REVIEW

Hall of Fame www.advmat.de

Data Storage Using DNA

Shaopeng Wang, Xiuhai Mao, Fei Wang, Xiaolei Zuo,* and Chunhai Fan*

mainstream storage devices is struggling

The exponential growth of global data has outpaced the storage capacities of to keep up.[2] Consequently, the amount of
current technologies, necessitating innovative storage strategies. DNA, as a data being discarded due to inadequate stor-
natural medium for preserving genetic information, has emerged as a highly age capacity is skyrocketing. In terms of
promising candidate for next-generation storage medium. Storing data in long-term data storage, the durability and
maintenance costs of storage mediums are
DNA offers several advantages, including ultrahigh physical density and
also one of the primary considerations.
exceptional durability. Facilitated by significant advancements in various Driven by the ever-changing demands for
technologies, such as DNA synthesis, DNA sequencing, and DNA data storage, several generations of storage
nanotechnology, remarkable progress has been made in the field of DNA data media, ranging from magnetic tape to 3D
storage over the past decade. However, several challenges still need to be NAND flash memory, have been developed
so far with the aim of enhancing the per-
addressed to realize practical applications of DNA data storage. In this review,
formance, durability, reliability, and capa-
the processes and strategies of in vitro DNA data storage are first introduced, bility while keeping costs under control.[3]
highlighting recent advancements. Next, a brief overview of in vivo DNA data However, the rapid iteration of storage tech-
storage is provided, with a focus on the various writing strategies developed nologies also introduces another issue to
to date. At last, the challenges encountered in each step of DNA data storage the field: rapid obsolescence. Data stored on
are summarized and promising techniques are discussed that hold great older mediums may become difficult to ac-
cess as necessary software or hardware to
promise in overcoming these obstacles.
read the data may no longer be available.[4]
Therefore, there is an urgent need to de-
velop novel storage media with the poten-
tial to serve as mainstream storage for a
1. Introduction long period of time. To achieve this goal, candidate media must
overcome as many limitations of current systems as possible,
The digitization of the world has brought about a revolutionary
with clear advantages in terms of storage density, durability, and
transformation in our daily lives, yielding tremendous benefits
maintenance costs.
for society. However, this digital revolution has also precipitated
DNA, a biomolecule that nature has chosen to store ge-
an unprecedented surge in data generation, leading to significant
netic information in most living organisms, holds great promise
storage challenges. Based on current data expansion trends, it is
as a substrate for in vitro data storage[5] Compared to other
projected that the global data volume will reach 175 zettabytes by
biomolecules like RNA and proteins, DNA offers several distinct
2025.[1] To put that into perspective, the amount of data would
advantages for storing biological information in vivo. First, it ex-
require enough DVDs to circle the Earth 222 times.[1] As the vol-
hibits greater stability. Secondly, DNA has higher replication fi-
ume of data continues to grow at a rapid pace, the capacity of
delity due to the evolution of DNA replication mechanism.[6] Ad-
ditionally, DNA is composed of four different nucleotides that can
S. Wang, X. Mao, X. Zuo, C. Fan be combined in a programmable manner, providing the poten-
Institute of Molecular Medicine
Shanghai Key Laboratory for Nucleic Acids Chemistry and Nanomedicine
tial to carry enormous amounts of genetic information. These
Renji Hospital unique properties make DNA an attractive substrate for in vitro
School of Medicine data storage.[5,7] The stability of DNA makes long-term data stor-
Shanghai Jiao Tong University age significantly more feasible. In fact, it has been demonstrated
Shanghai 200127, China that DNA can remain stable for thousands of years under suit-
E-mail: [email protected]; [email protected]
able conditions.[8] Moreover, long-term DNA storage consumes
F. Wang, X. Zuo, C. Fan
School of Chemistry and Chemical Engineering little to no power, thereby reducing maintenance costs. The high
New Cornerstone Science Laboratory fidelity of DNA replication by DNA polymerase is beneficial for
Frontiers Science Center for Transformative Molecules the processes of data copying and retrieval. Furthermore, the abil-
Zhangjiang Institute for Advanced Study and National Center for ity to combine the four nucleotides arbitrarily provides an op-
Translational Medicine
Shanghai Jiao Tong University
portunity to increase data density. It has been calculated that
Shanghai 200240, China the theoretical data density of DNA is 1.7 × 1010 GB g−1 .[9]
With this capacity, it is estimated that the total global data
The ORCID identification number(s) for the author(s) of this article of 175 zettabytes in 2050 could be stored in just 10 kg of
can be found under https://doi.org/10.1002/adma.202307499
DNA.
DOI: 10.1002/adma.202307499

Adv. Mater. 2023, 2307499 2307499 (1 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 1. Overview of the major strategies and processes involved in data storage using DNA. There are two primary strategies for storing data in DNA:
the DNA sequence-based scheme and the DNA nanotechnology-based scheme. Six major steps are involved in DNA data storage, including encode,
write, preservation, retrieval, read, and decode.

The development of DNA data storage has advanced in par- 2. In Vitro DNA Data Storage
allel with our increasing comprehension of DNA and the ad-
vancements in biotechnology. The concept of storing data in Although DNA-based data storage differs significantly from con-
DNA was initially proposed in the mid-1960s,[10] shortly after a ventional data storage technologies, their fundamental storage
thorough understanding of the structure and chemical proper- processes share similarities. The process of storing data into
ties of DNA was attained. However, it was not until 1996 that DNA involves six major steps (Figure 1): 1) Encoding digi-
the first experimental demonstration of DNA data storage was tal data into specific formats (Converting information into nu-
achieved,[11] a decade after the automation and commercializa- cleotide sequence or structured patterns); 2) Writing data into
tion of DNA oligo synthesis.[12] Over the past two decades, re- storage medium (DNA oligos synthesis or DNA structure pro-
markable progress has been made in biotechnology, particularly gramming); 3) Data preservation (ensuring data accessibility, in-
in large-scale de novo DNA synthesis[12] and high-throughput tegrity, and authenticity by maximizing the long-term stability of
DNA sequencing.[13] These developments bring the opportunity DNA); 4) Accessing requested data (random access and data re-
for DNA data storage technology to flourish. Indeed, by lever- trieval); 5) Data reading (determining the precise order of nu-
aging next-generation DNA synthesis and sequencing technolo- cleotides or unique patterns); and 6) Decoding the underlying
gies, Church et al. successfully stored 5.27 megabits (MB) of data digital data (converting nucleotides or structures patterns back
in 54898 oligos in 2012, which were accurately retrieved through into digital information).[5,7a,16]
high-throughput sequencing.[14] Following a similar approach,
Goldman et al. stored 739 kilobytes (KB) of data in DNA oli- 2.1. Processes of In Vitro DNA Digital Data Storage
gos and suggested that scaling the system up could meet the de-
mand for global volume information storage.[15] These two pio- 2.1.1. Encoding
neering works have laid the foundation for next generation DNA
data storage research and have inspired numerous efforts world- As the initial step of DNA data storage, encoding refers to the
wide aimed at enhancing its performance in recent years. Con- process of mapping binary digital information into specific for-
sequently, substantial progress has been achieved in the field to mats suitable for storage in DNA using computer algorithms.
date.[5,7] Similar to encoding processes in other data storage technolo-
This review will commence by providing a brief yet compre- gies, one of the primary considerations for developing optimal
hensive overview of in vitro DNA-based data storage, covering encoding algorithms for DNA-based data storage is the storage
the key steps and various strategies employed. Subsequently, we capacity.[5,17] However, the unique nature of DNA presents addi-
will delve into in vivo DNA data storage, providing a concise sum- tional practical obstacles that must be considered in designing
mary of recent advancements in developing in vivo writing strate- encoding algorithms. These factors include challenges in syn-
gies. Furthermore, we will highlight the current challenges asso- thesizing DNA with long repetitive sequences (homopolymers)
ciated with each step of DNA data storage and introduce existing or unbalanced GC content, as well as the formation of unde-
cutting-edge technologies capable of overcoming these obstacles. sired secondary structures.[17] To date, many efforts have been
We hope that this review will provide readers with valuable infor- dedicated to overcoming those restrictions and developing en-
mation and ignite research enthusiasm in the field, ultimately coding algorithms to approach the maximal Shannon capacity of
driving the advancement of DNA-based data storage. 2 bits per nucleotide.[18] For instance, DNA fountain has been

Adv. Mater. 2023, 2307499 2307499 (2 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

demonstrated to achieve an information density of 1.57 bits per dehydrated DNA has been successfully stored on glass for scal-
nucleotide with desirable accuracy and fidelity.[18] For a more able DNA data storage[21] (Figure 2A), and remains stable at room
comprehensive overview of different encoding algorithms, we temperature for several years.[20b] While lower temperatures can
recommend referring to recent reviews.[5,19] increase the half-life of DNA,[22] the associated high cost of long-
term storage significantly outweighs this benefit. Other strate-
gies, such as storing in ethanol,[23] have also proven effective in
2.1.2. Write preserving DNA samples through dehydration. Furthermore, the
addition of additives (also known as DNA stabilizers)[20b,22a] and
Once the data has been encoded, it needs to be physically stored the use of specific storage plates[24] can offer additional protection
in a suitable medium using an appropriate method. In DNA- to DNA.
based storage, two commonly used writing schemes are em- Encapsulating DNA in an inorganic matrix emerges as one of
ployed: writing into DNA sequences and writing using DNA the most promising preservation methods for enhancing DNA
nanotechnology[5,7b] (Figure 1). Writing information into DNA stability. Silica is the most commonly used material for en-
sequences is the most straightforward approach and can be capsulating DNA, and it has been shown to significantly in-
achieved through de novo DNA synthesis. However, significant crease DNA stability by safeguarding it from environmental in-
advancements in DNA nanotechnology, such as DNA origami sults such as heat and oxidation.[25] Accelerated aging experi-
and in vitro DNA modification, have made it increasingly fea- ments have demonstrated that silica encapsulation can extend
sible to store data independent of DNA sequences. A detailed the lifespan of DNA data to 2000 years at 9.4 °C.[8a] Moreover,
comparison of these two storage schemes will be discussed be- combining silica-protected DNA with 3D printing technology
low. Nevertheless, when considering the writing methods, accu- also provides the opportunity to store digital data in arbitrary
racy, speed, and cost are key factors that need to be taken into shapes[26] (Figure 2B). Using magnetic nanoparticles as the in-
account. ner core, silica-protected DNA can also be handled more easily
(Figure 2C), in addition to its extended lifespan.[27] However, en-
capsulation also has drawbacks for DNA storage. Encapsulation
2.1.3. Preservation of DNA into inorganic matrix significantly decreases the infor-
mation density. With optimized assembly method, the best stor-
The lifespan of a storage medium is a critical factor in determin- age density achieved to date for silica encapsulation is 3.4 wt%.[28]
ing the reliability, convenience, and costs associated with long- Additionally, the processes of embedding and isolating DNA
term data storage, particularly for archival data. Current storage data can be intricate, time-consuming, and may involve the use
media, including magnetic, optical, and electrical storage devices, of highly toxic chemicals. Recently, Mao et al. achieved rapid
typically have a limited lifespan ranging from a few decades to encapsulation of DNA by metal-organic frameworks (MOFs)
150 years.[3] This short lifespan makes frequent data copying and for data storage[29] (Figure 2D). By integrating microfluidics,
transferring necessary for archival data storage. As mentioned the DNA encapsulation and extraction processes can be auto-
above, the stability of DNA outperformances all the storage media mated, completing within a few minutes. More importantly,
for several orders of magnitude under desirable conditions. How- MOFs provide the advantage of enhancing DNA chemical and
ever, owing to its unique properties, DNA is vulnerable to specific thermal stability while exhibiting negligible toxicity.[29] Regard-
factors such as UV irradiation, ionizing radiation, DNA enzymes, ing storage density, coprecipitation with salt has been demon-
among others. Those factors mainly alter the integrity of DNA by strated to achieve a remarkable DNA loading density of 38 wt%
causing strand breaks, hydrolytic damage, and nucleobase mod- (Figure 2E). This method also makes retrieval of the DNA signifi-
ifications. Therefore, it is important to design preservation ap- cantly easier, albeit with a decrease in DNA lifespan compared to
proaches accordingly to extend the lifespan of DNA. Fortunately, encapsulation.[30] Ongoing efforts in the field are expected to fur-
biologists have developed various strategies to increase the in- ther enhance loading density and procedural convenience in the
tegrity of DNA in vitro. However, since the demands of storing future.
biological DNA samples and DNA containing digital information Preserving data in living organisms presents another feasible
differ significantly, these strategies need to be optimized before strategy for DNA-based data storage. Data-containing DNA frag-
being applied to the preservation of DNA used for data storage. ments can be assembled into artificial chromosomes or loaded
Ideally, storage methods for data containing DNA should con- into plasmids to be stored in yeast or bacteria. Due to the high
sume little to no energy (such as storage at room temperature), fidelity of DNA replication in living organisms, the amplifica-
have minimal impact on data density, and facilitate random data tion of digital data stored in them is believed to be more accu-
access. To date, there are three main categories of strategies used rate and efficient than other in vitro amplification methods. In
for preserving DNA data: 1) dehydration; 2) encapsulation; and 2003, it was demonstrated that digital data can be stored in bac-
3) in vivo preservation. teria, although the volume of stored data at that time was low.[31]
Water can be detrimental to DNA stability as it can acceler- With the advancement of high-throughput DNA synthesis tech-
ate the hydrolytic damage and enzymatic degradation of DNA. nology, 445 kb data have been successfully stored in a popula-
Therefore, methods inducing dehydration of DNA hold potential tion of bacteria through plasmid[32] (Figure 2F). Using the CIR-
for extending its lifespan. Studies have demonstrated that dry- SPR/Cas technique, digital data encoded into DNA was directly
state DNA is more stable than DNA in solution,[20] making the stored in the genome of a population of bacteria.[33] Artificial
storage of dehydrated DNA at room temperature an appealing chromosomes offer the opportunity to increase the volume of
approach for preserving digital data encoded into DNA. Indeed, data encoded in a single living organism. Using yeast artificial

Adv. Mater. 2023, 2307499 2307499 (3 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 2. Preservation strategies for DNA data storage. A) Scalable DNA data storage by storing dehydrated DNA in physically isolated spots on glass.
Adapted with permission.[21] Copyright 2019, Springer Nature. B) Data preservation through encapsulation of DNA within glass particles. Silica-protected
DNA data can be further blended into PCL filament and 3D printed into various shapes, such as the Stanford Bunny. Adapted with permission.[26]
Copyright 2020, Springer Nature. C) DNA absorption onto the surface of positively charged magnetic nanoparticles, followed by silica encapsulation,
enables DNA recovery through magnetic separation. Adapted with permission.[27] Copyright 2014, American Chemical Society. D) Enhanced stability
of data-encoded DNA through fast MOFs encapsulation of DNA immobilized on optically addressable SiO2 NPs. The entire process can be integrated
into microfluidic platforms for automated data storage. Adapted with permission.[29] Copyright 2023, American Chemical Society. E) Long-term DNA
data storage achieved by co-precipitation of DNA with earth alkaline salts. Adapted with permission.[30] Copyright 2020, Royal Society of Chemistry.
F) Large-scale data storage in a mixed culture of bacteria cells. DNA oligos are assembled into plasmids and further stored into bacteria. Adapted
with permission.[32] Copyright 2020, Springer Nature. G) Construction of an artificial chromosome for large-scale in vivo data storage. Adapted with
permission.[34] Copyright 2021, Oxford University Press.

chromosome, it has been demonstrated that more than 241 000 2.1.5. Read
base pair DNA sequences can be used to store information in
a single cell[34] (Figure 2G). Collectively, those data indicate that Once retrieved, data are ready to be read accurately and com-
in vivo preservation holds great potential for DNA-based data pletely. The reliability of the reading techniques is crucial to en-
storage. sure that the recovered data is accurate and free from corruption.
Additionally, the speed of the reading process is another key fac-
tor to consider for the practical applications of the storage tech-
2.1.4. Random Access nique. Depending on the storage media and the chosen writing
strategies, the reading process can vary significantly. For DNA
Random access refers to the efficient and rapid retrieval of re- sequence-based storage, sequencing methods, including all three
quested data from large storage pools and is a crucial element generations of sequencing techniques, are typically used to read
of any data storage system. While achieving random access is the data. For data stored using DNA nanotechnology, direct vi-
relatively straightforward in traditional storage media using ad- sualization techniques such as fluorescence microscopy, atomic
dressing schemes and data indexing, it presents greater chal- force microscopy, electron microscopy, and gel electrophoresis,
lenges in DNA-based data storage. This limitation becomes par- along with advanced nanopore techniques, have been employed
ticularly relevant when storing data that requires frequent ac- to read the data according to the chosen writing strategies. In gen-
cess. Nevertheless, significant progress has been made in im- eral, the reading techniques of DNA data storage largely depend
proving the random-access capability of DNA data storage. For on the writing strategies and will be introduced in detail together
DNA sequence-based data storage, methods such as PCR ampli- with the writing strategies later.
fication using unique primers[9,35] and physical isolation[36] have
been demonstrated to enable random access. In contrast, random
access for DNA nanotechnology-based data storage is not yet as 2.1.6. Decoding
well developed. The benefits and drawbacks of different access-
ing strategies, as well as potential approaches for enhancing ran- Decoding is the reverse process of encoding that transforms the
dom access in DNA nanotechnology-based data storage, will be readout from the previous step back into the original file us-
discussed in detail below. ing corresponding algorithms. In fact, the algorithms used for

Adv. Mater. 2023, 2307499 2307499 (4 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 3. Writing digital data into DNA sequence. A) Enhancing information capacity using degenerate bases. Schematic representation of data storage
with degenerate bases (left) and the oligo pool synthesis process (right). Adapted with permission.[37b] Copyright 2019, Springer Nature. B) Natural and
artificial nucleotides. Adapted with permission.[38] Copyright 2019, American Association for the Advancement of Science. C) Improved coding efficiency
in digital data storage using artificial nucleotides as additional encoding characteristics. Adapted with permission.[37a] Copyright 2022, John Wiley and
Sons. D) Digital data writing into DNA sequence using the TdT enzyme. Adapted with permission.[40a] Copyright 2019, Springer Nature.

encoding and decoding are closely related. As mentioned above, 2.2.1. DNA Sequence-Based Data Storage
achieving high data density is a primary consideration of encod-
ing/decoding algorithm. An ideal encoding/decoding algorithm The primary advantage of using DNA sequence to store digital
for DNA-based data storage should also possess error-correction data lies in its high loading density. Each nucleotide in the four
capabilities as errors occur inevitably in many steps, especially natural bases can theoretically represent 2 bits (A = 00, C = 01,
during writing and reading. To develop error-correcting schemes, G = 10, T = 11).[18] By introducing more coding blocks in ad-
adding logical redundancy is one of the most commonly used dition to the natural bases, the information capacity can be fur-
strategies, which has been demonstrated in numerous studies. ther increased (the theoretical coding efficiency is log2 (number
Additionally, the utilization of physical redundancy also benefits of coding characters)).[37] For example, an information capacity
the error-correction process, albeit with some sacrifice in data of 3.37 bits/character was achieved experimentally by introduc-
density. As technology continues to advance, the error rate dur- ing eleven degenerate nucleotide combinations as additional cod-
ing each step of DNA storage is likely to decrease further. Those ing characters[37b] (Figure 3A). Using artificial nucleotides is an-
advancements will change the scheme of developing suitable en- other way to increase the coding density. P, Z, S, and B are four
coding/decoding algorithms. However, the fundamental criteria artificial nucleotides designed to form P–Z and S–B orthogonal
for a good algorithm, ensuring data accuracy while maximizing base pairs that are compatible with their natural counterparts[38]
data loading density, will remain unchanged. (Figure 3B). By incorporating these four artificial nucleotides
as additional coding characters, the coding efficiency has been
shown to increase to 2.35 bits per nucleotide[37a] (Figure 3C).
2.2. Strategies for In Vitro DNA Digital Data Storage Write: The process of writing digital information into DNA
sequence involves DNA synthesis, with chemical synthesis being
As mentioned above, the strategies used to store digital data in the most commonly used method in vitro.[39] Currently, digital in-
DNA in vitro can be classified into two categories: DNA sequence- formation is usually segmented and stored in a large pool of DNA
based scheme and DNA nanotechnology-based scheme. Storing strands, each with a length of less than 200 nucleotides.[5] Spe-
data into DNA sequence is achieved through the unique arrange- cific DNA strands are built up in a nucleotide-by-nucleotide man-
ment of the four nucleotides. On the other hand, DNA nanotech- ner during traditional chemical synthesis procedure, which is
nology is also emerging as a promising strategy due to several time-consuming for large-scale synthesis. The emergence of var-
advantages. These two strategies primarily differ in the write and ious microarray formats provides the opportunity to synthesize
read steps. In addition, the preservation and the random access distinct DNA strands in parallel, significantly increasing writing
of DNA nanotechnology-based storage schemes are less well- speed[39] and making large-scale data storage possible. In fact,
studied and differ to some extent from those of DNA sequence- both Church et al. and Goldman et al. used Agilent’s Oligo Li-
based storage. In this section, we will provide a detailed descrip- brary Synthesis microarray platform to synthesize the oligo pool
tion and comparison of those two strategies. in their pioneering works.[14,15]

Adv. Mater. 2023, 2307499 2307499 (5 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 4. PCR-based random-access strategies. A,B) Selective data retrieval by PCR. A) Schematic representation of the selective retrieval process. B)
Design and utilization of a random access primer library for data retrieval. Adapted with permission.[35b] Copyright 2018, Springer Nature. C,D) Selective
data retrieval by nested PCR. C) Schematic representation of the nested PCR-based retrieval process. D) Large-scale DNA memory based on nested PCR.
Adapted with permission.[35c] Copyright 2008, Springer Nature.

Enzymatic synthesis is considered as an attractive alternative Lin et al. that specific data can be extracted through magnetic sep-
approach to synthesis DNA strands in vitro. Terminal deoxynu- aration independent of PCR (Figure 5B). More importantly, this
cleotidyl transferase (TdT) is a particularly promising enzyme for system enabled in-storage file operations, including lock-unlock,
DNA synthesis due to its ability to efficiently elongate polynu- rename, and deletion.[44]
cleotide chains independent of a template.[39] Using TdT to syn- Encapsulating data into microcompartment also facilitates
thesize DNA strands with short homopolymeric extensions, re- random access. For example, Bögels et al. recently demon-
searchers have successfully stored digital information into a DNA strated that storing DNA data in thermoresponsive microcap-
strand by employing the transition of non-identical nucleotides sules enables multiplexed and repeated random data access[45]
as the encoding scheme[40] (Figure 3D). TdT-based enzymatic (Figure 5C). They achieved this by enhancing the fidelity of PCR
DNA synthesis has also demonstrated its capability to record tem- amplification through constraining the reaction into microreac-
poral signals by harnessing the nucleotide selectivity of TdT un- tors. Furthermore, their platform is compatible with fluorescent
der different conditions.[41] Upon optimization, the incorpora- barcoding and sorting, further enhancing the capability of ran-
tion of nucleotides by TdT can be engineered in a controlled man- dom data retrieval.[45] Banal et al. achieved random access by
ner to synthesize user-defined DNA sequence.[39,42] It is widely barcoding silica-encapsulated DNA data with three 25-nucleotide
anticipated that those strategies, which have been embraced by single-stranded DNA (ssDNA) sequences[46] (Figure 5D). As a
numerous companies,[39] will advance the enzymatic synthesis result, specific files can be extracted with high selection sensi-
of DNA, thereby facilitating the progress of DNA data storage. tivity (one in 106 files) through fluorescent sorting. Moreover,
Random Access: PCR is the predominate method for access- the barcodes represent file metadata, enabling content-based file
ing information in DNA sequence-based storage. By introduc- searches.[46] In fact, content-based filter queriers are attractive
ing orthogonal primer pairs, specific datasets with unique primer forms of random access in practical DNA data storage. Using
pairs can be extracted in a convenient and multiplexed way[9,35a,b] probe hybridization and magnetic separation, Bee et al. demon-
(Figure 4A). Through the design of primer libraries, it has strated the successful retrieval of similar images from a DNA-
been shown that 35 distinct files stored in more than 13 mil- based database of 1.6 million images[47] (Figure 5E). Addition-
lion DNA oligos can be individually retrieved without errors[35b] ally, the introduction of content-based barcodes has also been pro-
(Figure 4B). Systematic investigations have also revealed that posed recently to achieve content-based filter queries.[48]
PCR-based random access can faithfully recover files as long as Read: DNA sequencing is the sole method for reading out
there are more than ten copies of the template present in the in- the data stored in DNA sequences. The past two decades have wit-
formation pool.[9] Nested PCR (Figure 4C), which uses hierarchi- nessed significant advances in DNA sequencing. Until now, three
cal primers, may further increase the random accessibility of data generations of sequencing techniques have been developed.[49]
retrieval.[35c,43] Through careful design, Yamamoto et al. created All three generations of DNA sequencing techniques have been
16.8 million addresses and successfully demonstrated the extrac- employed in DNA data storage. Sanger sequencing is a represen-
tion of data embedded into a unique address through nested tative first-generation DNA sequencing method with a maximum
PCR[35c] (Figure 4D). read length of approximately 1000 bp. It has been used to read
In addition to PCR, physical isolation methods have also been DNA data stored into individual DNA strand.[35a,50] However, it is
applied for random access. For example, Tomek et al. achieved not suitable for reading large-scale data stored in an oligo pool, as
random access of specific files by adding chemical handles (in- even the co-sequencing of two DNA strands significantly impacts
cluding biotin–streptavidin, fluorescein–antibody, digoxigenin– the readouts.[50] Second-generation DNA sequencing techniques
antibody, polyA–polyT oligomers) to unique primers (Figure 5A). massively increase throughput through parallelization. Illumina
After emulsion PCR amplification, the corresponding data could sequencing, the most popular second-generation method, typi-
be extracted using magnetic beads.[36] Using single-strand over- cally has a read length of less than 300 bp.[49a] Since the DNA
hangs as the physical address, it has also been demonstrated by oligos used for large-scale in vitro data storage are currently

Adv. Mater. 2023, 2307499 2307499 (6 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 5. Physical isolation-based random-access strategies. A) Selective data retrieval by the addition of primers with chemical handles followed by
physical isolation. Adapted with permission.[36] Copyright 2019, American Chemical Society. B) Dynamic DNA data storage system utilizing a T7 promoter
and a single-stranded overhang domain. The single-stranded overhang domain allows for selective retrieval and in-storage file operations such as lock,
unlock, rename, and delete. Repeated random access to data stored by this strategy can be achieved through non-PCR-based magnetic separation, in vitro
transcription, and reverse transcription. Adapted with permission.[44] Copyright 2020, Springer Nature. C) Repeated random multiplexed data access
of DNA data stored in thermoresponsive microcapsules. Compartmentalization of DNA files into thermoresponsive and semipermeable microcapsules
enables data amplification using multiplex PCR and data recovery using magnetic separation after PCR. Selective access to files can be achieved through
fluorescent barcoding. Adapted with permission.[45] Copyright 2023, Springer Nature. D) Content-addressable DNA data storage system. Addressing
DNA data encapsulated in silica capsules with content barcodes using orthogonal 25-nucleotide ssDNA strands allows for scalable indexing and Boolean
logic selection and retrieval. Adapted with permission.[46] Copyright 2021, Springer Nature. E) Similarity search in DNA data storage. Through feature-
to-sequence encoding, hybridization probes, used as queriers in the system, preferentially bind to targets representing visually similar images. After
targets filtering with magnetic beads, images with similar features can be retrieved for decoding. Adapted with permission.[47] Copyright 2021, Springer
Nature.

less than 200 nucleotides, Illumina sequencing is an ideal plat- date. They differ in terms of throughput and read length. These
form for reading such data and has been widely adopted.[9,14,15] factors should be carefully considered when selecting a suitable
The nanopore technique, a third-generation DNA sequencing reading method for a specific set of data.
method, enables direct sequencing of single DNA molecules in
a high-throughput manner.[49a] With its ability to provide ultra-
long reads of up to 4 megabases, nanopore sequencing is an at- 2.2.2. DNA Nanotechnology-Based Data Storage
tractive approach for reading data stored in DNA. Indeed, many
studies have demonstrated the use of nanopore sequencing for The advancement of DNA nanotechnology has led to a revo-
reading DNA data.[35b,51] In addition, Pacific Biosciences (PacBio) lutionary breakthrough in the precise control and manipula-
sequencing technology, a single-molecule real-time (SMRT) se- tion of DNA molecules, enabling unprecedented capabilities.
quencing method capable of producing read lengths exceed- This exceptional ability to manipulate DNA at the nanoscale has
ing 10 kb, represents another compelling option among third- opened up new avenues for DNA data storage. In the realm
generation sequencing techniques for reading DNA data.[52] In of DNA nanotechnology-based data storage, specific patterns
summary, all three generations of DNA sequencing techniques are meticulously designed and generated to store digital data,
have been used to read digital data stored in DNA sequences to taking advantage of the remarkable programmability in DNA

Adv. Mater. 2023, 2307499 2307499 (7 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

nanotechnology. To date, various encoding blocks, such as DNA presence or absence can be detected by nanopore sensing, mak-
strands (including modified versions like biotin and fluorescence ing nanopore sensing a predominant data reading approach. For
modifications), DNA nanostructures, and DNA topology, have example, the Keyser group demonstrated the use of DNA dumb-
been utilized to generate customized patterns tailored to individ- bell hairpins as representations of the bit state for data storage.[62]
ual data storage needs. As mentioned above, the reading strate- They utilized a 7228 bp single-standard M13 DNA as the scaffold
gies in DNA nanotechnology-based data storage are intricately and assembled oligos with or without hairpin structures on it.
linked with the writing methods. In this section, we will pro- The presence of the hairpin structures caused a signal change
vide a comprehensive overview of the writing and reading steps when passing through a nanopore, enabling the readout of data
involved in DNA nanotechnology-based data storage. Addition- by nanopore sensing (Figure 6G). Leveraging this strategy, they
ally, we will discuss the key features of framework nucleic acids developed a 3-bit barcoding system that could be accurately de-
(FNAs) that can be explored for DNA data storage. coded using solid state nanopores.[62] Subsequently, the same
Read and Write: SsDNA is one of the simplest elements group conducted a series of optimizations to improve the capa-
that can be used to construct specific geometric patterns on bility of data storage. By increasing detection sensitivity using
substrates through assembly. Once assembled, the ssDNA can nanopores with smaller diameters, they were able to distinguish
serve as encoding characters to represent binary states or as ad- individual small DNA structures, including 8 bp hairpins.[63]
dress sites for further assembly of additional encoding charac- Consequently, the density of DNA nanostructures assembled on
ters. Early studies have demonstrated information storage by as- the scaffold could be significantly increased, thereby improving
sembling partially complementary DNA strands to form struc- the data loading density. Using 8 bp and 16 bp hairpins to rep-
tured scaffolds with single-stranded overhangs. These overhangs resent “0” and “1” respectively, the Keyser group successfully en-
were designated as “0” and could be converted to “1” when chang- coded 56-bit data into a 7228 bp M13 DNA and decoded it through
ing to double stranded duplexes by introducing complementary nanopore sensing.[63] The storage capacity could be further en-
strands (Figure 6A). Using 3 overhangs, a total of 8 memory hanced by linking information carriers together.[63] In addition
states were achieved and read out using gel electrophoresis.[53] to hairpins, other nanostructures can also serve as the encoding
Apart from double-standard DNA (dsDNA), DNA origami can characters (Figure 6G). For example, multi-way junctions of vary-
also serve as the scaffold. For example, six-helix bundle DNA ing sizes have been assembled on DNA carriers through toehold-
nanorods, with a length of ≈800 nm, were used to position mediated strand displacement reactions to store digital data such
fluorescent-labeled DNA strands (Figure 6B). While initially used as 2D grayscale image.[64] Expanding the repertoire of nanostruc-
for creating barcodes, this system could also be adopted for in- tures also contributes to increased storage capacity. Furthermore,
formation storage and subsequently read out using total inter- beyond DNA, alternative scaffolds can be employed as carriers
nal reflection fluorescence microscopy and DNA-PAINT super- of DNA nanostructures for data storage. For example, Zhang
resolution microscopy.[54] By assembling biotin-modified DNA et al. have demonstrated the use of carbon nanotubes as scaf-
onto DNA origami to create specific patterns, Zhang et al. demon- folds for patterning tubular nucleic acids to achieve information
strated the storage of diverse data formats, including text, musi- storage[65] (Figure 6H).
cal notes, and images (Figure 6C). Those data could be securely The exceptional programmability of base pairing in DNA of-
read out by AFM due to the capability of encryption.[55] Data en- fers the opportunity to induce conformational changes through
coded by biotinylated DNA strands have also been read out using simple DNA strand hybridization. Exploiting such conforma-
nanopore sensing, through the using of streptavidin to cause sig- tional changes has been employed for data storage purposes.
nal changes during nanopore measurement[56] (Figure 6D). Dick- Gel electrophoresis, which effectively resolves the conforma-
inson et al. encoded data using ssDNA with docking-site domains tional differences of DNA, serves as a convenient method for
(representing “1”) or without (representing “0”) on DNA origami data readout in such storage systems. For example, the close
(Figure 6E). The data was read out by monitoring the binding and open conformations of dsDNA have been utilized as me-
of fluorescent imager probes using DNA-PAINT.[57] DNA can chanical bits for data storage. Using 8 oligos to represent
also be easily assembled on other substrates, such as silicon 8 different bits, an 8-bit ASCII character encoded text mes-
wafers[58] and gold surfaces,[59] to form unique patterns. In the- sage “Hello world” was successfully stored and decoded by gel
ory, those substrates can also be used to pattern DNA strands electrophoresis.[66] Chandrasekaran et al. demonstrated the cre-
for data storage. For example, ssDNA has been immobilized on ation of multi-bit memory devices by assembling ssDNA on
a microfluidics chip for data storage, where binary states were M13 scaffold. By incorporating 6 address sites, a 5-bit memory
represented by fluorescence signals and read out using a fluo- system was developed, enabling the storage of text message by
rescence microarray scanner.[60] More importantly, by immobi- controlling the conformational states of DNA nanostructures[67]
lizing ssDNA on individually addressable electrodes, data writing (Figure 6I). Modifying the topology of dsDNA has also been
can be controlled by electric field-induced hybridization of fluo- harnessed for data storage. Tabatabaei et al. achieved success-
rescent probes (Figure 6F). Using this approach, a 3-bit memory ful data storage by introducing nicking positions on dsDNA[68]
device was created, and the stored data was read through fluores- (Figure 6J). By using the DNA-guided programmable restric-
cence imaging.[61] tion enzyme Pyrococcus furiosus Argonaute (PfAgo), nicking
Using ssDNA as address sites to immobilize other encoding sites can be generated in a programmable and multiplexed man-
characters, especially DNA nanostructures, is another popular ner. Therefore, by converting digital data into nicking posi-
strategy for data storage. To date, a wide range of DNA nanos- tions on the recording DNA, data can be efficiently written us-
tructures has been employed to create specific patterns on ds- ing PfAgo and subsequently read out through high-throughput
DNA. Due to the unique properties of those nanostructures, their sequencing.[68]

Adv. Mater. 2023, 2307499 2307499 (8 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 6. Representative strategies of digital data storage based on DNA nanotechnology. A) Rewritable 8-state DNA memory device. The DNA scaffold
contains three independent and reversible addressing sites at 7 nm spacing, enabling data recording and erasing by DNA hybridization. The correspond-
ing states can be read out by gel electrophoresis. Adapted with permission.[53] Copyright 2004, American Chemical Society. B) Submicrometric fluores-
cent barcodes based on DNA nanorods. The fluorescent barcodes can be easily decoded by epifluorescence or DNA-PAINT super-resolution imaging.
Adapted with permission.[54] Copyright 2012, Springer Nature. C) Digital data storage and encryption through nanopatterning of biotin-modified ssDNA
on DNA origami. Adapted with permission.[55] Copyright 2019, Springer Nature. D) DNA hard drive using biotin-modified ssDNA as the encoding char-
acters. Decoding of the data can be achieved through nanopore sensing by adding streptavidin. Adapted with permission.[56] Copyright 2020, American
Chemical Society. E) Digital nucleic acid memory with data encoding into DNA origami and reading through DNA-PAINT. Adapted with permission.[57]
Copyright 2021, Springer Nature. F) Random access writing and reading operations through electric field-induced hybridization on addressable electrode
arrays. Adapted with permission.[61] Copyright 2018, Springer Nature. G) Digital data storage through DNA nanostructures assembly. Multiple DNA
nanostructures, including DNA dumbbell, DNA hairpin, and DNA multi-way junctions, can be used as the encoding characters for data storage. Adapted
with permission.[63,64] Copyright 2019, American Chemical Society; Copyright 2021, John Wiley and Sons. H) Encoding Carbon Nanotubes with tubular
nucleic acids for information storage. Adapted with permission.[65] Copyright 2019, American Chemical Society. I) Rewritable multi-bit DNA memory
system with information encoded in the conformations of a DNA nanoswitch. Adapted with permission.[67] Copyright 2017, Oxford University Press. J)
DNA punch cards for digital data storage through topological modification via enzymatic nicking. Adapted with permission.[68] Copyright 2020, Springer
Nature. K) RNase H-assisted decoding and encryption of information stored in DNA nanoswitch libraries. Adapted with permission.[70] Copyright 2023,
American Chemical Society.

There are several advantages of storing data using DNA oped which also facilitate the design DNA sequences to enhance
nanotechnology. 1) High writing speed: As introduced above, the writing speed.[69] 2) Convenient data erasing and rewrit-
DNA strand hybridization is a prominent method employed for ing: The erasing and rewriting of data in DNA nanotechnology-
data writing. Benefit from the high programmability of DNA based storage system is relatively easy and fast. For data stored
Hybridization, writing at different sites can be implemented via DNA hybridization, toehold-mediated DNA strand displace-
simultaneously.[61,63] The fast kinetics of DNA hybridization also ment reactions have been harnessed to achieve convenient data
makes the writing at each site relatively fast. Moreover, predic- erasing.[56,64,67] Following erasure, new encoding characters can
tion algorithms of DNA hybridization kinetics have been devel- be hybridized onto the scaffolds to store another set of data. In

Adv. Mater. 2023, 2307499 2307499 (9 of 23) © 2023 Wiley-VCH GmbH

 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

addition to toehold-mediated strand displacement, electric field- of dimensional forms, including linear, planar, and volumetric
induced strand displacement has also been successfully applied shapes.[75,76] In theory, by employing meticulous design based
in data rewriting.[61] In the case of data encoded by nicking on base pairing, FNAs can be custom-built with various sizes
sites on dsDNA, erasure can be achieved by DNA ligation us- and shapes to meet specific requirements.[76] On the other hand,
ing enzymes such as T4 DNA ligase.[68] 3) Data encryption: The FNAs can serve as versatile platforms for the precise organiza-
self-assembly nature in DNA nanotechnology-based data stor- tion of functional molecules at the nanometer scale.[75,76] Addi-
age provides the opportunity for data encryption. Encryption can tionally, FNAs exhibit enhanced stability and structural rigidity
be achieved by simply omitting key elements during the data compared to unstructured DNA strands.[75] All those key features
writing process.[55,56,66] Alternatively, Talbot et al. developed a make FNAs promising platforms for DNA data storage.
stimuli-responsive system that can be used for data encryption[70] The exceptional addressability of FNAs with nanometer res-
(Figure 6K). In their system, both DNA and RNA strands were olution has been convincingly demonstrated by researchers for
used to induce conformational change of DNA nanostructures precise organization of a diverse range of molecules, includ-
for data writing. RNase H, an endonuclease that selectively ing proteins, nucleic acids, and inorganic nanoparticles, among
cleaves the RNA in a DNA–RNA hybrid duplex, allows the con- others.[75] This capability has been successfully applied to con-
formational changes induced by RNA strands to be erased. Thus, trol the spatial arrangement of enzymes within multienzyme
data can be encrypted by the use of DNA or RNA strands and complexes, leading to optimized catalytic activity. For example,
decrypted through RNase H digestion.[70] 4) In-memory compu- Fu et al. employed FNAs as a platform and successfully show-
tation: DNA strand displacement reactions, extensively explored cased the precise positioning of enzyme pairs with nanometer-
and successfully employed in constructing biological computing scale proximity[77] (Figure 7A). The controlled organization of in-
circuits, are compatible with DNA nanotechnology-based data organic nanoparticles further attests to the high controllability
storage.[71] Consequently, these computing circuits can be inte- of FNAs. Utilizing FNAs, well-ordered arrays of Au nanoparti-
grated into the storage systems to construct in-memory compu- cles have been assembled with interparticle distances of less than
tation systems. Indeed, logic operations[61,67] and cascade logic 2 nm.[78] Indeed, a typical planar DNA nanostructure, with a sur-
gates[72] have been successfully incorporated into DNA data face area of 8000 to 10000 nm2 , is folded using approximately
storage systems. Thus, DNA nanotechnology-based data storage 200 staples[76,79] (Figure 7B). Each of those staples can serve as a
holds great promise as a versatile system for in-memory compu- unique addressable point. This design enables minimal distances
tation. between two addressable sites to be as small as 5 nm under the
Preservation and Random Access: In comparison to DNA current settings.
sequence-based data storage, the preservation and random access The remarkable addressability of FNA can be leveraged to ad-
of DNA nanotechnology-based data storage are still relatively less dress the writing speed challenges faced by DNA data storage.
developed. The preservation of data stored by DNA nanotechnol- As discussed earlier, data writing speed is primarily constrained
ogy presents a greater challenge due to the more stringent condi- by the capacity of parallel DNA synthesis. Increasing the syn-
tions required to maintain the integrity of DNA nanostructures thesis density is a practical approach to improving DNA syn-
compared to DNA strands. However, significant progress has thesis throughput while maintaining a realistic infrastructure
been made in developing storage conditions that can effectively footprint.[39] One of the most effective strategies to enhance syn-
preserve the integrity of DNA origami, such as lyophilization[22b] thesis density is by minimizing the physical distance between
and cryopreservation.[73] These methods hold promise in meet- synthesis sites. For example, Microsoft achieved the synthesis of
ing the specific preservation needs of data stored by DNA nan- DNA with a density of 2.5 × 107 unique sequences per square
otechnology. centimeter by reducing the distance between two synthesis sites
Although the random access of data stored by DNA nanotech- to 2 μm, with the potential for further density increases by de-
nology remains relatively unexplored, a few studies have demon- creasing the inter-microelectrode distance.[80] By capitalizing on
strated its feasibility. For instance, Boskovic et al. showed that the minimal distance of 5 nm between addressable sites in FNAs,
data stored by assembling DNA dumbbell on ssDNA could be the theoretical DNA synthesis density can reach 4 × 1012 unique
randomly accessed through PCR amplification.[74] sequences per square centimeter when using large-area arrays
To achieve this, they first ligated the nicks between the indi- of FNAs[81] as the synthesis platform. This substantial increase
vidual encoding DNA dumbbells, resulting in an intact double- in synthesis throughput will also contribute to reducing the cost
stranded structure. They demonstrated that after ligation, the associated with DNA synthesis,[39] which is another challenge en-
data could be amplified without significantly affecting the read- countered in DNA data storage.
out accuracy. By introducing content-specific indexes, random Another advantageous feature of FNAs applicable to DNA data
access was achieved through PCR amplification using unique storage is their capability to generate diverse and customizable
primer pairs, enabling targeted retrieval of specific data.[74] In structures.[76,79,82] Utilizing computer-aided design, researchers
addition to PCR, toeholds have also been harnessed to achieve worldwide have successfully created a wide range of 2D and 3D
random access when data was stored by topological modification FNAs (Figure 7C,D). Furthermore, through programmed assem-
of DNA.[68,72] bly, small DNA nanostructures can be organized into intricate
Framework Nucleic Acids for DNA Data Storage: FNAs refer DNA lattices, further expanding the category of FNAs available
to a class of well-defined nanostructures based on nucleic acids, for use.[76b] For instance, Wei et al. utilized single-stranded tiles
constructed utilizing diverse DNA nanotechnologies, particularly as the initial framework and successfully constructed an impres-
DNA origami.[75] These nanostructures, ranging in size from sive array of 107 distinct and complex 2D shapes[83] (Figure 7E).
tens of nanometers to sub-micrometers, exhibit a wide range In the context of DNA nanotechnology-based data storage, the

Adv. Mater. 2023, 2307499 2307499 (10 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 7. Key advantages of framework nucleic acid that holds potential in DNA data storage. A) Precise spatial organization of multi-enzyme complexes
using DNA nanostructures. Adapted with permission.[77a] Copyright 2014, Springer Nature. B) Principle of complex nanostructure assembly using DNA
origami technology. Adapted with permission.[76a] Copyright 2021, Springer Nature. C) Representative 2D planar shapes of DNA origami. Adapted with
permission.[79] Copyright 2006, Springer Nature. D) Representative 3D shapes of DNA origami. Adapted with permission.[82] Copyright 2015, Springer
Nature. E) Assembly of complex shapes from single-stranded DNA tiles. Adapted with permission.[83] Copyright 2012, Springer Nature. F) Metallization
of DNA origami with low-valence metal ions (Cu2+ and Ag+ ). Adapted with permission.[85a] Copyright 2019, Springer Nature. G) Thickness controllable
mineralization of DNA nanostructures with silica. Adapted with permission.[86] Copyright 2018, Springer Nature.

incorporation of nanostructures with varying geometries and loading density, as compared to existing chemical encapsulation
sizes enriches the pool of encoding blocks. Considering that the strategies.
theoretical coding efficiency is determined by the logarithm to
the base 2 of the number of distinct coding blocks, FNAs hold
great potential for increasing the information storage capacity in 3. In Vivo DNA Data Storage
DNA nanotechnology-based data storage.[55]
The third feature of FNAs that can be applied to DNA data As mentioned above, data encoded into in vitro synthesized DNA
storage is their enhanced stability. Experimental evidence has can be preserved in living organisms. This can be perceived as the
demonstrated that FNAs exhibit greater stability compared to nascent stage of in vivo DNA data storage. Thanks to the advances
their single-stranded counterparts under diverse conditions.[75,84] in synthetic biology, in vivo DNA data storage is rapidly evolving
Furthermore, researchers have achieved precise metallization[85] towards a more intelligent direction, involving direct collection
(Figure 7F) and mineralization[86] (Figure 7G) of FNAs. DNA en- and recording of new data in response to stimuli. In this section,
capsulation provides a protective shield that can protect DNA we will briefly introduce the various strategies employed for in
from degradation in harsh conditions.[87] Importantly, compared vivo DNA data storage, along with their applications.
to other chemical encapsulation methods, the precise metalliza- The foundation of using DNA as a molecular recorder in vivo
tion and mineralization of FNAs may have significantly lesser relies on the presence of specific writers capable of engineer-
impact on the DNA loading density. For example, the controlled ing DNA in a pre-defined manner. Fortunately, several enzymes,
silicification on FNAs only results in the formation of a thin such as recombinases and CRISPR-Cas nucleases, have demon-
silica coating layer with an average thickness of approximately strated their capability in the modification and manipulation of
3 nm.[86] Therefore, these techniques hold immense promise for DNA. Indeed, both recombinases and CRISPR-Cas nucleases
DNA data preservation while exerting minimal influence on data have been successfully employed for in vivo DNA data storage.

Adv. Mater. 2023, 2307499 2307499 (11 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 8. In vivo information writing using site-specific DNA recombinase. A) Depending on the orientation of the recombining sites, DNA recom-
bination can result in either excision or inversion of DNA segments. Recombinase-mediated cassette exchange enables precise substitution of DNA
fragments at specific sites. B) Finite-state machine-like double inversion recombination switch for heritable data recording. Adapted with permission.[89]
Copyright 2008, Public Library of Science. C) Rewritable digital data storage achieved by recombinase-mediated inversion and restoration of specific
DNA sequences. Adapted with permission.[90] Copyright 2012, The National Academy of Sciences. D) Recombinase-based state machines (RSMs) for
information recording based on DNA excision and inversion operations. Adapted with permission.[92] Copyright 2016, American Association for the
Advancement of Science. E) Precise in vivo DNA writing for analog information recording using recombinase-mediated DNA fragment substitution.
Adapted with permission.[93] Copyright 2014, American Association for the Advancement of Science.

DNA recombinases are enzymes that recognize specific short three orthogonal recombinases, they successfully recorded the
DNA sequences (sites) and catalyze recombination events be- temporal order of inputs within a 16-state RSMs in living cells.[92]
tween those two sites.[88] Depending on the orientation of these Recombinase-mediated DNA fragment substitution is another
recognizing sites, the outcome of the reaction can be excision or writing strategy. Beta recombinase, derived from bacteriophage
inversion of a segment of DNA[88] (Figure 8A). Recombinases- lambda (𝜆), is capable of mediating homologous recombination
mediated DNA inversion, achieved by placing the two recogniz- between DNA molecules with short regions of homology. Farzad-
ing sites in an inverted manner, has been widely used in data fard and Liu developed a platform for generating ssDNA in vivo
recording. To date, various recombinases have been employed for and achieved in vivo writing at specific genomic DNA addresses
in vivo DNA data storage. For example, Ham et al. utilized two by leveraging beta recombinase[93] (Figure 8E). The successful
orthogonal recombination systems derived from different bacte- writing replaced a stop codon in the kanamycin resistance gene,
ria to create a finite state machine in E. coli capable of encoding rendering the cell resistant to kanamycin. As the information was
state information into DNA[89] (Figure 8B). Bacteriophage Bxb1, written in a population of bacteria, the system was capable of
a phage serine integrase, has also been utilized for data writ- recording analog signals instead of binary signals.[93]
ing in vivo by inducing DNA fragment inversion.[90] More im- CRISPR-Cas systems have revolutionized the field of DNA en-
portantly, the inversion generated by Bxb1 could be reversed by gineering and have also been leveraged for data writing in in vivo
co-expressing an excisionase cofactor, enabling rewritable digital DNA data storage. As an adaptive immune system of bacteria,
information storage in live cells[90] (Figure 8C). However, the data the CRISPR-Cas system protects bacteria by acquiring and inte-
capacity of this writing strategy is limited, and one approach to grating invaders’ DNA into the CRISPR array.[94] Since the in-
expand memory capacity is by using orthogonal recombinases. tegration of foreign DNA into the CRISPR array follows a time-
For example, Yang et al. identified 34 phage integrases[91] and defined manner, with new spacers integrated ahead of older ones,
created 11 memory switches capable of recording 1.375 bytes of the CRISPR-Cas system is an excellent data recording system in
information in living cells using this set of new recombinases.[91] bacteria. Indeed, Cas1-Cas2 has been used for data recording in
In addition to DNA fragment inversion, DNA recombi- bacteria[95] (Figure 9A) and demonstrated the storage of a digi-
nases can also induce DNA fragment excision or substitution tal movie.[33] Furthermore, specific cas1-cas2 complexes with re-
(Figure 8A). Excision occurs when the two recognizing sites verse transcriptase activity have enabled information acquisition
are placed unidirectionally.[88] By carefully arranging orthog- from RNA[96] (Figure 9B).
onal recombinase recognition sites, Roquet et al. developed CRISPR-Cas systems have been adopted as tools to manip-
recombinase-based state machines (RSMs) that integrated both ulate the genomes of eukaryotes due to their capability of in-
inversion and excision as writing strategies (Figure 8D). Using ducing double-stranded breaks (DSBs) in a programmable and

Adv. Mater. 2023, 2307499 2307499 (12 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 9. In vivo information writing using CRISPR/Cas techniques. A) Recording arbitrary DNA sequences into a genomic CRISPR array through Cas1-
Cas2 spacer acquisition. Adapted with permission.[95a] Copyright 2016, American Association for the Advancement of Science. B) Transcriptional memory
via CRISPR spacer acquisition from RNA. Adapted with permission.[96] Copyright 2018, Springer Nature. C) Multi-purpose molecular recording within
synthetic DNA target sites. Adapted with permission.[97a] Copyright 2019, Springer Nature. D) Analog memory device for localized genetic recording
with continuously evolving stgRNA. Adapted with permission.[98a] Copyright 2016, American Association for the Advancement of Science. E) Temporally
ordered analog recording by repeated insertion of random nucleotides at a single locus. Adapted with permission.[99] Copyright 2021, Springer Nature. F)
Durable and analog cellular event recording in mammalian cells using base editors. Adapted with permission.[103] Copyright 2018, American Association
for the Advancement of Science. G) DNA typewriter for time-resolved molecular recording via sequential prime editing. Adapted with permission.[106]
Copyright 2022, Springer Nature.

Adv. Mater. 2023, 2307499 2307499 (13 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

precise fashion. Several groups have demonstrated in vivo infor- as deciphering clonal dynamics in disease and development (lin-
mation recording by putting synthetic recorders containing ar- eage tracing).[97,109]
rays of distinct CRISPR guide RNA target sites into cells and us-
ing Cas9 to generate mutations within the synthetic recorder[97]
(Figure 9C). Perli et al. took a different approach by modify- 4. Challenges of DNA Data Storage
ing the sgRNA-expressing DNA locus to be the recording site
(Figure 9D). They added a PAM immediately downstream of Despite the significant progress that has been made in recent
the 20-nucleotide specificity-determining sequence, converting years, the field of DNA data storage still faces numerous chal-
it into a self-targeting guide RNA (stgRNA). This allowed con- lenges. Addressing these challenges will further improve the
tinuous information recording as the ever-evolving stgRNA re- feasibility and practicality of DNA data storage for large-scale
peatedly directed Cas9 nuclease to its DNA locus, inducing data storage applications. Similar to other emerging technolo-
mutations.[98] By utilizing a stgRNA library, Park et al. developed gies, challenges are present at every stage of DNA data storage.
“DNA clocks” capable of recording temporal information in both
cells and animals.[98b] However, since deletions are more com-
mon than insertions in CRISPR-mediated indels, these recording 4.1. Encoding and Decoding
systems primarily act by erasing DNA. Recently, Loveless et al.
achieved DNA writing in the stgRNA expressing locus by using The encoding and decoding algorithms currently employed in
the TdT enzyme to randomly add nucleotides at Cas9-induced DNA data storage are predominantly adapted from established
DSBs[99] (Figure 9E). file compression and error correction algorithms utilized in other
CRISPR-Cas systems have also been engineered extensively to storage technologies.[15,18] However, due to the distinctive strat-
perform genome editing in dependent of DSBs and DNA repair egy and unique practical challenges inherent in DNA storage,[19]
processes. These techniques include base editing[100] and prime these algorithms are suboptimal and only address partial issues.
editing,[101] both of which have been utilized in in vivo data stor- Consequently, more effort is needed to develop highly efficient
age. DNA base editors consist of a catalytically disabled Cas nu- coding algorithms tailored specifically for DNA data storage.
clease fused to a nucleobase deaminase enzyme, enabling the
generation of precise point mutations in genomic DNA.[100] In
vivo data recording using DNA base editors has been demon- 4.2. Write
strated in various organisms, including bacteria[102] and mam-
malian cells[103] (Figure 9F). The high precision and versatility The primary advantage of storing digital data in DNA sequences
of base editing have facilitated single-nucleotide resolution com- lies in its ultrahigh storage density. Despite considerable ad-
puting and memory in living cells.[104] By linking signals of in- vancements in DNA synthesis technologies, the massive synthe-
terest to the expression of specific guide RNAs, dcas12a base ed- sis of DNA oligo pools continues to pose challenges in terms of
itor has enabled multiplexed in parallel recording of up to four speed and cost. It has been demonstrated that parallel synthe-
signals.[105] sis can now achieve the synthesis of 8.4 million unique oligonu-
Prime editors consist of a programmable nickase fused to cleotide sequences. However, to store tera/zettabytes of data in
a polymerase enzyme and are capable of making virtually any DNA within a 24 hour timeframe, the synthesis of hundreds of
substitution, small insertion, and small deletion in genomic billions of distinct oligonucleotides is required, a scale that ex-
DNA.[101] Prime editors rely on an extended guide RNA that both ceeds the current capacity for parallelization by several orders of
specifies the target site and templates the desired genome edit. magnitude. In addition to speed, the cost of synthesis poses a
Recently, Choi et al. developed a DNA Typewriter, capitalizing on significant challenge for DNA data storage. For instance, with
the prime editing technique[106] (Figure 9G). They constructed a current arrayed technologies, the reported price of DNA syn-
DNA tape comprised of a tandem array of partial CRISPR-Cas9 thesis ranges from 0.01 to 0.1 USD per kb.[39] Consequently,
target sites, where only the first target site was active, while all to store 1 terabyte of digital data with a conversation rate of 1
other sites were inactive due to 3 bp truncation at their 5′ ends. bit/nucleotide, the cost would amount to 80–800 million USD,[7b]
Prime editing at the first site was designed to introduce 5 bp a figure that renders commercial utilization of DNA data storage
insertions, with the first 2 bp used as a recording barcode and unfeasible. The third challenge of DNA writing pertains to the
the last 3 bp used to activate the next target site along the array. trade-off between oligo length and synthesis accuracy. Existing
Using this strategy, they achieved sequential and unidirectional approaches typically utilize DNA strands shorter than 200 nu-
information recording in DNA, successfully demonstrating the cleotides for data storage. This is mainly because the synthesis
recording of thousands of symbols, complex event histories, and accuracy and yield drop significantly beyond this range.[39] Stor-
short text messages.[106] ing information in shorter oligos requires the inclusion of index-
CRISPR-Cas systems not only provide versatile ways to write ing information within each oligo to facilitate data reconstruc-
information in living cells for information recording but also en- tion. Thus, the shorter the oligos, the more indexing is needed,
able data rewriting in vivo owing to their ability to substitute DNA consequently reducing the information density.
fragments. For example, Liu et al. have demonstrated the reliable Despite the relatively faster writing speed and lower de-
target-specific rewriting of digital data stored in live cells.[107] In mand for DNA synthesis compared to DNA sequence-based
addition to storing information in vivo,[106,107] CRISPR-Cas based data storage, the primary limitation of DNA nanotechnology-
in vivo information recording is also capable of recording exoge- based data storage lies in its low data capacity. Currently, main-
nous and endogenous stimuli.[98] cellular transcripts,[108] as well stream DNA nanotechnology-based strategies can only store a

Adv. Mater. 2023, 2307499 2307499 (14 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

limited volume of data, typically ranging from a few bits[54,62,67] pairs, it is crucial to prevent conflicts between the primers and
to approximately 14 kb, as demonstrated by DNA punch card the data payload.[35b] Therefore, for a specific dataset, the de-
systems.[68] Therefore, to realize the practical implementation of sign of primer pairs suitable for random access needs to be tai-
DNA nanotechnology-based data storage, there is a crucial need lored accordingly.[9,35b] Additionally, PCR-based random access
to significantly expand the data capacity by several orders of mag- irreversibly consumes a small fraction of the data during each
nitude. retrieval process.[18,35b] The presence of PCR bias also presents
In vivo DNA data storage is also confronted with the challenge challenges for multiplexed data retrieval.[115] While strategies like
of limited storage capacity. Additionally, the writing speed of in emulsion PCR[35b,36] and the use of specific microcapsules[45]
vivo data storage is comparatively sluggish. For instance, it has can help mitigate these issues, their implementation may pose
been estimated that it requires approximately 6 h for a single obstacles, and they cannot resolve all the challenges associ-
dCas9-RNA complex to locate a target in Escherichia coli.[110] As ated with PCR-based random access. As alternative random ac-
a result, the writing time for in vivo DNA data storage can range cess methods, various physical isolation strategies have been
from hours to potentially days. In fact, most of the in vivo storage developed.[45–47] Although they have proven useful in certain
experiments examined the outcomes 3–24 hours after the com- cases, most of these methods still require the use of specific
mencement of the writing process.[93,103] primers for data recognition,[45–48] which is also influenced by the
availability of unique primers.
Another important aspect of random access is the ability to
4.3. Preservation achieve file preview, a feature common in conventional storage
technologies. In DNA-based data storage, file preview is complex
The long-term stability of DNA is a key factor that contributes and requires sophisticated designs. For instance, Tomek et al.
to its appeal as a medium for data storage. However, regardless achieved file preview for four different image files in the presence
of the preservation methods, DNA degradation or damage does of a randomized, nonspecific 1.5 GB background by manipulat-
occur.[111] In addition to the potential loss of stored information, ing PCR stringency and employing unique data organization.[116]
other factors such as long-term storage costs and the impact of As this system relies on the use of specific PCR primers, it also
loading density should also be taken into account. As mentioned faces the problem of primer crosstalk with other primers or the
above, dehydration has been utilized as a preservation strategy for data payload. Indeed, the authors did encounter problematic se-
DNA data. While it offers convenience in preparation, the half- quences within the data payload region and acknowledged that
life of DNA under dehydrated conditions at room temperature is primer crosstalk would be more problematic under more promis-
relatively short compared to other preservation strategies.[111] On cuous conditions.[116]
the other hand, maintaining optimal humidity levels is necessary In DNA nanotechnology-based data storage, random access is
for maximum DNA dehydration protection,[111] which entails en- currently less urgent due to the limited data capacity of these
ergy consumption and increased costs. In contrast, preserving techniques, which has received less attention in research. How-
DNA data by DNA encapsulation requires less involvement in ever, advancements in these fields are expected to increase the
environmental control. However, DNA encapsulation has limita- data capacity, thereby making random access an appealing issue.
tions including a significant decrease in data loading density and Nevertheless, the capability of random access in DNA data stor-
a complex implementation procedure.[28] age needs improvement to support the feasibility of commercial
Preserving data within living organisms is currently limited DNA data storage.
to DNA sequence-based data storage and is not compatible with
DNA nanotechnology-based data storage. Although it has been
demonstrated that data stored in cells can be preserved even 4.5. Read
after cell death,[89] and freeze-drying cells can maintain DNA
integrity,[112] the failure to maintain cell viability abrogates the Data storage in DNA sequence can only be accessed through
significant advantage of preserving DNA data in living organ- DNA sequencing. Due to the low throughput, the first-
isms, which is the potential for high-fidelity data amplifica- generation sequencing is not suitable for reading large-scale
tion. While certain bacteria, particularly bacterial spores, can be datasets.[49] Second-generation sequencing technologies, partic-
preserved by drying them on sterile filter paper,[113] the most ularly Illumina sequencing, have been widely adopted for data
common long-term preservation methods for most bacteria and retrieval.[14,15] Driven by the high demand for genome sequenc-
mammalian cells still require specific equipment and tempera- ing in life sciences, the field of DNA sequencing has made signif-
ture control,[113,114] resulting in increased costs. Overall, despite icant progress, resulting in substantial cost reductions. Illumina
the development of various preservation methods for DNA data achieved the milestone of a $1000 genome in 2014 and continues
storage, each approach still possesses its own limitations that to reduce costs.[117] However, for DNA storage to become com-
must be overcome in the future. mercially viable, the goal is to achieve a $1 genome. The speed of
DNA sequencing is also a limiting factor for the commercializa-
tion of DNA data storage. It has been estimated that to make DNA
4.4. Random Access data storage competitive with mainstream cloud archival storage
systems, a 2–3 orders of magnitude improvement in terms of the
When it comes to large-scale data storage, a major limitation sequencing throughput is needed.[7a]
of PCR-based random access is the requirement for orthogo- Nanopore sequencing, a promising third-generation DNA se-
nal primer pairs. Apart from avoiding crosstalk among primer quencing technique, has also been employed for DNA data

Adv. Mater. 2023, 2307499 2307499 (15 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

retrieval.[35b] It offers several advantages such as long sequenc- cleavage site, thereby avoiding the creation of scars (conserved
ing reads, high sequencing speeds, and the use of smaller nucleotides) commonly observed with other DNA enzymes.[119b,c]
and often portable equipment.[49b,118] However, the accuracy of While these methods still rely on the in vitro synthesis of de-
nanopore sequencing is currently lower compared to Illumina sired DNA sequences through chemical synthesis to construct
sequencing.[118] Additionally, due to limitations in DNA synthe- the template, after sequence verification, ssDNA can be produced
sis, digital data are typically stored in DNA oligos less than 200 nt at milligram-to-gram scale with >98.5% purity and lengths of up
in length.[5] To sequence these data using nanopore technology, to 7000 nucleotides.[119b] This represents a significant improve-
short DNA strands must be concatenated into longer strands.[51] ment compared to large-scale DNA production by chemical syn-
This additional step makes nanopore sequencing less appeal- thesis, where the purity of 51-nucleotide ssDNA is estimated to
ing in data retrieval at present. However, it is foreseeable that be around 70%, and purity decreases significantly with increas-
advancements in DNA synthesis technologies will enable the ing length.[39,119b] Leveraging this system, Zhang et al. success-
storage of digital data in longer DNA strands, thereby making fully demonstrated the storage of a 6.7 KB image in six long
nanopore sequencing more attractive. sequences that can be massively generated.[119b] Moreover, the
The readout strategies for DNA nanotechnology-based data ability to generate diverse ssDNA for constructing FNAs[119a,c]
storage are relatively diverse.[55,57] However, many characteriza- holds great potential in DNA nanotechnology-based data stor-
tion methods, such as AFM[55] and DNA-PAINT,[57] require ex- age, as it can significantly reduce the cost of generating encoding
pensive equipment, limiting their widespread application. In blocks.
contrast, nanopore sensing is a feasible readout methodology that
has been utilized for data retrieval in DNA nanotechnology-based
storage. A limitation of nanopore sensing is that the nanopore 5.2. Enzyme Evolution
used for different storage methods needs to be customized, in-
cluding its size, to achieve accurate data reading.[56,62,63] Enzymes play a crucial role in various processes involved in DNA
data storage, including writing (e.g., TdT, DNA restriction en-
zymes, DNA recombinases, CRISPR-Cas nucleases), random ac-
5. Promising Biotechnologies for DNA Data cess (DNA polymerase), and reading (DNA polymerase). There-
Storage fore, techniques that can enhance the performance of these en-
zymes hold immense potential in advancing DNA data storage.
The field of biotechnology has witnessed remarkable advance- Direct evolution has proven to be an effective strategy to engineer
ments over the past few decades, leading to groundbreaking enzymes with desired properties or improved functionality.[120]
discoveries and transformative applications across various disci- While the methodologies for direct enzyme evolution may
plines. As the toolbox continues to expand, certain techniques vary, the underlying principle involves iterative cycles of gener-
have emerged as promising solutions to address the challenges ating genetic diversity and selecting protein variants with de-
faced in DNA data storage, even though they were originally de- sired properties.[120a] In early-generation protein evolution strate-
veloped for other purposes. In this section, we will describe some gies, these steps were often performed separately, making the
of these promising techniques and their potential applications in entire process time-consuming and labor intensive. The advent
DNA data storage. of continuous direct evolution methods, where all steps of the
evolution cycle are performed continuously without manual in-
tervention, has significantly increased the efficiency of protein
5.1. In Vitro Production of DNA with Large Quantity and High evolution.[120] Among those continuous evolution techniques,
Purity phage-assisted continuous evolution (PACE) has emerged as
one of the most successful methods used for evolving various
Although chemical synthesis remains the predominant method proteins[121] (Figure 10C). In the PACE system, the desired func-
for de novo DNA production, various alternative technologies tion of a protein of interest is linked to the life cycle of phages,
have emerged to produce DNA at scale and low cost.[39] Among enabling rapid protein evolution (approximately 200 rounds of
these alternatives, bacteriophage-derived biotechnological syn- protein evolution over an 8 d period).[121] Therefore, PACE holds
thesis has shown promise in the large-scale production of high- promise as a technique to evolve enzymes for DNA data storage
purity, unique ssDNA at a low cost.[119] For example, Ducani et al. by carefully designing the linkage between the desired properties
demonstrated the stoichiometric production of high-quality ss- and the phage life cycle.
DNA by incorporating hairpin regions encoding recognition se- The successful evolution of Cas9 serves as compelling evi-
quences for a restriction enzyme, amplifying phagemid DNA in dence for the feasibility of this concept. Cas9 has been evolved to
E. coli, and subsequently releasing the ssDNA through enzyme generate variants with varied protospacer-adjacent motif (PAM)
digestion.[119a] Furthermore, the production of ssDNA molecules compatibilities (Figure 10D), overcoming the limitations im-
of varying lengths can be utilized as staples or scaffolds to fold posed by PAM restrictions.[122] Considering the demonstrated
DNA origami structures[119a,c] (Figure 10A). More recently, Zhang utility of Cas9 in in vivo DNA data storage, the reduction of PAM
et al. developed a technique enabling the mass production of ss- sequence restrictions holds significant promise for improving in
DNA with 100% arbitrary sequence, independent of restriction vivo data writing. Similarly, it is anticipated that enzyme evolu-
enzyme digestion[119b] (Figure 10B). This was achieved through tion will offer potential benefits to all processes involved in the
the identification of a pair of self-hydrolyzing DNA enzymes with use of specific enzymes in DNA data storage in the future. For
minimal sequence requirements upstream or downstream of the instance, the evolution of the TdT enzyme with improved activity

Adv. Mater. 2023, 2307499 2307499 (16 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Figure 10. Prospective techniques for DNA data storage. A) Scalable and cost-efficient production of ssDNA using bacteriophages for the genera-
tion of DNA origami nanorods. Adapted with permission.[119c] Copyright 2017, Springer Nature. B) Generation of ssDNA with arbitrary sequences
through paired-end cutting assisted by DNA enzymes (PECAN). Adapted with permission.[119b] Copyright 2022, John Wiley and Sons. C) Overview
of phage-assisted continuous evolution (PACE) for protein evolution. Adapted with permission.[120a] Copyright 2015, Springer Nature. D) Evolution
of Cas9 variants compatible with various PAMs. Adapted with permission.[122a] Copyright 2020, Springer Nature. E) De novo design of enzymes from
scratch. Adapted with permission.[123] Copyright 2022, Springer Nature. F) Mirror-image DNA polymerase and DNA. Adapted with permission.[125b,126a]
Copyright 2021, Springer Nature; Copyright 2016, Springer Nature. G) Strategies for nucleic acid enrichment in targeted sequencing. Adapted with
permission.[130] Copyright 2021, Elsevier. H) Adaptive sampling in nanopore technology for selective sequencing of DNA of interest.

and/or substrate selectivity will undoubtedly provide significant evolved to generate efficient enzymes with the desired activity.[123]
benefits to the writing process of DNA data storage. Significant progress has been made in this field over the past
decades, leading to the successful creation of various artificial
enzymes.[123,124] This rapid advancement suggests that enzymes
5.3. Protein Design suitable for DNA data storage may indeed be designed and cre-
ated to fulfill industrial requirements.
In addition to top-down engineering of natural enzymes through
enzyme evolution to enhance their activity, the de novo design
of enzymes from scratch using a bottom-up approach is emerg- 5.4. Mirror-Image DNA
ing as a promising method for creating enzymes with desir-
able activity.[123] In a typical process of de novo enzyme de- L-DNA, as an enantiomer of naturally occurring D-DNA, ex-
sign, active sites are computationally designed based on desired hibits identical physical and chemical properties to its na-
catalytic activities and then docked into structurally character- tive D-counterpart. However, L-DNA is completely orthogo-
ized protein scaffolds (Figure 10E). Subsequent optimization us- nal to the stereospecific environment of native biology, mak-
ing various computational tools and algorithms further refines ing it highly resistant to biodegradation and contamination.[125]
these novel proteins, which are then experimentally validated and Therefore, L-DNA holds tremendous potential for DNA data

Adv. Mater. 2023, 2307499 2307499 (17 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

storage as it significantly enhances data durability while preserv- unique features inherent in nanopore sequencing, including
ing all the advantages of DNA data storage, including high in- real-time data analysis and the ability to reverse the driving volt-
formation capacity.[125b] Recent advancements in the field have age across the pores to reject sequencing fragments.[118] By using
made data storage in L-DNA increasingly feasible. L-DNA se- specific algorithms, individual DNA fragments passing through
quences of up to 150 nucleotides in length can be chemically the nanopores are analyzed and determined whether to be fur-
synthesized.[125b] Moreover, the development of chirally inverted ther sequenced based on a small initial portion that has already
versions of the enzymes, particularly various mirror-image DNA been sequenced.[132a] This feature, known as adaptive sampling
polymerases,[125b,126] has enabled the engineering of L-DNA[127] or “Read Until”[132a] (Figure 10H), has been demonstrated in
(Figure 10F). A noteworthy accomplishment is the chemical syn- numerous studies.[133] Moreover, with the incorporation of ad-
thesis of the hyper-thermostable and high-fidelity mirror-image vanced algorithmic frameworks, software, and high-performance
Pfu DNA polymerase (Figure 10F), which can be utilized sim- processors, adaptive sampling in nanopore sequencing has
ilarly to its D-counterparts for L-DNA amplification and large evolved towards a more intelligent direction, enabling dynamic
fragment assembly.[125b] Leveraging this enzyme, Fan et al. suc- decision-making based on information obtained from previously
cessfully encoded digital information in 11 double-stranded L- sequenced fragments.[132b] Thus, nanopore adaptive sampling
DNA segments, each consisting of 220 base pairs, which could techniques undoubtedly provide an effective and straightforward
be read out using L-DNA sequencing.[125b,128] In addition to en- method for achieving random access in DNA data reading, de-
hancing the durability of digital information, L-DNA also offers serving significant attention in future research investigations.
a unique avenue for steganography, thereby bolstering informa-
tion security.[125b]
L-DNA has also been employed in DNA-PAINT to enhance 5.7. Plasmonic Photothermal PCR
imaging performance as it does not hybridize to natural D-
DNA.[129] Considering that DNA-PAINT has demonstrated its As mentioned previously, PCR is a pivotal technique in DNA stor-
value as a data reading strategy in DNA nanotechnology-based age that plays an important role in various processes such as
storage, the utilization of L-DNA in this context holds significant data copy and retrieval. However, the long thermocycling time
promise. associated with conventional PCR significantly impedes the effi-
ciency of these procedures. Ultra-fast photonic PCR, harnessing
the photothermal properties of specific nanomaterials, has show-
5.5. Targeted Sequencing Techniques cased its capacity to substantially reduce PCR reaction time.[134]
Photothermal materials, particularly gold nanoparticles and gold
The ability to selectively sequence specific regions of the genome nanofilms, improve the thermocycling rate by inducing volumet-
has garnered significant attention in the field of life science as it ric heating and enhancing heat transfer within the solution.[134]
provides an effective and cost-efficient way for genetic studies.[130] It has been demonstrated that plasmonic PCR employing gold
It is also relevant to DNA data storage as targeted sequencing nanorods can complete 30 PCR cycles in just 54 seconds.[135]
of specific subsets of DNA aligns with the concept of random Of greater significance, this shortened extension time aids in
access. Therefore, techniques that have been developed to en- preventing imperfect primer binding, thereby enhancing PCR
rich specific DNA regions in targeted sequencing (Figure 10G) specificity.[135] In addition to gold nanomaterials, other nano-
can potentially be adapted for random access in DNA data stor- materials, including magnetic nanoparticles and carbon-based
age with appropriate modifications. Indeed, two well-established nanomaterials, also exhibit potential for developing ultrafast PCR
strategies for DNA enrichment, namely PCR and hybridization techniques.[134] The application of these innovations in DNA data
capture,[130] have already been utilized to achieve random access storage undoubtedly holds the potential to elevate overall data
in DNA data storage. The emergence of CRISPR/Cas as an intrin- storage performance.
sically programmable nucleic acid targeting system has further
advanced the field of targeted sequencing and led to the develop-
ment of various more efficient and convenient target enrichment 6. Conclusions and Perspectives
strategies.[130] For example, Lee et al. introduced the CRISPR-Cap
system, which enabled simple and scalable enrichment of target DNA has emerged as one of the most promising materials for
regions in the genome without significant bias.[131] By reasonably next-generation data storage, holding the potential to address the
modifying these CRISPR-based target enrichment strategies to daunting challenge of managing and storing the exponential vol-
suit the specific characteristics of DNA used for data storage, it umes of digital information generated globally. The field of DNA
is theoretically possible to enhance the performance of random data storage has made significant strides in recent years, pro-
access in DNA data storage. pelled by the rapid advancement of various technologies. To date,
tremendous progress has already been achieved, with the abil-
ity to easily store digital data at the megabyte scale within DNA
5.6. Adaptive Sampling in Nanopore Sequencing sequences.[35b] Moreover, digital data storage based on DNA nan-
otechnology has emerged as a promising alternative strategy, of-
In addition to above mentioned targeted sequencing strategies, fering distinct advantages including in-memory computation ca-
nanopore sequencing is another promising method that enables pabilities. In the realm of in vivo DNA data storage, various strate-
selective sequencing with minimal requirements of specialized gies have been developed to achieve intelligent data recording,
library preparation procedures.[132] This is attributed to several leading to novel insights into biological processes.

Adv. Mater. 2023, 2307499 2307499 (18 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Despite the significant achievements over the past decade, the sequencing on a single electrode, thereby demonstrating inte-
field of DNA data storage still faces many challenges. Among grated digital data storage.[136] The retrieval of desired data, in-
them, the most significant issue is that, when compared to exist- cluding enhanced capabilities for previewing and searching spe-
ing storage technologies, DNA storage falls considerably short in cific files, should also be improved to meet user needs effectively.
terms of cost, speed, and random-access capability. These draw- In terms of large-scale computation on DNA data, the integration
backs hinder the practical commercial use of DNA data storage of DNA storage and DNA computing holds tremendous promise.
and restrict its application primarily to archival data storage. An- For instance, Wang et al. have demonstrated the development
other challenge in DNA data storage is the lack of automation. of DNA stand displacement-based software for parallel molec-
Unlike traditional storage methods, the various steps involved ular computation on digital data stored in DNA.[72] Notably, Lv
in DNA data storage are relatively disjointed. For instance, in et al. have showcased that, through careful design, DNA strand
DNA sequence-based data storage, digital information is written displacement reaction are well-suited for general-purpose DNA
using DNA synthesis, followed by preservation under appropri- computing.[137] Those efforts lay the foundation for the develop-
ate conditions. To retrieve the data, the nucleotide composition ment of appropriate software to seamlessly integrate DNA com-
is characterized using sequencing platforms and then decoded putation with DNA storage. Additionally, it is important to recog-
to obtain the original information. The transition of materials nize that instrumental performance requirements in DNA data
among the instruments involved in different steps still requires storage differ from those in biological applications. Therefore,
human involvement, at least in the current stage. Moreover, the more efforts should be devoted to optimizing techniques specifi-
requirement for sophisticated and costly instruments, which are cally tailored to the needs of DNA data storage. Furthermore, im-
typically only available in larger companies and laboratories, fur- proving scalability and portability are vital aspects for achieving
ther restricts the widespread application of DNA data storage in widespread application of DNA data storage in everyday life.
everyday life. Furthermore, conducting large-scale computations
on data stored in DNA remains a formidable challenge. The ab-
sence of compatible “software” necessitates a process involving Acknowledgements
the DNA sequencing, in silico computation, and subsequent syn- This work was supported by the National Key R&D Program of
thesis of new DNA oligos—an approach that is both costly and China (2021YFF1200300); National Natural Science Foundation of China
time-intensive. Overall, there remain significant challenges in (T2188102, 22025404); Innovative Research Team of High-Level Local
DNA data storage that need to be addressed to make DNA a prac- Universities in Shanghai (SHSMU-ZLCX20212602); Shanghai Municipal
tical next-generation storage medium. Health Commission (2022JC027).
Fortunately, numerous existing techniques show great
promise in overcoming the challenges encountered in DNA Conflict of Interest
data storage. Among them, FNAs provide a unique platform to
potentially address the challenges associated with data writing The authors declare no conflict of interest.
in both DNA sequence-based and DNA nanotechnology-based
data storage, owing to their remarkable addressability and ability
to form arbitrary structures. In addition to FNAs, various other
Keywords
biotechnologies can also be applied to DNA data storage. The data storage, DNA, DNA nanotechnology, framework nucleic acids
large-scale generation of high-purity ssDNA will significantly
reduce costs in DNA data storage. Enzyme evolution and de novo Received: July 27, 2023
enzyme design techniques will benefit all steps involving the Revised: October 1, 2023
use of specific enzymes in DNA data storage. The introduction Published online:
of mirror-image systems will not only enhance data preservation
but also contribute to steganography. Promising solutions for
random access include targeted sequencing and nanopore adap-
[1] D. R.-J. G.-J. Rydning, J. Reinsel, J. Gantz, The Digitization of the World
tive sampling. On the other hand, the challenges faced by each
from Edge to Core, Vol. 16, International Data Corporation, Framing-
step in DNA data storage are interconnected, and solving one ham, MA 2018.
problem may have a cascading effect on another problem. For [2] W. A. Bhat, Computer 2018, 51, 54.
instance, the ability to synthesize longer DNA sequences with [3] A. Anzel, D. Heider, G. Hattab, Comput. Struct. Biotechnol. J. 2021,
higher accuracy and lower cost at high throughput will not only 19, 4904.
increase writing speed and reduce costs but also enhance data [4] J. Rothenberg, Avoiding Technological Quicksand: Finding a Viable
density. Technical Foundation for Digital Preservation. A Report to the Council
Moving forward, the field of DNA data storage needs to on Library and Information Resources, Council on Library and Infor-
progress towards a more automated and intelligent direction in mation Resources, Washington, DC 1999.
addition to overcoming current challenges. Through the integra- [5] Y. Hao, Q. Li, C. Fan, F. Wang, Small Struct. 2021, 2, 2000046.
[6] P. Forterre, J. Filée, H. Myllykallio, in The genetic code and the origin
tion of different instruments used in each step or developing de-
of life, (Ed: L. Ribas de Pouplana), Springer, New York, 2004, 10.
vices capable of meeting the multistep requirements for DNA [7] a) L. Ceze, J. Nivala, K. Strauss, Nat. Rev. Genet. 2019, 20, 456;
storage, DNA data storage should be able to store digital infor- b) A. Doricchi, C. M. Platnich, A. Gimpel, F. Horn, M. Earle, G.
mation with minimal human involvement in the intermediate Lanzavecchia, A. L. Cortajarena, L. M. Liz-Marzán, N. Liu, R. Heckel,
steps, akin to current storage strategies. For example, Xu et al. R. N. Grass, R. Krahne, U. F. Keyser, D. Garoli, ACS Nano 2022, 16,
have pioneered a system that combines DNA synthesis and DNA 17552.

Adv. Mater. 2023, 2307499 2307499 (19 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

[8] a) R. N. Grass, R. Heckel, M. Puddu, D. Paunescu, W. J. Stark, Angew. 242; c) M. Yamamoto, S. Kashiwamura, A. Ohuchi, M. Furukawa,
Chem., Int. Ed. 2015, 54, 2552; b) M. Hofreiter, Nucleic Acids Res. Nat. Comput. 2008, 7, 335.
2001, 29, 4793. [36] K. J. Tomek, K. Volkel, A. Simpson, A. G. Hass, E. W. Indermaur, J.
[9] L. Organick, Y.-J. Chen, S. Dumas Ang, R. Lopez, X. Liu, K. Strauss, M. Tuck, A. J. Keung, ACS Synth. Biol. 2019, 8, 1241.
L. Ceze, Nat. Commun. 2020, 11, 616. [37] a) Y. Ren, Y. Zhang, Y. Liu, Q. Wu, J. Su, F. Wang, D. Chen, C. Fan, K.
[10] a) N. Wiener, US News & World Report 1964, 56, 84; b) M. Neiman, Liu, H. Zhang, Small Methods 2022, 6, 2101335; b) Y. Choi, T. Ryu,
Radiotekhnika 1965, 6, 8. A. C. Lee, H. Choi, H. Lee, J. Park, S.-H. Song, S. Kim, H. Kim, W.
[11] J. Davis, Art J. 1996, 55, 70. Park, S. Kwon, Sci. Rep. 2019, 9, 6582.
[12] S. Kosuri, G. M. Church, Nat. Methods 2014, 11, 499. [38] S. Hoshika, N. A. Leal, M.-J. Kim, M.-S. Kim, N. B. Karalkar, H.-
[13] A. M. Giani, G. R. Gallo, L. Gianfranceschi, G. Formenti, Comput. J. Kim, A. M. Bates, N. E. Watkins, H. A. Santalucia, A. J. Meyer,
Struct. Biotechnol. J 2020, 18, 9. S. Dasgupta, J. A. Piccirilli, A. D. Ellington, J. Santalucia, M. M.
[14] G. M. Church, Y. Gao, S. Kosuri, Science 2012, 337, 1628. Georgiadis, S. A. Benner, Science 2019, 363, 884.
[15] N. Goldman, P. Bertone, S. Chen, C. Dessimoz, E. M. Leproust, B. [39] A. Hoose, R. Vellacott, M. Storch, P. S. Freemont, M. G. Ryadnov,
Sipos, E. Birney, Nature 2013, 494, 77. Nat. Rev. Chem. 2023, 7, 144.
[16] L. C. Meiser, P. L. Antkowiak, J. Koch, W. D. Chen, A. X. Kohll, W. J. [40] a) H. H. Lee, R. Kalhor, N. Goela, J. Bolot, G. M. Church, Nat.
Stark, R. Heckel, R. N. Grass, Nat. Protoc. 2020, 15, 86. Commun. 2019, 10, 2383; b) H. Lee, D. J. Wiegand, K. Griswold,
[17] Z. Ping, S. Chen, G. Zhou, X. Huang, S. J. Zhu, H. Zhang, H. H. Lee, S. Punthambaker, H. Chun, R. E. Kohman, G. M. Church, Nat. Com-
Z. Lan, J. Cui, T. Chen, W. Zhang, H. Yang, X. Xu, G. M. Church, Y. mun. 2020, 11, 5246.
Shen, Nat. Comput. Sci. 2022, 2, 234. [41] N. Bhan, A. Callisto, J. Strutz, J. Glaser, R. Kalhor, E. S. Boyden, G.
[18] Y. Erlich, D. Zielinski, Science 2017, 355, 950. Church, K. Kording, K. E. J. Tyo, J. Am. Chem. Soc. 2021, 143, 16630.
[19] M. G. T. A. Rutten, F. W. Vaandrager, J. A. A. W. Elemans, R. J. M. [42] a) S. Palluk, D. H. Arlow, T. De Rond, S. Barthel, J. S. Kang, R. Bector,
Nolte, Nat. Rev. Chem. 2018, 2, 365. H. M. Baghdassarian, A. N. Truong, P. W. Kim, A. K. Singh, N. J.
[20] a) J. Bonnet, M. Colotte, D. Coudy, V. Couallier, J. Portier, B. Morin, Hillson, J. D. Keasling, Nat. Biotechnol. 2018, 36, 645; b) X. Lu, J. Li,
S. Tuffet, Nucleic Acids Res. 2010, 38, 1531; b) L. Organick, B. H. C. Li, Q. Lou, K. Peng, B. Cai, Y. Liu, Y. Yao, L. Lu, Z. Tian, H. Ma, W.
Nguyen, R. Mcamis, W. D. Chen, A. X. Kohll, S. D. Ang, R. N. Grass, Wang, J. Cheng, X. Guo, H. Jiang, Y. Ma, ACS Catal. 2022, 12, 2988.
L. Ceze, K. Strauss, Small Methods 2021, 5, 2001094. [43] a) X. Song, S. Shah, J. Reif, Theor. Comput. Sci. 2021, 894, 190; b)
[21] S. Newman, A. P. Stephenson, M. Willsey, B. H. Nguyen, C. N. S. Kashiwamura, M. Yamamoto, A. Kameda, T. Shiba, A. Ohuchi,
Takahashi, K. Strauss, L. Ceze, Nat. Commun. 2019, 10, 1706. presented at DNA Computing: 8th International Workshop on DNA-
[22] a) K. Matange, J. M. Tuck, A. J. Keung, Nat. Commun. 2021, 12, 1358; Based Computers, DNA8 Sapporo, Japan, June 10–13, 2002 Revised
b) B. Zhu, Y. Zhao, J. Dai, J. Wang, S. Xing, L. Guo, N. Chen, X. Qu, Papers 8, 2003; c) B. Lau, S. Chandak, S. Roy, K. Tatwawadi, M.
L. Li, J. Shen, ACS Appl. Mater. Interfaces 2017, 9, 18434. Wootters, T. Weissman, H. P. Ji, Sci. Rep. 2023, 13, 8514.
[23] a) M. A. Murphy, L. P. Waits, K. C. Kendall, S. K. Wasser, J. A. [44] K. N. Lin, K. Volkel, J. M. Tuck, A. J. Keung, Nat. Commun. 2020, 11,
Higbee, R. Bogden, Conserv. Genet. 2002, 3, 435; b) M. Srinivasan, 2981.
D. Sedmak, S. Jewell, Am. J. Pathol. 2002, 161, 1961. [45] B. W. Bögels, B. H. Nguyen, D. Ward, L. Gascoigne, D. P. Schrijver,
[24] a) A. G. Kansagara, H. E. McMahon, M. E. Hogan, Nat. Methods A.-M. M. Pistikou, A. Joesaar, S. Yang, I. K. Voets, W. J. Mulder, Nat.
2008, 5, 4; b) N. V. Ivanova, M. L. Kuzmina, Mol. Ecol. Resour. 2013, Nanotechnol. 2023, 18, 912.
13, 890. [46] J. L. Banal, T. R. Shepherd, J. Berleant, H. Huang, M. Reyes, C. M.
[25] D. Paunescu, R. Fuhrer, R. N. Grass, Angew. Chem., Int. Ed. 2013, Ackerman, P. C. Blainey, M. Bathe, Nat. Mater. 2021, 20, 1272.
15, 4269. [47] C. Bee, Y.-J. Chen, M. Queen, D. Ward, X. Liu, L. Organick, G. Seelig,
[26] J. Koch, S. Gantenbein, K. Masania, W. J. Stark, Y. Erlich, R. N. Grass, K. Strauss, L. Ceze, Nat. Commun. 2021, 12, 4764.
Nat. Biotechnol. 2020, 38, 39. [48] A. El-Shaikh, B. Seeger, Sci. Rep. 2023, 13, 7053.
[27] M. Puddu, D. Paunescu, W. J. Stark, R. N. Grass, ACS Nano 2014, 8, [49] a) J. M. Heather, B. Chain, Genomics 2016, 107, 1; b) J. Shendure, S.
2677. Balasubramanian, G. M. Church, W. Gilbert, J. Rogers, J. A. Schloss,
[28] W. D. Chen, A. X. Kohll, B. H. Nguyen, J. Koch, R. Heckel, W. J. R. H. Waterston, Nature 2017, 550, 345.
Stark, L. Ceze, K. Strauss, R. N. Grass, Adv. Funct. Mater. 2019, 29, [50] B. Zakeri, P. A. Carr, T. K. Lu, PLoS One 2016, 11, 0152774.
1901672. [51] R. Lopez, Y.-J. Chen, S. Dumas Ang, S. Yekhanin, K. Makarychev, M.
[29] C. Mao, S. Wang, J. Li, Z. Feng, T. Zhang, R. Wang, C. Fan, X. Jiang, Z. Racz, G. Seelig, K. Strauss, L. Ceze, Nat. Commun. 2019, 10, 2933.
ACS Nano 2023, 17, 2840. [52] T. Hu, N. Chitnis, D. Monos, A. Dinh, Hum. Immunol. 2021, 82, 801.
[30] A. X. Kohll, P. L. Antkowiak, W. D. Chen, B. H. Nguyen, W. J. [53] J.-S. Shin, N. A. Pierce, Nano Lett. 2004, 4, 905.
Stark, L. Ceze, K. Strauss, R. N. Grass, Chem. Commun. 2020, 56, [54] C. Lin, R. Jungmann, A. M. Leifer, C. Li, D. Levner, G. M. Church, W.
3613. M. Shih, P. Yin, Nat. Chem. 2012, 4, 832.
[31] P. C. Wong, K.-K. Wong, H. Foote, Commun. ACM 2003, 46, 95. [55] Y. Zhang, F. Wang, J. Chao, M. Xie, H. Liu, M. Pan, E. Kopperger, X.
[32] M. Hao, H. Qiao, Y. Gao, Z. Wang, X. Qiao, X. Chen, H. Qi, Commun. Liu, Q. Li, J. Shi, L. Wang, J. Hu, L. Wang, F. C. Simmel, C. Fan, Nat.
Biol. 2020, 3, 416. Commun. 2019, 10, 5469.
[33] S. L. Shipman, J. Nivala, J. D. Macklis, G. M. Church, Nature 2017, [56] K. Chen, J. Zhu, F. Boskovic, U. F. Keyser, Nano Lett. 2020, 20, 3754.
547, 345. [57] G. D. Dickinson, G. Md. Mortuza, W. Clay, L. Piantanida, C.
[34] W. Chen, M. Han, J. Zhou, Q. Ge, P. Wang, X. Zhang, S. Zhu, L. Song, M. Green, C. Watson, E. J. Hayden, T. Andersen, W. Kuang, E.
Y. Yuan, Natl. Sci. Rev. 2021, 8, nwab028. Graugnard, R. Zadegan, W. L. Hughes, Nat. Commun. 2021, 12,
[35] a) S. M. H. Tabatabaei Yazdi, Y. Yuan, J. Ma, H. Zhao, O. Milenkovic, 2371.
Sci. Rep. 2015, 5, 14138; b) L. Organick, S. D. Ang, Y.-J. Chen, [58] G. Pardatscher, D. Bracha, S. S. Daube, O. Vonshak, F. C. Simmel,
R. Lopez, S. Yekhanin, K. Makarychev, M. Z. Racz, G. Kamath, P. R. H. Bar-Ziv, R. H. Bar-Ziv, Nat. Nanotechnol. 2016, 11, 1076.
Gopalan, B. Nguyen, C. N. Takahashi, S. Newman, H.-Y. Parker, [59] a) S. Wang, X. Cai, L. Wang, J. Li, Q. Li, X. Zuo, J. Shi, Q. Huang, C.
C. Rashtchian, K. Stewart, G. Gupta, R. Carlson, J. Mulligan, D. Fan, Chem. Sci. 2016, 7, 2722; b) F. Huang, H. Xu, W. Tan, H. Liang,
Carmean, G. Seelig, L. Ceze, K. Strauss, Nat. Biotechnol. 2018, 36, ACS Nano 2014, 8, 6849.

Adv. Mater. 2023, 2307499 2307499 (20 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

[60] H. H. Nguyen, J. Park, S. Hwang, O. h. S. Kwon, C.-S. Lee, Y.-B. Shin, [90] J. Bonnet, P. Subsoontorn, D. Endy, Proc. Natl. Acad. Sci. USA 2012,
T. H. Ha, M. Kim, Sci. Rep. 2018, 8, 337. 109, 8884.
[61] Y. Song, S. Kim, M. J. Heller, X. Huang, Nat. Commun. 2018, 9, 281. [91] L. Yang, A. A. K. Nielsen, J. Fernandez-Rodriguez, C. J. Mcclune, M.
[62] N. A. W. Bell, U. F. Keyser, Nat. Nanotechnol. 2016, 11, 645. T. Laub, T. K. Lu, C. A. Voigt, Nat. Methods 2014, 11, 1261.
[63] K. Chen, J. Kong, J. Zhu, N. Ermann, P. Predki, U. F. Keyser, Nano [92] N. Roquet, A. P. Soleimany, A. C. Ferris, S. Aaronson, T. K. Lu, Science
Lett. 2018, 19, 1210. 2016, 353, aad8559.
[64] J. Zhu, N. Ermann, K. Chen, U. F. Keyser, Small 2021, 17, 2100711. [93] F. Farzadfard, T. K. Lu, Science 2014, 346, 1256272.
[65] Y. Zhang, F. Li, M. Li, X. Mao, X. Jing, X. Liu, Q. Li, J. Li, L. Wang, C. [94] P. Horvath, R. Barrangou, Science 2010, 327, 167.
Fan, X. Zuo, J. Am. Chem. Soc. 2019, 141, 17861. [95] a) S. L. Shipman, J. Nivala, J. D. Macklis, G. M. Church, Science 2016,
[66] K. Halvorsen, W. P. Wong, PLoS ONE 2012, 7, 44212. 353, aaf1175; b) R. U. Sheth, S. S. Yim, F. L. Wu, H. H. Wang, Science
[67] A. R. Chandrasekaran, O. Levchenko, D. S. Patel, M. MacIsaac, K. 2017, 358, 1457.
Halvorsen, Nucleic Acids Res. 2017, 45, 11459. [96] F. Schmidt, M. Y. Cherepkova, R. J. Platt, Nature 2018, 562, 380.
[68] S. K. Tabatabaei, B. Wang, N. B. M. Athreya, B. Enghiad, A. G. [97] a) M. M. Chan, Z. D. Smith, S. Grosswendt, H. Kretzmer, T. M.
Hernandez, C. J. Fields, J.-P. Leburton, D. Soloveichik, H. Zhao, O. Norman, B. Adamson, M. Jost, J. J. Quinn, D. Yang, M. G. Jones, A.
Milenkovic, Nat. Commun. 2020, 11, 1742. Khodaverdian, N. Yosef, A. Meissner, J. S. Weissman, Nature 2019,
[69] J. X. Zhang, J. Z. Fang, W. Duan, L. R. Wu, A. W. Zhang, N. Dalchau, 570, 77; b) A. Mckenna, G. M. Findlay, J. A. Gagnon, M. S. Horwitz,
B. Yordanov, R. Petersen, A. Phillips, D. Y. Zhang, Nat. Chem. 2018, A. F. Schier, J. Shendure, Science 2016, 353, aaf7907.
10, 91. [98] a) S. D. Perli, C. H. Cui, T. K. Lu, Science 2016, 353, aag0511; b) J.
[70] H. Talbot, K. Halvorsen, A. R. Chandrasekaran, ACS Synth. Biol. Park, J. M. Lim, I. Jung, S.-J. Heo, J. Park, Y. Chang, H. K. Kim, D.
2023, 12, 978. Jung, J. i. H. Yu, S. Min, S. Yoon, S.-R. Cho, T. Park, H. H. Kim, Cell
[71] H. Lv, Q. Li, J. Shi, C. Fan, F. Wang, ChemPhysChem 2021, 22, 1151. 2021, 184, 1047.
[72] B. Wang, S. S. Wang, C. Chalk, A. Ellington, D. Soloveichik, Proc. [99] T. B. Loveless, J. H. Grotts, M. W. Schechter, E. Forouzmand, C. K.
Natl. Acad. Sci. USA 2023, 120, 2217330120. Carlson, B. S. Agahi, G. Liang, M. Ficht, B. Liu, X. Xie, C. C. Liu, Nat.
[73] Y. Xin, C. Kielar, S. Zhu, C. Sikeler, X. Xu, C. Möser, G. Grundmeier, Chem. Biol. 2021, 17, 739.
T. Liedl, A. Heuer-Jungemann, D. M. Smith, A. Keller, Small 2020, [100] H. A. Rees, D. R. Liu, Nat. Rev. Genet. 2018, 19, 770.
16, 1905959. [101] P. J. Chen, D. R. Liu, Nat. Rev. Genet. 2023, 24, 161.
[74] F. Boskovic, A. Ohmann, U. F. Keyser, K. Chen, Small Struct. 2021, [102] Z.-P. Zou, B.-C. Ye, ACS Synth. Biol. 2020, 9, 2440.
2, 2000144. [103] W. Tang, D. R. Liu, Science 2018, 360, eaap8992.
[75] Z. Ge, H. Gu, Q. Li, C. Fan, J. Am. Chem. Soc. 2018, 140, 17808. [104] F. Farzadfard, N. Gharaei, Y. Higashikuni, G. Jung, J. Cao, T. K. Lu,
[76] a) S. Dey, C. Fan, K. V. Gothelf, J. Li, C. Lin, L. Liu, N. Liu, M. A. D. Mol. Cell 2019, 75, 769.
Nijenhuis, B. Saccà, F. C. Simmel, H. Yan, P. Zhan, Nat. Rev. Methods [105] H. R. Kempton, K. S. Love, L. Y. Guo, L. S. Qi, Nat. Chem. Biol. 2022,
Primers 2021, 1, 13; b) N. C. Seeman, H. F. Sleiman, Nat. Rev. Mater. 18, 742.
2017, 3, 17068. [106] J. Choi, W. Chen, A. Minkina, F. M. Chardon, C. C. Suiter, S. G.
[77] a) J. Fu, Y. R. Yang, A. Johnson-Buck, M. Liu, Y. Liu, N. G. Walter, N. Regalado, S. Domcke, N. Hamazaki, C. Lee, B. Martin, R. M. Daza,
W. Woodbury, H. Yan, Nat. Nanotechnol. 2014, 9, 531; b) J. Fu, M. J. Shendure, Nature 2022, 608, 98.
Liu, Y. Liu, N. W. Woodbury, H. Yan, J. Am. Chem. Soc. 2012, 134, [107] Y. Liu, Y. Ren, J. Li, F. Wang, F. Wang, C. Ma, D. Chen, X. Jiang, C.
5516. Fan, H. Zhang, K. Liu, Sci. Adv. 2022, 8, eabo7415.
[78] S. Takabayashi, W. P. Klein, C. Onodera, B. Rapp, J. Flores-Estrada, [108] C. Jiao, C. Reckstadt, F. König, C. Homberger, J. Yu, J. Vogel, A. J.
E. Lindau, L. Snowball, J. T. Sam, J. E. Padilla, J. Lee, W. B. Knowlton, Westermann, C. M. Sharma, C. L. Beisel, Nat. Biotechnol. 2023, 41,
E. Graugnard, B. Yurke, W. Kuang, W. L. Hughes, Nanoscale 2014, 6, 1107.
13928. [109] a) J. J. Quinn, M. G. Jones, R. A. Okimoto, S. Nanjo, M. M. Chan, N.
[79] P. W. K. Rothemund, Nature 2006, 440, 297. Yosef, T. G. Bivona, J. S. Weissman, Science 2021, 371, eabc1944; b)
[80] B. H. Nguyen, C. N. Takahashi, G. Gupta, J. A. Smith, R. Rouse, P. D. Yang, M. G. Jones, S. Naranjo, W. M. Rideout, K. H. (J.). Min, R.
Berndt, S. Yekhanin, D. P. Ward, S. D. Ang, P. Garvan, H.-Y. Parker, R. Ho, W. Wu, J. M. Replogle, J. L. Page, J. J. Quinn, F. Horns, X. Qiu,
Carlson, D. Carmean, L. Ceze, K. Strauss, Sci. Adv. 2021, 7, eabi6714. M. Z. Chen, W. A. Freed-Pastor, C. S. Mcginnis, D. M. Patterson, Z.
[81] a) M. Zhao, Y. Chen, K. Wang, Z. Zhang, J. K. Streit, J. A. Fagan, J. Gartner, E. D. Chow, T. G. Bivona, M. M. Chan, N. Yosef, T. Jacks,
J. Tang, M. Zheng, C. Yang, Z. Zhu, W. Sun, Science 2020, 368, J. S. Weissman, Cell 2022, 185, 1905.
878; b) A. Gopinath, C. Thachuk, A. Mitskovets, H. A. Atwater, D. [110] D. L. Jones, P. Leroy, C. Unoson, D. Fange, V. Curic, M. J. Lawson, J.
Kirkpatrick, P. W. K. Rothemund, Science 2021, 371, eabd6179. Elf, Science 2017, 357, 1420.
[82] E. Benson, A. Mohammed, J. Gardell, S. Masich, E. Czeizler, P. [111] X. Tan, L. Ge, T. Zhang, Z. Lu, Russ. Chem. Rev. 2021, 90, 280.
Orponen, B. Högberg, Nature 2015, 523, 441. [112] M. Zhang, H. Oldenhof, B. Sydykov, J. Bigalk, H. Sieme, W. F.
[83] B. Wei, M. Dai, P. Yin, Nature 2012, 485, 623. Wolkers, Sci. Rep. 2017, 7, 6198.
[84] P. Wang, T. A. Meyer, V. Pan, P. K. Dutta, Y. Ke, Chem 2017, 2, 359. [113] R. L. Gherna, C. Reddy, in Methods for General and Molecular Micro-
[85] a) S. Jia, J. Wang, M. Xie, J. Sun, H. Liu, Y. Zhang, J. Chao, J. Li, L. biology (Ed: T. J. Beveridge) ASM Press, Washington, DC 2007, p.
Wang, J. Lin, K. V. Gothelf, C. Fan, Nat. Commun. 2019, 10, 5597; b) 1019.
Z. Chen, C. Liu, F. Cao, J. Ren, X. Qu, Chem. Soc. Rev. 2018, 47, 4017. [114] K. Phelan, K. M. May, Curr. Protoc. Cell Biol. 2015, 66, 1.
[86] X. Liu, F. Zhang, X. Jing, M. Pan, P. Liu, W. Li, B. Zhu, J. Li, H. Chen, [115] Y.-J. Chen, C. N. Takahashi, L. Organick, C. Bee, S. D. Ang, P. Weiss,
L. Wang, J. Lin, Y. Liu, D. Zhao, H. Yan, C. Fan, Nature 2018, 559, B. Peck, G. Seelig, L. Ceze, K. Strauss, Nat. Commun. 2020, 11,
593. 3264.
[87] D. Paunescu, M. Puddu, J. O. B. Soellner, P. R. Stoessel, R. N. Grass, [116] K. J. Tomek, K. Volkel, E. W. Indermaur, J. M. Tuck, A. J. Keung, Nat.
Nat. Protoc. 2013, 8, 2440. Commun. 2021, 12, 3518.
[88] A. Akopian, W. M. Stark, Adv. Genet. 2005, 55, 1. [117] J. Preston, A. VanZeeland, D. A. Peiffer, Innovation at Illumina: The
[89] T. S. Ham, S. K. Lee, J. D. Keasling, A. P. Arkin, PLoS One 2008, 3, Road to the $600 Human Genome, Nature Portfolio, Berlin, accessed
2815. data: October 10, 2023.

Adv. Mater. 2023, 2307499 2307499 (21 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

[118] Y. Wang, Y. Zhao, A. Bollas, Y. Wang, K. F. Au, Nat. Biotechnol. 2021, [127] J. Chen, M. Chen, T. F. Zhu, Nat. Biotechnol. 2022, 40, 1601.
39, 1348. [128] X. Liu, T. F. Zhu, Cell Chem. Biol. 2018, 25, 1151.
[119] a) C. Ducani, C. Kaul, M. Moche, W. M. Shih, B. Högberg, Nat. Meth- [129] H. J. Geertsema, G. Aimola, V. Fabricius, J. P. Fuerste, B. B. Kaufer,
ods 2013, 10, 647; b) Q. Zhang, K. Xia, M. Jiang, Q. Li, W. Chen, M. H. Ewers, Nat. Biotechnol. 2021, 39, 551.
Han, W. Li, R. Ke, F. Wang, Y. Zhao, Y. Liu, C. Fan, H. Gu, Angew. [130] Z. Schultzhaus, Z. Wang, D. Stenger, Biotechnol. Adv. 2021, 46,
Chem., Int. Ed. 2023, 62, 202212011; c) F. Praetorius, B. Kick, K. L. 107672.
Behler, M. N. Honemann, D. Weuster-Botz, H. Dietz, Nature 2017, [131] J. Lee, H. Lim, H. Jang, B. Hwang, J. H. Lee, J. Cho, J. i. H. Lee, D.
552, 84. Bang, D. Bang, Nucleic Acids Res. 2019, 47, 1.
[120] a) M. S. Packer, D. R. Liu, Nat. Rev. Genet. 2015, 16, 379; b) M. S. [132] a) M. Loose, S. Malla, M. Stout, Nat. Methods 2016, 13, 751; b) L.
Morrison, C. J. Podracky, D. R. Liu, Nat. Chem. Biol. 2020, 16, 610. Weilguny, N. De Maio, R. Munro, C. Manser, E. Birney, M. Loose,
[121] K. M. Esvelt, J. C. Carlson, D. R. Liu, Nature 2011, 472, 499. N. Goldman, Nat. Biotechnol. 2023, 41, 1018.
[122] a) S. M. Miller, T. Wang, P. B. Randolph, M. Arbab, M. W. Shen, T. [133] a) I. Stevanovski, S. R. Chintalaphani, H. Gamaarachchi, J. M.
P. Huang, Z. Matuszek, G. A. Newby, H. A. Rees, D. R. Liu, Nat. Ferguson, S. S. Pineda, C. K. Scriba, M. Tchan, V. Fung, K. Ng, A.
Biotechnol. 2020, 38, 471; b) T. P. Huang, Z. J. Heins, S. M. Miller, B. Cortese, H. Houlden, C. Dobson-Stone, L. Fitzpatrick, G. Halliday,
G. Wong, P. A. Balivada, T. Wang, A. S. Khalil, D. R. Liu, Nat. Biotech- G. Ravenscroft, M. R. Davis, N. G. Laing, A. Fellner, M. Kennerson,
nol. 2023, 41, 96. K. R. Kumar, I. W. Deveson, Sci. Adv. 2022, 8, eabm5386; b) A.
[123] S. L. Lovelock, R. Crawshaw, S. Basler, C. Levy, D. Baker, D. Hilvert, Patel, H. Dogan, A. Payne, E. Krause, P. Sievers, N. Schoebe, D.
A. P. Green, Nature 2022, 606, 49. Schrimpf, C. Blume, D. Stichel, N. Holmes, P. Euskirchen, J. Hench,
[124] a) A. Quijano-Rubio, H.-W. Yeh, J. Park, H. Lee, R. A. Langan, S. E. S. Frank, V. Rosenstiel-Goidts, M. Ratliff, N. Etminan, A. Unterberg,
Boyken, M. J. Lajoie, L. Cao, C. M. Chow, M. C. Miranda, J. Wi, H. J. C. Dieterich, C. Herold-Mende, S. M. Pfister, W. Wick, M. Loose, A.
Hong, L. Stewart, B.-H. Oh, D. Baker, Nature 2021, 591, 482; b) L. Von Deimling, M. Sill, D. T. W. Jones, M. Schlesner, F. Sahm, Acta
Cao, B. Coventry, I. Goreshnik, B. Huang, W. Sheffler, J. S. Park, K. M. Neuropathol. 2022, 143, 609; c) S. Martin, D. Heavens, Y. Lan, S.
Jude, I. Markovic, R. U. Kadam, K. H. G. Verschueren, K. Verstraete, Horsfield, M. D. Clark, R. M. Leggett, Genome Biol. 2022, 23, 11.
S. T. R. Walsh, N. Bennett, A. Phal, A. Yang, L. Kozodoy, M. Dewitt, [134] M. You, Z. Li, S. Feng, B. Gao, C. Yao, J. Hu, F. Xu, Trends Biotechnol.
L. Picton, L. Miller, E.-M. Strauch, N. D. Debouver, A. Pires, A. K. 2020, 38, 637.
Bera, S. Halabiya, B. Hammerson, W. Yang, S. Bernard, L. Stewart, [135] P. J. R. Roche, M. Najih, S. S. Lee, L. K. Beitel, M. L. Carnevale, M.
I. A. Wilson, H. Ruohola-Baker, et al., Nature 2022, 605, 551. Paliouras, A. G. Kirk, M. A. Trifiro, Analyst 2017, 142, 1746.
[125] a) B. E. Young, N. Kundu, J. T. Sczepanski, Chemistry 2019, 25, 7981; [136] C. Xu, B. Ma, Z. Gao, X. Dong, C. Zhao, H. Liu, Sci. Adv. 2021, 7,
b) C. Fan, Q. Deng, T. F. Zhu, Nat. Biotechnol. 2021, 39, 1548. eabk0100.
[126] a) Z. Wang, W. Xu, L. Liu, T. F. Zhu, Nat. Chem. 2016, 8, 698; b) A. [137] H. Lv, N. Xie, M. Li, M. Dong, C. Sun, Q. Zhang, L. Zhao, J. Li,
Pech, J. Achenbach, M. Jahnz, S. Schülzchen, F. Jarosch, F. Bordusa, X. Zuo, H. Chen, F. Wang, C. Fan, Nature 2023, https://doi.org/10.
S. Klussmann, Nucleic Acids Res. 2017, 45, 3997. 1038/s41586-023-06484-9.

Shaopeng Wang received his Ph.D. degree from Shanghai Institute of Applied Physics (SINAP), Chi-
nese Academy of Sciences (CAS) (2016). He was a postdoctoral fellow at Johns Hopkins University
School of Medicine, USA (2016–2022). Now, he is an associate research fellow at the Institute of
Molecular Medicine, Renji Hospital, School of Medicine, Shanghai Jiao Tong University. His research
interests include DNA nanotechnology, super-resolution imaging, and DNA data storage.

Xiuhai Mao is an associate professor in the Institute of Molecular Medicine at Shanghai Jiao Tong Uni-
versity. He obtained his Ph.D. degree from Shanghai Institute of Applied Physics, Chinese Academy
of Sciences in 2013. His research field covers functional nucleic acids nanostructure and smart
nanomedicine.

Adv. Mater. 2023, 2307499 2307499 (22 of 23) © 2023 Wiley-VCH GmbH
 15214095, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adma.202307499 by Peking University Health, Wiley Online Library on [26/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advmat.de

Fei Wang obtained her B.S. from University of Science and Technology of China (USTC) in 2013. She
obtained her Ph.D. in inorganic chemistry at Shanghai Institute of Applied Physics (SINAP), Chinese
Academy of Sciences (CAS) in 2018. She conducted postdoctoral research in Shanghai Jiao Tong Uni-
versity (SJTU) and joined SJTU as a Tenure Track Associate Professor in 2021. Her research interests
are focused on DNA computing and DNA-based data storage.

Xiaolei Zuo received his Ph.D. degree from Shanghai Institute of Applied Physics, Chinese Academy of
Sciences (2008). He was a postdoctoral fellow at University of California, Santa Barbara, USA (2008–
2010), and at Los Alamos National Laboratory, USA (2010–2012). Now he is a professor of the Insti-
tute of Molecular Medicine, Renji Hospital, School of Medicine, Shanghai Jiao Tong University. His
research interests include biosensors, 3D DNA frameworks, and DNA data storage.

Chunhai Fan is a K. C. Wong Chair Professor, new cornerstone investigator and dean in the School of
Chemistry and Chemical Engineering at Shanghai Jiao Tong University (SJTU), and executive dean of
the National Center for Translational Medicine. Dr. Fan’s research is focused on DNA nanotechnology,
biosensors and bioimaging. He has published more than 600 papers in peer-reviewed journals, and
been recognized as highly cited researchers by Clarivate Analytics since 2014.

Adv. Mater. 2023, 2307499 2307499 (23 of 23) © 2023 Wiley-VCH GmbH