LETTERS
https://doi.org/10.1038/s41587-019-0356-z
1
Functional Materials Laboratory, Department of Chemistry and Applied Biosciences, ETH Zurich, Zurich, Switzerland.
2
Complex Materials, Department
of Materials, ETH Zurich, Zurich, Switzerland.
3
Erlich Lab LLC, Raanana, Israel.
4
These authors contributed equally: Yaniv Erlich, Robert N. Grass.
*e-mail:
erlichya@gmail.com; robert.grass@chem.ethz.ch
DNA storage offers substantial information density
17
and
exceptional half-life
3
. We devised a ‘DNA-of-things’ (DoT)
storage architecture to produce materials with immutable
memory. In a DoT framework, DNA molecules record the data,
and these molecules are then encapsulated in nanometer
silica beads
8
, which are fused into various materials that are
used to print or cast objects in any shape. First, we applied
DoT to three-dimensionally print a Stanford Bunny
9
that con-
tained a 45 kB digital DNA blueprint for its synthesis. We syn-
thesized five generations of the bunny, each from the memory
of the previous generation without additional DNA synthesis
or degradation of information. To test the scalability of DoT,
we stored a 1.4 MB video in DNA in plexiglass spectacle lenses
and retrieved it by excising a tiny piece of the plexiglass and
sequencing the embedded DNA. DoT could be applied to store
electronic health records in medical implants, to hide data in
everyday objects (steganography) and to manufacture objects
containing their own blueprint. It may also facilitate the devel-
opment of self-replicating machines.
The worlds data are growing at exponential rates in today’s
information age
5
. However, attempts to further miniaturize tradi-
tional storage architectures, such as hard-drives and magnetic tapes,
are becoming increasingly difficult
10,11
. These devices are reaching
their physical limitations and can hardly keep pace with digital
storage requirements. Due to these challenges, there is great inter-
est in the potential of using DNA molecules as an architecture for
long-term cold storage. Previous studies have reported that DNA
storage can reach 215 PB g
1
, orders of magnitude higher physical
densities than traditional devices
1,5
, and can have a half-life of thou-
sands of years
3,12
.
Beyond exceptional density and endurance, DNA storage is vir-
tually the only storage architecture that can take any shape. This
stands in stark contrast to traditional storage architectures such as
tapes and hard-drives, where the actual shape of the device is often
critical to its functionality. Previous studies of DNA storage have
largely overlooked the virtue of the absence of shape constraints.
However, this property allows the realization of new storage archi
-
tectures beyond today’s conventional designs.
Here, we explored a storage architecture, dubbed DoT, in which
DNA molecules are fused to a functional material to create objects
with immutable memory. The DoT architecture starts with encod
-
ing the data in DNA molecules in a manner that is robust to errors.
Due to the low density of the DNA molecules in the embedding
material, we expect that some DNA fragments will not be recov-
ered from the object, creating fragment dropouts. To ameliorate
that, we used DNA Fountain
5
as our encoding scheme. This scheme
provides high flexibility in setting virtually any redundancy level
that can correct dropout errors and has been shown to perfectly
retrieve data from minute quantities of material
5
. One challenge for
the DoT architecture is that simply mixing the DNA with the func-
tional material results in quickly degraded DNA due to hydrolysis
stress and elevated temperatures during the preparation of the mix-
ture (Supplementary Fig. 1). To mitigate this, the DoT architecture
first encapsulates the DNA insilica nanoparticles
8
, resulting insilica
particle-encapsulated DNA (SPED). The SPED shell sequesters the
DNA molecules, prolonging their half-life and facilitating the mix-
ing of DNA with the embedding material. Finally, we mix the SPED
with the desired material and shape it using either three-dimen-
sional (3D) printing or casting techniques.
To empirically test data storage using the DoT architecture, we
created a 3D object that embeds DNA that encodes the blueprint
for creating itself (Fig.
1a and Supplementary Video 1). This con
-
figuration is reminiscent of biological organisms, in which the
instructions for making an object reside within the matter itself. As
for the object, we selected the Stanford Bunny
9
, which is a common
computer graphics 3D test model. First, we compressed the binary
stereolithography (stl) file of the bunny from 100 kB to 45 kB. Next,
we used DNA Fountain to encode the file in 12,000 DNA oligo
-
nucleotides (oligos), which is the maximal number of oligos pro-
duced by a single CustomArray chip (Supplementary Note 1). With
this number of oligos compared with the file size, DNA Fountain
encoding yields a redundancy level of 5.2×, meaning that we can
tolerate a dropout of even 80% of the DNA oligos and still correctly
decode the file. The length of the oligos was 145 nucleotides (nt),
consisting of 104 nt of payload and 41 nt for PCR annealing sites
(Supplementary Fig. 2). Next, we loaded the PCR-amplified oligos
into SPED beads (Supplementary Fig. 3 and Supplementary Note 2)
and embedded the beads in polycaprolactone (PCL). PCL is a bio-
degradable thermoplastic polyester that offers low melting tempera-
ture and high solubility in various organic solvents, making it an
ideal material for blending and printing under mild conditions. To
prepare 3D-printing filaments, we mixed the SPED capsules with
dissolved PCL and extruded the mix into a 2.85 mm filament com
-
patible with desktop 3D printers. Notably, the filament contained
SPED beads in a concentration of 100 mg kg
1
(100 ppm), which
did not create any detectable changes to the mechanical properties,
weight or color of the filament
13
. The DNA loading within SPED
was 2 mg of DNA per gram of SPED beads, which translates to a
DNA concentration of 0.2 mg kg
1
(0.2 ppm) of the PCL filament
(Fig.
1b and Supplementary Note 3), well below the concentration
A DNA-of-things storage architecture to create
materials with embedded memory
Julian Koch
1
, Silvan Gantenbein
2
, Kunal Masania
2
, Wendelin J. Stark
1
, Yaniv Erlich
3,4
* and
Robert N. Grass
1,4
*
NATURE BIOTECHNOLOGY | VOL 38 | JANUARY 2020 | 39–43 | www.nature.com/naturebiotechnology
39