 1) and talked about what we were trying to do - that was published in Science.
Swedlow and colleagues work on cell imaging inspired the Open Microscopy Environment (OME). Image: Jason Swedlow
We also went to see microscope companies with the idea and many said: "oh that's very nice". But when we showed them it working and many of them said: "ah, now that is interesting". Ilya Goldberg gets a lot of credit for that first version, and my colleague Peter Sorger.
The next major step came with Kevin Elicieri from LOCI [Laboratory for Optical and Computational Instrumentation] at University of Wisconsin-Madison suggested building Bio-Formats and linking that to the OME Data Model. We initiated that project together, again with no funding, and have taken it forward so that now it is incredibly successful and used world-wide.
How long did it take to get Bio-Formats up and running?
The first version was available within a year. Then we were giving posters at meetings and starting to form collaborations. That's how the collaboration with LOCI happened.
How do you stay ahead of the hundreds of file types in use when everything is moving so quickly?
Swedlow and Open Microscopy Environment team. Back row (l to r): Aleksandra Tarkowska, Scott Loynton, Colin Blackburn, Jean-Marie Burel, Simon Wells, Will Moore. Front row (l to r): Brian Loranger, Chris Allan, Jason Swedlow, Andrew Patterson. Image: Jason Swedlow
We get data from the community and we reverse engineer the data it and immediately release software that reads that data.
For example, different microscopes and their software versions use different file types – the microscope companies never call us and say they're releasing a new version – all we know is that suddenly there's a file type that doesn't work on existing computer programs out there. So the scientist sends us the data that needs to be supported, we examine the files, determine what's changed and update and release our software. So it's completely community driven; we just get real data from the community and people say, "great, thanks very much!"
What else do you make for end users?
OMERO is a data management engine, an operating system for scientific images that enables access, remote access and visualisation – all of that stuff. The two flagship tools that we release are Bio-Formats and OMERO.
What kind of feedback do you get from end users?
We get constant feedback because its open source and people always want it to be faster and improved. If you go to openmicroscopy.org and see the community and forums, it's constantly "What about this? What about that?" – that's the way open source works.
Is the open source model the future of research?
I'm not sure if I'd say that, but science is moving so quickly it's certainly a very powerful model for software that supports scientific research.
Can you use your software for all data or just images?
Today, our focus is on multidimensional image data. We try as much as we can to support information attached to image data, so analytic results, protocols, maybe a paper or pdf. We certainly seek to broaden data types we work with.
More recently, in partnership with a number of users and non-profit and commercial entities, we've been building customised versions for the scientific community, such as the JCB Data Viewer, to enable publishing complex multi-dimensional data. On the commercial side, Columbus®, released by PerkinElmer is essentially OMERO in a box. PerkinElmer then distributes our software to major pharma and biotech companies integrated with their own products.
On the commercial side, if it's open source how do you make money?
We build and release the software so it's open and people can see it. So they can assess if our tools are of use to them and if they're interested they come to us. If they want a commercial license then they can incorporate our software into their commercial products, a so-called OEM [original equipment manufacturer] arrangement. Then there are people who need the technology and want a custom solution and we build one up.
The basis of IP [intellectual property] and commercial success is the technology and know-how. We control the code base and we are making almost all the additions to it and that control helps a little because it's really complex. So the source code is ours, copyright by the University of Dundee and Glencoe Software and Glencoe Software provides commercial licenses and customization.
Swedlow's innovation allows scientists to share image data more easily, such as this fluorescent light micrograph of cell division. The mitotic spindle can be seen in red, chromosomes in blue and the larger bright yellow/green dots are the centrosome around which the spindle is formed. Image: Jason Swedlow
We do have some patches and suggested code that come in, especially for Bio-Formats, but almost always it's people saying "I want you to do this" or "I want you to change that ". We don't provide the total solution - we provide the infrastructure and users can build their tools on top of it. But we don't own people's data, they own their data.
What were the hardest parts along the way?
A few things. One, and this is true across science, is keeping the projects funded and as important and relevant to the community as possible.
Two, learning how to manage a software project is very, very different from running a research laboratory; it's learning how software is built and how it is conceived, developed, released, supported and maintained.
If I had something important to say, it's that modern bioscience depends a lot on great software tools and the skills of managing software are not trivial and not part of the repertoire of most biologists that my colleagues or I are usually exposed too.
Finally, creating this commercial model. We are using very well established templates from other open source software projects, but it has been a huge amount of work.
But you must be very satisfied now?
The dream of many scientists is that the work they put in is of some utility and relevance. It's wonderful to be discovering new things and developing new tools but it is especially gratifying to see them used, particularly by other researchers across a wide range of fields – it's inspiring for everyone on the team.
Where does the BBSRC funding fit in?
BBSRC has funded critical components of the project where we've been looking to develop a specific piece of important technology. We're very grateful to BBSRC and their panels.
We have two grants running currently, one funds our efforts to take Bio-Formats and OMERO and move toward the electron microscopy community and it is terrific to receive that funding and establish collaborations with the European Bioinformatics Institute.
The other grant is a collaboration with the National e-Science Centre NESC. The concept is simple: I have a sizable image set, somebody else has a sizable computer resource, how do we link those up?
More image data for sure, and we want to move the applications into a broader repertoire of data types. We are very interested in clinical data. Overall we have great ambitions – if it's scientific data we're interested in it – obviously that's overly ambitious but why not shoot high?
Two cells formed from mitosis; the two daughter cells' chromosomes can be seen in red. Image data may just be the beginning for the open source approach. Image: Jason Swedlow
The other thing we want to do from a technical standpoint is bring access to large distributed computation to a biologist's desktop, initially for their microscopy data. When we do imaging experiments, and we can get 50 images in an afternoon's work, that data has to be processed, annotated and understood and that involves computation. How is it that those 20, 50 or 1000s of images are going to be processed and reduced to quantities that can be measured and understood?
And there are large computer resources [clusters] around the country. If I have a large amount of data, where is the most efficient place to have it processed? Maybe my local cluster, or my collaborator has a resource somewhere else. Moving data is expensive, but sometimes worth the time and cost penalty because the collaborator has better algorithms that can be used, for example. It's about using resources efficiently and stimulating collaboration.
Does the Open Microscopy Environment affect publishing?
We want biologists to be able to publish data in any way they see fit. My lab can put up a web page containing results for discussion, and we want that data to be available to the community, so will be releasing software to say "this data is available from a URL".
We take a pretty broad view of what publication might mean; certainly a lot of publication must be through peer review, but the lesson of the web is that peer reviewed science is not the only form of scientific publication. It's very important and the core of what we do, but not the only one and we want to enable as many different forms as possible.
The web has changed the face of peer-reviewed publication. But I think the technology of publication needs to transform as well. If you look at a published paper as it appears on the web it's still a pretty traditional format and we'd like to help change that.
These are not new concepts – this is what the sequence bioinformatics is all about and look at the power, the knowledge and the science that's generated. We can talk for hours about the challenges but that's where we want to go.
The UK has perhaps the richest history of scientific innovation of any country in the world, and the Royal Society report The Scientific Century: securing our future prosperity shows that innovation and commercialisation are flourishing in Britain.
For example, from 2006-10 university spinout companies have floated on the stock market or been taken over for a combined total of £3.5Bn and employ 14,000 people in the UK. Furthermore, between 2000 and 2008, patents granted to UK universities increased by 136% and university spin outs had a turnover of £1.1Bn in 2007/08 (ref 2).
Science can be a big moneyspinner.
The perception that the UK is not successful when it comes to commercialising science, or as some have put it: "Britain invents; the world profits" is therefore clearly outdated, and that strategies to harness and increase innovation are working.
In addition to the benefits it brings, it is argued that present £7.5Bn science budget pays for itself many times over as technology is developed and then taxed as it is sold. The Medical Research Council estimates every pound it spends brings a 39p return each year (ref 3). Moreover, independent studies have shown that for maximum market sector productivity and impact, government innovation policy should focus on direct spending on research councils (ref 4).
Finally, the UK produces more publications and citations for the money it spends on research than any other G8 nation. Specifically, the UK produces 7.9% of the world's publications, receives 11.8% of citations, and 14.4% of citations with the highest impact, even though the UK consists of only 1% of the world's population (ref 2).
- Informatics and quantitative analysis in biological imaging (external link)
- The Scientific Century: securing our future prosperity (external link)
- Medical Research: What's it worth? (PDF, external link)
- Public support for innovation, intangible investment and productivity growth in the UK market sector (PDF, external link)
tel: 01793 413329