Stephen's talk and the science of manifold data
Stephen Friend, the president, co-founder, and director of Sage Bionetworks, gave a Hematology/Oncology Grand Rounds talk at Duke on 13 February. He called it "Integrating Genomes and Networks to Understand Health and Disease: If Not Now, Then When?" Although to say so oversimplifies the real message, Stephen urges us to move beyond symptoms-based view of health and disease toward a more complex view derived from complex and comprehensive data. This shift to manifold data -- not just "big" but also "broad" data -- shapes scientific inquiry: we begin to migrate from hypothesis-driven science toward data-driven science, capturing a more complex picture of biology in the process.
"Doing this kind of science doesn't become any easier just because you acknowledge complexity," he cautioned.
Disease, and especially cancer, is not simple, and our descriptions and explanations oversimplify. Stephen has a perspective that's been informed by stints in academe, pharmaceuticals, and biotech. He's played an influential role in deepening biomedical data over the past decade. And yet, he said, as impressive as the insights these data reveal, they are but "narrow slices." Explaining complexity of disease with them, Stephen said (in a delightful simile), "is almost like you're trying to tell the story of a movie by looking at a single frame."
I agree. We are doing a new kind of science -- or, perhaps more accurately, we're thinking about a new way of explaining the Scientific Project: from a rhetoric of the neatly ordered processes of hypothesis-driven steps toward a way of thinking about that discovery as less ordered, more accommodating of influences and the products of interaction, even unnervingly accepting of serendipity. And, of course, that new explanation will accept the remit of manifold data and the difficulties and ambiguities that come with their complexity.
Be clear about one thing, though: this change of science is not just about method. It's not just a matter of doing analysis of different kinds of data or using more mind-bending statistical methods or talking with colleagues more frequently (in order to "cross pollinate" and feel all "interdisciplinary").
The change evokes a shift of authority grounding the whole enterprise. That's the challenge, especially for large and established academic research organizations.
The Latin word auctoritas denotes the kind of force that "augments" action. It guides in a manner "more than advice and less than command" (Theodore Mommsen). It ordains, in the sense of ordering and setting to rights. And, in Stephen's talk, the issues of auctoritas came forth as I had hoped they would.
Publication and auctoritas. Of course, true to etymology, publication authorship reflects the shifting of auctoritas. "Patients who give you their data will tell you to publish as a group," Stephen observed as he described a new and highly involved vision of patient engagement in research. No more jockeying for first or last author or languishing in the mediocre middle. Following the example of GitHub and open source software development, contributions to a biomedical research project would be associated with individuals in a much more detailed fashion: credit is given atomically and is based on activity, not bestowed after the fact by a principal investigator or senior author.
Stephen juxtaposed the title page from the first Philosophical Transactions of the Royal Society (1666) with GitHub. While the comparison is interesting, a focus on the medium of communication is, I think, really not at the crux of the matter. The Philosophical Transactions embody an entirely different set of authorities -- reflected in the last two words of the publication name, of course -- from those embodied in GitHub. And the contrast even has a bit of humor: GitHub is a hub or meeting place for "gits" (a nest of ninnies to use seventeenth century coinage). Using "gits" as a source of talent and progress is really the revolutionary thing here.
Data, the public, and auctoritas. At one time, medical researchers kept their careers, quite literally, in the freezer -- where biological samples for investigations were preserved. Now, careers are also kept on hard drives, where data derived from samples resides. The result in academic research culture is a set of "tenure feudal states" where fiefdoms secure and protect their resources, derive their stature from their use, and create alliances by leveraging those resources.
But what happens when emerging scientific questions overmatch the resources of any "feudal state"?
In part the appeal for more comprehensive data ends up breaking the borders of the feudal states. Stephen seeks to open up data much as he wants to change authorship. The fiefdom of the academic lab is replaced by an appeal to "citizens." Consent is a key, and on this front Sage Bionetworks has been quite ingenious with "Portable Legal Consent." (See weconsent.us for a new model of research consent developed by John Wilbanks, Sage's "chief commons officer" and his colleagues.) The involvements of the public in the frameworks of research that Stephen set forth involve not just surrendering samples to research overlords. The new framework calls upon "citizens" to play roles of "citizens as patients," "citizens as scientists," and "citizens as funders."
I like to think of this as a republic of data, though I think the structures still need some working out. (The classicist in me recalls Cicero's concern about auctoritas, and his observation that in the Roman Republic, "Cum potestas in populo auctoritas in senatu sit" -- "Power is in the people, while authority is in the Senate.") Stephen recognizes where the power resides (and he's a Ciceronean in that regard!); the mechanisms of auctoritas are less clear, and the emerging innovations like Portable Legal Consent, patient/citizen foundations and research organizations are exploring avenues (a few examples here, here, here, and here).
As the old chestnut goes, Rome wasn't built in a day. Stephen said that he saw the changes taking "years if not decades," so I do think he sees the enormity of the change and the kind of deep-rooted resistance that changes like this inevitably encounter.
Clearly, the kind of shift that Stephen laid out could be an existential threat to existing frameworks for academic science. Questions after the presentation brought up that threat, and Stephen's responses (I think) were more reassuring and sanguine than mine would have been. By resituating authority that "augments" and undergirds scientific endeavor, you change things radically. The face of academic research institutions will have to change if scientists in those institutions want to take part in "integrating genomes and networks to understand human health and disease." Stephen hopes that the depth of expertise that today's framework enables and nurtures can be complemented by this new direction. I hope so, too, but its authority will not be undiminished.
I wrote to Stephen after his talk, "I was pleased that you used the theme of institutional renewal as a parallel theme to the science. Of course, these are intertwined, though scientists like to talk about 'findings' and 'results' absent of the machinery used to uncover them." I hope we can proceed with a feeling of institutional renewal, for creating a new auctoritas for this new science is a good and necessary undertaking. It's been done before. It's actually nice that I post this on Galileo's birthday, but his circumstance also serves as a caution: He spent his later years under house arrest for the changes he was forced to recant.
