SEATTLE – Very soon, it will be economically feasible to sequence human genomes and collect massive amounts of different types of health data as standard medical practice. Already, there are remarkable examples of how these new genetic data are changing our thinking about disease and diagnosis.
Consider the Beery twins, born in 1996 in San Diego, California. They suffered from chronic vomiting, seizures, and muscle weakness, sending them and their parents on an odyssey of medical examinations and tests. The first diagnosis was cerebral palsy. Then they received a diagnosis of dystonia, a rare neurological disorder. But treatments based on these diagnoses did not alleviate the children’s symptoms.
Frustrated, their parents had the twins’ genomes sequenced. The results revealed that the twins had been incompletely diagnosed. Their previously diagnosed dystonia was being caused by a genetic mutation that was interfering with the neurotransmitter serotonin. The twins’ doctors found that the dystonia could be fully treated with a readily available serotonin replacement.
So why haven’t success stories like that of the Beery twins, together with the Internet’s power and increasingly affordable collection of molecular data, led to the construction of a knowledge network of disease? Why aren’t scientists and doctors turning in droves to data-intensive science in order to build better “disease maps”?
One possible answer is that there are still technical barriers that block the construction and use of such networks. With our ability to generate ever-rising oceans of molecular data – now approaching the zetabyte scale (that’s a one followed by 21 zeros) – comes the challenge of storing and deciphering this information. But scientists and software engineers regularly face such daunting challenges, and, with DNA serving as the reference language of modern biomedical research, the technical barriers to constructing disease networks will be short-lived.
Cultural barriers are the real stumbling block. As humans, we are highly evolved to adjust to our surroundings: we tend to adapt to a culture, well-conceived or not, and lose sight of its failings. But when we glimpse an alternative, our culture’s inadequacies (and even insanities) are immediately apparent, which may prompt a cultural shift, collective action, and change. The fall of the Berlin Wall in November 1989 and, more recently, the Arab Spring are clear examples of this dynamic.
Similarly, the example of the Beery twins shows us that an alternative to symptom-based medicine can be realized: the advent of genomics technology can change not just what is known, but, more importantly, how we think of ourselves.
But, in order to build the disease networks of tomorrow, we will need to move beyond the current linear approaches to science and to how scientists work. We all like a good story that unfolds in a straightforward way, but the story of disease plays out across a poly-nodal information network, similar to what an air traffic controller might track in the skies above a major airport. Biomedical researchers’ “lock and key” and linear-pathway representations are incomplete, and should be supplemented with disease maps that can now be built using molecular data.
We must also build the infrastructure and cultivate the relationships needed to share disease maps with basic researchers, practicing doctors, drug developers, and even the public at large. And that could prove to be even more difficult, because the current closed nature of the medical-information system and its self-directed incentive structure block such sharing. Patents, trademarks, and competition for resources (people, money, and accolades) seal off information and prevent molecular data from being analyzed and shared. Rewards in biomedical research go to “solo workers,” and do nothing to acknowledge the work that can be done only by multi-functional groups.
Despite these imposing obstacles to progress, there are reasons to believe that a cultural shift is afoot: researchers from geographically distant labs are forming non-traditional “federations” to combine their data sets, work on them collaboratively, and post the results for other scientists to analyze. Crowd-sourced competitions like DREAM Challenges and FoldIt show that important scientific findings can emerge from outside of universities and pharmaceutical companies. And public-private partnerships between drug developers, basic researchers, and patient groups that share information pre-competitively (that is, with no or limited patent filings) are an increasingly popular way to translate scientific findings into potentially meaningful clinical benefits.
But a few successful federations, competitions, and partnerships may not be enough to transform biomedical research. Disruptive change may be required, and here each of us can make a profound difference. Patient groups are already organized; their members can report their symptoms online and self-enroll in clinical trials. We can already obtain portions of our own genetic information and use it to make informed medical decisions, join existing patient groups, or create new ones. We can provide our genetic samples to data-driven trials to learn about our likelihood to respond to particular therapies. We can even organize to self-fund future studies or join only those studies that give us the legal right to say how and where our data are used.
In other words, patients can and should stop being the passive “sick” and actively engage to pressure clinicians, researchers, and drug developers to adapt or perish. Democratized medicine represents the fullest flowering of the biomedical information revolution. There are few worthier goals than a future in which citizen-patients are active participants in managing their own health.