Preservation In Vivo
BY MADELEINE MENDELL
Like others researching the storage of digital media and data onto DNA, I have become fascinated by the implications this new storage format will have for our relationships to the living world, to ourselves, and to our data. As a digital preservationist and audiovisual archivist-in-training, my research has primarily focused on the use of silicon-based synthetic DNA as an off-line storage solution for cultural heritage institutions. At these archives, preservation consists of a series of practices that seek to ensure digital object integrity while providing access over time. Eventually, a new format must conform to these objectives if it is to become appropriate for archival storage, which encompasses the housing, maintenance, and retrieval of these objects [1]. In vitro, or synthetic, DNA has become an increasingly realistic option for institutions coping with massive amounts of data. A single gram of synthetic DNA could potentially store 215 petabytes of information, which would be the equivalent of almost 12,000 LTO-9 18-terabyte tapes, the current industry standard.
In researching synthetic DNA as a format that might soon be able to fulfill those capabilities (and which has recently gained interest from some archives), my focus has routinely become unmoored from the synthetic and re-oriented towards the living. Synthetic DNA has received large amounts of funding and press attention as the next big storage technology. Conversely, in vivo, or living, DNA data storage remains an experimental, fantastical, and improbable option. In vivo DNA storage relies on the usage of CRISPR gene-editing proteins to insert data into a living vector’s DNA strand and the cultivation of that vector – for example, an Escherichia coli bacteria colony – in order to amplify (make copies of) the information. Data stored on the cells of living bacteria necessarily must grow and mutate. Synthetic DNA, on the other hand, is isolated, static, and stable on its own. It does not change over time.
Melody Jue’s Wild Blue Media and Alexis Gumbs’ post-apocalyptic archaeological epic M Archive carve out frameworks, at once scientific and poetic, for exploring the archive in and after life. I am particularly drawn to Jue’s methodology for a “conceptual displacement as a method of defamiliarization to make our terrestrial orientations visible” [2]. Jue uses conceptual displacement to estrange herself from terrestrial ways of thinking; here, I wonder, alongside Jue’s oceanic database full of “eroded, washed away, or overgrown” inscriptions, how memories preserved in little cellular colonies might affect our perceptions of time and life. I am inspired by Jue’s and Gumbs’s own methodologies, and preservation is the system that has guided me in working through these questions of change and integrity. Positioning in vivo DNA within a digital preservation workflow not only asserts the aliveness - and therefore susceptibility to biological change - of the system but also has the capacity to re-orient preservation away from the practice of resisting death and towards embracing life.
Encoding data onto living cellular DNA has not been seen as an option for archives or large data centers. In the few conversations with archivists I have had, colleagues have expressed disinterest, ignorance, or dismissal of the format. I am unsurprised by their reactions. The lack of stability data might have in a bacteria colony due to the uncanny proximity to life do not make in vivo DNA a strong candidate for long-term archival storage. When a digital object enters an institution, archivists must find a way both to ensure that the object remains as close to its original state as possible and to give that object an afterlife in access. A living colony of information that must transform over time appears counterintuitive to these objectives, but also recalls a period during the 1990s in which forward-thinking librarians and archivists saw the future of stewarding digital objects as a radical change to the stability of the paper archive.
Librarian and scholar Clifford Lynch wrote in 1993 how the transformation of an “electronic work” necessary for it to be perceived – say, the rendering of a bitstream into a video on a screen – “can cause major changes in both the presentation and the intellectual content and thus threaten the integrity of the work” [3]. From the outset, preservationists have grappled with the mutability of digital objects in the archive: their ability to change and be changed, their instability, their simultaneous infidelity to any one state of being and obligation to various hardware and software dependencies. In the years since, archivists have addressed this through, for example, fixity checks, which monitor the integrity of an object over time. I bring up this history to underscore how the threat of change to an archival object’s afterlife causes a paradigm shift wherein an object does not simply move towards its logical, if prolonged, end. Instead, it can transform into something else. I offer up the potentiality of a living storage format as an alternative to the incessant labor of creating remedies for stymying change and prolonging only an object’s afterlife.
In 2017, a Harvard team’s research into encoding data onto E. coli bacteria colonies led to the first ‘movie’ stored on DNA. The ‘movie,’ really just several successive still photographs by Eadweard Muybridge, was chosen not to demonstrate the storage of data onto DNA but instead as “a good example of a complex piece of information that has both many parts to it – that is many pixel values – and a time component, so it’s organized over time” [4]. The “time component” not only lends the system of images more complexity as a bitstream but also demonstrates, in the sequencing of those nucleotides back into images, how the bacteria’s mutations have affected the images over time, creating little digital (or biological) artifacts in the files. As Deb Verhoeven discusses in her conversation with Mél Hogan about this project, “DNA is used by these scientists as a field for mining or harvesting categories of time and space” [5]. It is no coincidence to me that in these Muybridge stills the Harvard scientists chose to represent not an early movie, but an experiment in sequencing time and space by way of the technical image. In the cellular regime, time is the scale on which this data is embedded, grown, and reconceived of as data. Time can be seen in the changes between the ‘write’ and ‘read’ of the bitstream, which in turn reveals a metatextual index of cell division and DNA mutation in the bacteria.
The changes documented by the Muybridge photographs are less about the persistence of the object itself and more about the existence of this object in a new environment. They indicate to the sequencer what events happened to the object over the course of its time in the bacteria. In archiving, there is a language for these changes in preservation metadata – information about an object’s preservation – and a system for describing these changes in PREMIS, a data dictionary hosted by the Library of Congress. PREMIS helps archivists document changes in human-readable language through a model that includes the entities Agent, Event, Environment, and Object. These entities describe the preservation actions (Event) an institution, software, or person (Agent) performs on a piece of media (Object) as well as the various components required to render that object (Environment, loosely). In the Environment of an E. coli colony, DNA acts as an all-encompassing preservation metadata system, recording the events of its own change over time onto itself.
Preservation takes place over time. Preservation metadata documents these actions, not only revealing the ongoing afterlife of an object in care, but also the ideas and values of preserving institutions and technologies. This record of maintenance is what the in vivo cell produces: an indefinite log of an object’s residency in a new Environment. The difference between the living and synthetic DNA reveals a critical insecurity at the center of preservation’s relationship with time: that in order to make a (digital) object accessible, archivists must change the nature of that object, by migrating it, emulating it, virtualizing it, transforming it. Access and integrity necessarily must work together and against each other. At the center of that dialectic is the digital preservationist’s constant task of maintenance - the maintenance of a stable, unchanging object that in turn must be rendered consistently into the future. The fantasy of a storage format without need for constant care that will last thousands and thousands of years is the hook for those companies and institutions proffering synthetic DNA. A static container that won’t require any work, i.e. the work of archivists and preservationists, is the administrator’s dream.
On the other hand, the living DNA data storage holds a visual mirror up to the preservationist’s labor and confronts us with a question: what is it to maintain a record over time? Preservation necessarily hinges on a linear construct of time: that there is a future in which an object might be accessed, will need to be rendered, must be the same. A human record stored onto an E. coli colony displaces that object from our anthropomorphically centered experience of time. In the early weeks of this year, technological advances by Yim et al. opened the possibility of a direct digital-to-biological workflow for encoding DNA onto bacteria using an electrical stimulation to record “temporal biological events over time” [6]. Seth Shipman, of the original Harvard project, calls this “time imprinting” or “temporal stratification” [7]. Like any other time-based media that has a duration and unfolds over time, DNA is encoded over and by time.
How can DNA be time-based media? Is time how long it takes for the E. coli cell to divide and replicate? E. coli is often preferred for lab experiments because of its fast growth, but that cell division is not marked by clear intervals. E. coli cells perform “multifork replication,” or a process by which cells begin replicating anew before parental cells have finished writing out the new DNA. In order to quantify the cell replication time, scientists must impose artificial intervals of time onto a colony, but at what point can a cycle be complete if the bacteria continuously reproduces itself? What is time to an E. coli colony?
Scientific processes – similar to the moving image’s simulation of time through sequential stillness – must create synthetic intervals in order to quantify and research time. However, as Karen Barad argues – and maybe an E. coli colony might agree – “the past is never left behind, never finished once and for all, and the future is not what will come to be in an unfolding of the present moment; rather the past and the future are enfolded participants in matter’s iterative becoming. Becoming is not an unfolding in time, but the inexhaustible dynamism of the enfolding of mattering” [8]. Becoming and ongoing, the overlapping cellular replication of a bacteria colony on a petri dish emphasizes not only how time resists imposed sequential structures but also how time refuses stability and quantifiable stillness.
Imagine a world where preservationists, on the precipice of committing to synthetic DNA as the final resting place for our cultural heritage, instead choose the unruly, alive Environment of a bacteria colony (perhaps “radiation and heat-resistant spores” [9]). Preservation is a system unstuck from death by digital media, and can potentially be pushed into life by DNA. Preservation was also always alive, populated by human beings laboring to maintain the existence of objects in the afterlife. With unfolding, indefinite duration, DNA and preservation are time-based, dynamic, active, alive. An object stored onto in vivo DNA has an afterlife in life. Melody Jue writes of an oceanic database in which seawater is “both a storage medium and a transformative force that changes bodies” [10]. To choose a system that documents time through its material and spatial changes and inscribes those changes onto our species’ cultural objects would be to embrace the becoming of an object through its afterlife. It would be to see the generation of preservation metadata as a stream of information not only reflective of an institution’s values, but also subject to, not resisting, the conditions of the Earth, the Environment.
Consider the National Digital Stewardship Alliance (NDSA)’s definition for “archival storage” as an encompassing system that “provides the services and functions for the long-term storage, maintenance, and retrieval of digital objects” or the Open Archival Information System (OAIS) Reference Model’s, which defines archival storage as a component of an OAIS-type archive that “manages the long-term storage and maintenance of digital materials entrusted to the OAIS.”
Melody Jue, Wild Blue Media: Thinking Through Seawater (Durham: Duke University Press, 2020), 6.
Clifford Lynch, “Accessibility and Integrity of Networked Information Collections” (Washington, DC: US Office of Technology Assessment, August 1993), 61.
Seth Shipman, “Movie Replayed from Living Cells’ DNA Debuts ‘Molecular Recorder’” (National Institute of Mental Health), July 2017, YouTube Video, 7:40, https://www.youtube.com/watch?v=gK3dcjBaJyo.
Mél Hogan and Deb Verhoeven, “Sustainable DNA: In Conversation,” in Right Research: Modelling Sustainable Research Practices in the Anthropocene (Open Book Publishers, 2020).
Sung Sun Yim, Ross M. McBee, Alan M. Song, Yiming Huang, Ravi U. Sheth, and Harris H. Wang, “Robust direct digital-to-biological data storage in living cells,” Nature Chemical Biology (2021). https://doi.org/10.1038/s41589-020-00711-4.
Santi Bhattarai-Kline, Sierra K. Lear, and Seth L. Shipman, “One-steph data storage in cellular DNA,” Nature Chemical Biology (2021). https://doi.org/10.1038/s41589-021-00737-2.
Karen Barad, Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning (Durham: Duke University Press, 2007), 234.
Yim, et al., 3.
Jue, 113.
Madeleine Mendell is a second-year graduate student in New York University’s Moving Image Archiving and Preservation MA program. She is interested in the chemistry and materialism of media in the archive, from hand-processed film to the storage of digital media on DNA.
Banner photo by National Cancer Institute on Unsplash.
