Ripples After We Toss the Pebble

When it comes to academic research, digital transformation (Dx) has fast and slow tracks, and I've traveled on both. I've been the researcher—a passenger on the train—and I've been the conductor—the IT guy helping to keep the train on track. Digital transformations involving the management of institutional assets, student records, and new forms of instruction delivery, for example, have fairly clear purposes and processes. All of these areas have benefited from the application of digital tools, often without drama (or at least without tragedy).

Those of us "doing" Dx in academic research confront the slow and fast tracks. In the end, it's not just about flipping the switch on a digital tool or service. It's also about the ripples that emanate after the stone has been cast into the pond. Dx attends to non-digital matters as changes ripple across an institution. Those of us who work in information technology might not be charged with managing the implications that Dx foists on research and researchers, but we must be able to understand the implications and take part in responding to them.

Life in the Fast Lane

Dx is such old hat in some fields that researchers in those disciplines might think the term "digital transformation" is rather quaint. The reason for this is simple: research progress is hitched to digital tools and technologies for computational analysis and data production. These researchers have stepped beyond analog approaches, which still have valuable uses and have "gone digital." Examples are predominantly in the sciences, where floods of digital data have been transformative. For example, cosmologists are anticipating new, dense, and abundant digital imagery from the Large Synoptic Survey Telescope (LSST) project, with the data taps beginning to flow next fall and continuing afterward for ten years. LSST will take highly detailed and unparalleled pictures of the universe. New data and computational methods are also digitally transforming biological and biomedical research. The National Institutes of Health (NIH) underscored the volume of data and the computational muscle necessary to make use of it: "By 2025, the total amount of genomics data alone is expected to equal or exceed totals from the three other major producers of data: astronomy, YouTube, and Twitter."¹

My work as a researcher was in genomics, where I manned a keyboard rather than a pipette. It was and still is a "fast-track" field in transformation. Anyone who worked in genomics during the first decade of this century experienced the coupling of digital technologies, computing, and biological "bench science." In the early 2000s, our lab ran old-fashioned Sanger sequencing technologies that delivered what we then thought were loads of data. Sanger systems performed nothing like today's next-generation sequencing technologies. Digital research tools and genomics research were changing so quickly that the fact that research progress depended largely on the deft coordination of digital and lab-based activities was easy to see. Indeed, these activities converged, with interpretive and experimental processes running hand-in-hand.

Because of digital tools and methods, the science itself transformed. Through the first decade of the century, DNA became not only a form of description—a transcription of the "language of life"—but also a new type of signal that is now commonly used in studies in the life sciences. In essence, the gap narrowed between digital technologies and biological studies: computation—teasing meaning from the signal—moved nearer to the heart of scientific work.²

This is just one story of digital transformation in just one branch of the sciences. There are other stories in other disciplines, but here's the noteworthy feature: Dx has become inextricably tied to doing research. Transformations, digital or otherwise, usually result in some recognition of profound change, but in my lab, the change felt natural rather than contorting. We embraced digital tools because they so squarely fit the demands and challenges of the research.

That said, the changes that contorted lurked in other areas of scientific life that were tied to genomics. Probably the most apparent shifts were dislocations of what exactly it meant to conduct research in certain areas of biology. What data are essential to the field? What skills and competencies are required to work in the field? In simplest form, a faculty researcher might ask: "Should education in [your field here] include training in a programming language? Is 'data science' a part of the curriculum?" In effect, Dx in academic research concerns things other than digital appliances and computer systems. The transformation affects how scholars and scientists define their work, themselves, and their fields, and in ways that are more cultural than would normally be associated with the digital.

Not all scholars and scientists will have as easy a time with "transformation" as my genomics lab did, since we were engaged in a new field that didn't yet have many practices and protocols firmly in place. Of course, mismatched skills, lack (or ignorance) of tools, and just plain old fear hinder the adoption of digital approaches in research, but the toughest challenge is the alignment of the cultural values of a field and the cultural norms that are implied in digital scholarship. Dx makes assumptions about the culture of research: who does it, what tools they use to do it, and how the work is reported, evaluated, and maintained. In the sciences, Dx is often considered ho-hum largely because it already resonates within existing research cultures. Dx is part of the way that science works.

For instance, the sciences recognize the value of collaboration, acknowledging that individual talents cannot match the tools and experience that modern research often requires. And so, when a paper is submitted, co-authorship is common, maybe even the norm, and processes—or guidelines at least—for determining the order of the names are in place as well (although some frustration or yelling may be involved at times).³ The acceptance of collaborative research has allowed Dx to happen more easily in science fields because the talents of other areas like computer science became essential to address, say, biological questions.

Even in fast-track fields like genomics, however, gaps remain and to some extent have been created by the new opportunities that are introduced by technology. These new opportunities call for research and its products to be reimagined. The published paper no longer just reports "results"; it has also become an "advertisement" for the underlying data of a study—data that could be relevant to other projects. In part, this transformation has come about because digital forms of data are easily copied, shared, reused, and extended. The combination of data-driven science and new technological capabilities lend an urgency to rigorous data management and adherence to the FAIR Data Principles.⁴ The implication? With the embrace of digital approaches and tools, the scope of science has changed: Projects no longer "end" at the laboratory door. Indeed, projects don't—or might not—end at all if the powers of Dx are realized and practical steps are taken to enable and promote data availability, interoperability, and reuse.

Not "Can We?" but "Should We?"

I've had some experience in the slow track, too, where the mode of research is quite different but still exciting and fulfilling. I was trained in languages and history, specializing the medieval and Renaissance periods—perfect preparation for a career in information technology and science! Research practices and "outputs" (to use a term that is more science-related) were—and predominantly remain—products of individual efforts that are more likely conducted in a study than in a lab. Research "data" in the humanities have indeed taken new digital forms—that is, data have been "remediated," a term used (rather unfortunately, I think) to refer to the translation of texts, images, and other artifacts into digital forms. However, investigation and study using computational approaches have certainly been less intense and enthusiastic in the humanities than in the sciences, in part because collaboration has not yet found firm footing in the humanities, and, in some fields is discouraged. "The projects are nice, and they are appreciated," one colleague who works in the humanities told me about her digital work, "but my chair really stresses traditional single-author publication." In some well-established fields, single-author publication fits well into a tenure portfolio, and there really is such a thing as a "tenure book."

All of this is to say that the slow track of digital transformation might be slow, or sometimes stopped, in part because there isn't agreement about whether certain fields actually need to board the train.

Ted Underwood, a professor in the School of Information Sciences at the University of Illinois at Urbana-Champaign, uses lines from the movie Jurassic Park to open the chapter on "the risks of distant reading" in his book Distant Horizons: Digital Evidence and Literary Change. Of the first chapters in that book, he admits that, "like the scientists in Jurassic Park . . . [they] have been 'so preoccupied with whether or not they could' use numbers to learn something about literary history, that 'they didn't stop to think if they should.'"⁵ This quip sums up the reason for furrowed brows in some academic departments, most notably in the arts and the humanities. For researchers on the slow track of Dx, the push toward digital transformation in higher education more urgently presses the "if they should" question. Academic identities would certainly be transformed. But how would they be transformed? And once transformed, what would the benefits and the costs be? Should fields and researchers themselves be digitally transformed?

The general concern is the disruption of perspective and discourse—threads that unify disciplines across time. All disciplines have legacies that place today's researchers into an extended, generation-spanning dialogue, but Dx may interrupt those legacies. Scholars could lose beneficial discourse, tested and well-established methods, and, perhaps most alarming, coherence.

I suspect that such an interruption is most feared in fields where digital approaches and scholarly collaboration might undermine, or at least call into question, the value of connoisseurship—hard-won and often rare expertise arising from an individual's or a small group of scholars' long study. That expertise and judgment underlie the culture of single authorship and qualify the acceptance of collaborative research. Moreover, connoisseurship, even though it has been extremely valuable and useful to our understanding of the world, can restrain data access by warping scholarly incentives. In some areas of study, digital approaches challenge connoisseurship by democratizing methods that may rely on expert human judgments. Additionally, collaboration may "fuzzify" the traditional scholarly activity of a field of study by admitting "outside" talents into the mix and pressuring the release of data or forcing scholars to divert their attention to digital problems, such as applying computational or numerical methods to, say, artworks or literature.

In some circles, these are scary notions, since they would confound the evaluation of scholarship, upend habits of scholarly communication, and perhaps usurp authorities that preserve coherence in fields. In matters of digital transformation like these, the argument depends on cultural matters, not technical ones. In such cases, is there a pathway for Dx that preserves continuities and deepens fields that already have depth and history?

Two (Not So) Easy Ways to Promote Dx in Academic Research

There are a couple of ways to promote Dx in academic research. While these methods do not require unusual or extraordinary technical abilities (nor are they particularly futuristic), they can be challenging in their own right. Practices that are implied would cut across areas of academic research and constructively support the digital transformation of a broad range of fields, perhaps not without worry but with a forward-thinking and inclusive focus.

1. Assign published, well-curated data the same worth as an article published in a competitive journal. In the academic world, data are often more valuable than the initial results, which is probably one of the reasons that many publications require data publication. Further study, including reproducibility studies, can arise from shared and valuable data. Faculty evaluations should assign a value to well-curated data and give added weight to data that are reused by others. Such a policy would signal the institutional support for responsible data management and bolster the institution's reputation for solid science and scholarship. Fields that are not typically interested in "open data" should be encouraged to curate and share data more openly.
2. Build infrastructure and support systems for interinstitutional collaborations. Such collaborations are thriving in many fields and are often well supported by campus information technology. However, some fields—especially those that undervalue collaborative research—need encouragement, not because of technology but because of an ingrained (and, I think, outmoded) scholarly culture. In part, change comes from a concerted effort on the part of campus information technology as well as a frank institutional recognition of collaboration as a useful strategy promoting innovative research.

At most institutions, these are big changes—cataclysmic in some fields. That's why, under normal circumstances, digital transformers must move carefully and attentively.

Circumstantial Acceleration

The coronavirus pandemic has undoubtedly accelerated Dx across whole sectors of the world economy, including higher education. Changes that were hurriedly made to continue the teaching missions at colleges and universities have, even in these few months, matured to the point that some digital accommodations may be incorporated into "normal" teaching and learning practices. The same is true of higher education research missions, though the fields and projects that are rooted in digital practices have certainly had an easier time of it.

At Duke, when the circumstances made then-normal lab work riskier, researchers turned to computational tasks—at least in cases where digital workflows had been woven into research processes. Although everyone regretted the loss of lab camaraderie, the research continued because it could be handled remotely and the digital resources could be easily managed from afar. Research activities that could not turn to such methods—those that had not been digitally transformed—were (and largely remain) stuck in place.

Higher education institutions make much of "business continuity" scenarios where digital resources fail or become momentarily unavailable, but the recent experience with the coronavirus pandemic shows that Dx can also preserve and continue the important business of higher education and academic research.

Dx transforms the ways that academe responds to its core missions so that scholars and scientists and students can be more resilient and vibrant in our digital age.

Acknowledgments

I want to thank my colleagues Misha Angrist, of Duke's Initiative in Science and Society, and Holly Dressman and Shelley Rusincovitch, of the Duke School of Medicine, for their comments and suggestions as I worked on this article.

Notes

1. "NIH Strategic Plan for Data Science," National Institutes of Health (website), August 7, 2019. ↩
2. A delightful illustration of the digital qualities of DNA comes from Nick Goldman et al., who in 2013 described a method of storing digital information in a DNA sequence. About a cup of DNA can store around 100 million hours of high-definition video. (See Nick Goldman, et al. "Towards Practical, High-Capacity, Low-Maintenance Information Storage in Synthesized DNA." Nature 494, no. 7435 (February 2013): 77–80. Theirs was a flashy demonstration showing an interplay of DNA and digital bits, but the routine laboratory processes in use today use DNA as a medium to record and detect immune responses (for example) and other biological processes. DNA is a signal for capturing and communicating states. ↩
3. The so-called "Vancouver Rules" (or "Uniform Requirements for Manuscripts Submitted to Biomedical Journals (URMs)" influence the interpretation of co-authors' contributions. They date back to 1978. See "Roles and Responsibilities of Authors, Contributors, Reviewers, Editors, Publishers, and Owners," International Committee of Medical Journal Editors (website), accessed March 29, 2020; and Davide Castelvecchi, "Physics Paper Sets Record with More than 5,000 Authors," Nature, May 15, 2015. ↩
4. Mark D. Wilkinson, et al. "The FAIR Guiding Principles for Scientific Data Management and Stewardship." Scientific Data 3, no. 1 (March 15, 2016): 1–9. ↩
5. Ted Underwood, Distant Horizons: Digital Evidence and Literary Change (Chicago: The University of Chicago Press, 2019), p. 143. Underwood reassures his readers, saying that "historical research is less risky than cloning dinosaurs. But applying numbers to the literary past, in particular, remains controversial enough that an analogy to Jurassic Park is not absurd." ↩

---

Mark DeLong is the Director of Duke Research Computing at Duke University.