It’s Science, But Not as We Knew It
‘The simplest way to describe what I’m doing is building in silico simulations’. Already I’m wondering if I’m out of my depth. I’m talking to Sophie Kay (née Kershaw), a new generation scientist in her twenties completing her doctoral thesis in computational biology at Oxford University. She specializes in cancer modeling. I’ve come to see her to try to make sense of the idea that science may have entered a new phase, one that may turn out to be as significant as the original Scientific Revolution in the Sixteenth Century. She’s enthusiastically explaining her research which involves building a virtual simulation environment for colorectal tissue, and I’m not sure I can keep up. Fortunately Sophie is used to explaining her work to non-scientists.
Here’s what I understand. Sophie’s approach is based on a mathematical understanding of cells and the forces that act on and within them – her first degree is in mathematics, she originally wanted to be a medical doctor, but switched to computational biology as a graduate – and loves it. It’s her passion. Rather than running experiments in a laboratory, she’s much more likely to be found programming or exploring complex three-dimensional lattice-like images represented on a computer screen. Her representations can simulate the buckling and folding of tissue within a colorectal tissue, and mimic activities of the cells with a high degree of precision, including cell division. Through a sophisticated virtual recreation of human body parts, her research should help anticipate the effects of drug interventions, research that can feed back to more conventional researchers – biologists, clinicians, experimentalists - with whom she collaborates in a two-way interdisciplinary process that both provides the raw data for her programming, and suggests the most fruitful directions for further empirical research. This mixture of modeling and experimentation is intended to accelerate the pace of discovery and avoid some of the costly digressions in to cul-de-sacs that have been typical of conventional biomedical research.
Sophie is typical of a new wave of scientists. On her university webpage she’s wearing a Google t-shirt rather than a white coat, and she shares the contagious, almost limitless, enthusiasm about new technology and its potential that is a given in Palo Alto. Although not a ‘digital native’, someone who has never known a time before the Internet, she is young enough to have become a scientist in a decade when computer-processing power was increasing dramatically and online data sharing was becoming commonplace. Like many in her Generation Y cohort, she is completely comfortable with digital technology and uses it to build on the work of more conventional scientists. Her research wouldn’t have been imaginable twenty years ago. The exponential improvement in computer processing has opened up the possibility of detailed simulations, but it has also opened up global communication of results.
‘In silico’ , it turns out, is just the new jargon for the computer simulation of biological proceses, a term that parallels ‘in vitro’ (experiments on tissue taken from an organism) and ‘in vivo’ (experiments on tissue within an organism).
It is not just individual scientists like Sophie who are developing projects in this way. To take just one large-scale example, the Virtual Physiological Human, a simulation of an entire body, pooling current biomedical discoveries, is an important collaborative project that may transform our understanding of disease and the body’s reaction to different treatments. Specialists in various body parts and tissues contribute their segments of this human jigsaw. Scientists will be able to access all the relevant data and simulations from the Internet.
As a philosopher I’m used to people talking of scientific revolutions in relation to Thomas Kuhn’s notion of paradigm shifts. This is the idea spelt out in The Structure of Scientific Revolutions that from time to time dramatic changes in understanding occur, changes that are almost incomprehensible from within the traditional way of doing things. Most of the time scientists engage in ‘normal’ science, following through experiments from within an accepted paradigm, but when new paradigms emerge a dramatic shift occurs. The classic case is the so-called Copernican revolution - the flipover from the dominant view that the Earth was the centre of the universe, known as the Ptolemaic view, to the heliocentric model proposed by Copernicus, Gallileo and others. Similar paradigm shifts occurred with Newtonian physics and then the Einsteinian theory of special relativity. But the digital scientific revolution is nothing like this. It’s not a new model for understanding the world and our place within it, nor does it involve a dramatic change in the basic empirical assumptions of science. It is rather that computers have accelerated and amplified traditional scientific research to such a degree, that they have transformed research itself, its communication and its effects. The consequences are rippling throughout the scientific community and well beyond, into the social sciences, and even the humanities. Today any young scientist seeking funding will at some stage be required to provide a Data Management Plan – a DMP Prosaic as this sounds, it is a symptom of the transformation. A DMP typically includes information about how data will be collected, archived, and retrieved, the kind of meta-data applied, and, importantly, an indication of who will have responsibility for maintaining the data. A sound DMP guarantees that scientific research can be retrieved and re-used as part of a global and collaborative attempt to understand the world.
Timo Hannay is exceptionally well-placed to appreciate what’s happening in science today. A former biochemist and neurophysiologist who went on to become director of web publishing for the prestigious journal Nature, he now runs Digital Science , a tech company that designs software and apps that change the way science is done and communicated. He also finds time to travel to Googleplex every year for SciFoo, an ‘unconference’ co-organized by Google, the tech publishers O’Reilly, and Nature, where many of the world’s top scientists, computer experts, and other thinkers rub shoulders with the likes of Larry Page and Sergy Brin, and share their visions for tomorrow.
Timo agrees that we really are at the beginning of a revolution, one potentially as significant as the rise of science in the 16th Century. He diagnoses two aspects of this. First, scientists today are generating significantly more data than ever before, much of this a direct result of digital automation and processing. This increase in data is quite staggering. Secondly, we now have software capable of making sense of this data, of organizing, sharing, mining, indexing, recovering, and analyzing it. Without the software, scientists would be drowning in a sea of data; with it, they can organize it, find patterns, filter out irrelevancies, draw new. Businesses such as Figshare are building online tools and cloud storage for organizing, saving, and sharing complex data, videos, images, charts, and other products of research in retrievable analysable ways. As a result of these changes new specialists are emerging. Some, like Sophie Kershaw, build models based on available empirical research rather than collecting original data themselves; others find new correlations and patterns within data collected for completely different purposes. Such changes are also apparent in the social sciences: Steven Pinker’s recent counterintuitive claim that that violence is in decline, for instance, is founded entirely on a very wide range of other people’s research which he was able to access and sift using digital technology.
Science is essentially collaborative and cumulative. Scientists can’t afford to recapitulate the entire history of their discipline each time they address a new problem: they build on others’ work or overturn it. Consequently reliable scientific communication has always had a special part to play in the enterprise. In the 1640s an ‘invisible college’ of so-called natural philosophers, began to correspond and meet to discuss scientific topics. This lead to Robert Boyle, Christopher Wren and others forming the Royal Society in London. By the1660s the Society was publishing a journal Philosophical Transactions to share information: its subtitle indicates its purpose and scope, that of ‘giving some accompt of the present undertaking, studies, and labours, of the ingenious in many parts of the world.’ Since then the refereed journal article has been the principal unit of communication - and prestige - for scientists. The nature of this publishing didn’t change significantly for over three hundred years. But now, as we know, things are different. Open access is the buzz term. Grant bodies such as the Wellcome Trust are requiring scientists to make their results available through online open access journals as a condition of the awards. Learned societies are having to adapt or die. But it’s not just that traditional articles are getting published in new ways. Raw data can be shared online, as can negative results, which have, for the most part been excluded from journals. Journals thrive on positive findings. But discovering that something doesn’t work or hasn’t been replicated can be significant too.
Of course, the problem remains that some scientists are proprietorial over their findings and methods, jealously guarding them until published in a journal; others are funded by organisations such as pharmaceutical companies which are highly motivated to conceal their research until they are ready to release a new product on to the market. Although to some against the spirit of global co-operation that has allowed scientific breakthrough in so many fields, such caution about sharing is understandable, and can be commercially prudent, if not always defensible on broader moral grounds.
Within science, though, there is a strong and growing movement that champions digital openness with an evangelical fervour. Many insist on using open source software, and publish their data freely and frequently, including negative results, and encourage others to build on what they have done. In the area of human genome sequencing, this collaborative pooling of discovery has been remarkable. Following a 1996 summit of top scientists who met in Bermuda, the so-called Bermuda Principles were drawn up. These included the agreement to release all DNA sequencing data to publicly available data-bases within just 24 hours of discovery. The motivation behind this data-sharing was that it would benefit humanity. The real hope for humanity, particularly in the area of medical research, though, lies with the next generation of researchers.
Young scientists who spent their teenage years using Facebook, Twitter, Flickr, SMS text, Skype, and blogging, are adept users online communication. Social networking takes place around the edge of their science, and long-distance collaborations are simplified by Skype and email. Some choose to share work in progress on webblogs, despite the sneers of older colleagues. All these factors break down barriers of distance that have in the past impeded scientific interaction. A generation that grew up playing computer games has even discovered that some science can be turned into a game, a game that has already produced results. FoldIt, is a semi-addictive online multiplayer contest which sets compelling puzzles that attract some of the brightest game-obsessed individuals who team up to solve problems which actually contribute to the unraveling DNA structure of particular proteins – a first stage in understanding how drugs can be used to treat illnesses related to the proteins, such as HIV, Alzheimer’s and cancer. After playing some straightforward games that teach the basic principles, FoldIt citizen scientists/gamers are unleashed on real life problems that they help to solve by predicting the likely structures of protein molecules, in a competition against other players to get the highest score. It feels like a game because it is a game, but the side effect is genuine progress in science. This ingenious ploy which harnesses the human predilection for pattern-recognition and puzzle-solving has already yielded spectacular successes, including helping to work out the structure of an AIDS-causing monkey virus: it took gamers 10 days to come up with a solution that had eluded mainstream researchers for 15 years. In The Adventures of Tom Sawyer, Tom manages to get his friends to do his work of whitewashing a fence by turning it into a game. Perhaps he was on to something.