Review

DNA Data Storage

Tomasz Buko † , Nella Tuczko † and Takao Ishikawa *

Department of Molecular Biology, Institute of Biochemistry, Faculty of Biology, University of Warsaw,

Miecznikowa 1, PL-02-096 Warsaw, Poland
* Correspondence: [Link]@[Link]
† These authors contributed equally to this work.

Abstract: The demand for data storage is growing at an unprecedented rate, and current methods
are not sufficient to accommodate such rapid growth due to their cost, space requirements, and
energy consumption. Therefore, there is a need for a new, long-lasting data storage medium with
high capacity, high data density, and high durability against extreme conditions. DNA is one of the
most promising next-generation data carriers, with a storage density of 1019 bits of data per cubic
centimeter, and its three-dimensional structure makes it about eight orders of magnitude denser than
other storage media. DNA amplification during PCR or replication during cell proliferation enables
the quick and inexpensive copying of vast amounts of data. In addition, DNA can possibly endure
millions of years if stored in optimal conditions and dehydrated, making it useful for data storage.
Numerous space experiments on microorganisms have also proven their extraordinary durability in
extreme conditions, which suggests that DNA could be a durable storage medium for data. Despite
some remaining challenges, such as the need to refine methods for the fast and error-free synthesis of
oligonucleotides, DNA is a promising candidate for future data storage.

Keywords: bit; byte; long term data storage; next-generation information storage; oligonucleotide;
sequencing

Key Contribution: The latest achievements in DNA data storage are reviewed and summarized in
simple way so that principles are understandable for biologists without a background in data science.

Citation: Buko, T.; Tuczko, N.;

Ishikawa, T. DNA Data Storage. 1. Introduction
BioTech 2023, 12, 44. The demand for data storage is increasing by approximately 50% every year. In 2012,
[Link] the entire world’s total information storage was 2.7 ZB [1], in 2018 it reached 33 ZB, only to
rise two-fold in 2020. It is estimated that newly created data will take up about 175 ZB by
Academic Editor: Marco Mesiti
2025 [2]. This equals a 65-fold increase only in the period between 2012 and 2025.
The tremendous Global Datasphere expansion is a strong motivator for new develop-
Received: 20 April 2023 ments in data storage. Current data storage methods, such as magnetic (e.g., hard disk),
Revised: 22 May 2023 optical (e.g., Blu-ray disc), and solid-state (e.g., flash drive), are insufficient to accommodate
Accepted: 23 May 2023
such rapid growth [3]. The main problems with those methods are their cost, space, and
Published: 1 June 2023
energy consumption during the recording, storing, and reading of data. Moreover, their
durability reaches a maximum of 50 years in perfectly optimal conditions [4]. Humidity,
extreme temperatures (both high or low), magnetic fields, or mechanical failures are the
Copyright: © 2023 by the authors.
main reasons why those methods are not reliable for long-term data storage.
Licensee MDPI, Basel, Switzerland. Therefore, there is a great demand for a new, longevous data storage medium with a
This article is an open access article high capacity, high data density, and high durability against extreme conditions [1]. There
distributed under the terms and are a few prototypes of next-generation data carriers that may be able to cope with the
conditions of the Creative Commons above-mentioned challenges. Among them, DNA seems to be one of the most promising.
Attribution (CC BY) license (https:// The most distinguishing features of DNA from other storage media are its density and
[Link]/licenses/by/ durability against the extreme conditions.
4.0/).

BioTech 2023, 12, 44. [Link] [Link]

BioTech 2023, 12, 44 2 of 17

Escherichia coli has a storage density of 1019 bits of data per cubic centimeter [5]. This
means that 1.7 × 1019 bits can be stored in just 1 g of DNA. Due to its three-dimensional
structure, DNA is about eight orders of magnitude denser than other storage media.
Moreover, DNA replication during PCR or the cell’s proliferation enables the quick and
inexpensive copying of vast amounts of data [3].
For years, a DNA specimen collected from a 700,000-year-old horse was considered
to be the oldest extracted DNA. However, in 2021, this record was pushed to 1 million
years. DNA extracted from mammoth teeth was successfully extracted and sequenced [6].
Additionally, scientists managed to sequence 300,000-year-old mitochondrial DNA from
humans and bears [4]. These examples perfectly illustrate the longevity of DNA and
proves its usefulness for archeological purposes or data storekeeping. If stored in optimal
conditions and dehydrated, DNA can possibly endure for millions of years [1].
Numerous space experiments on microorganisms have proven their extraordinary
durability in extreme conditions. Due to solar UV radiance, the space vacuum, and extreme
temperature conditions, space is considered one of the most hostile environments [7].
UV radiance being the most deleterious parameter in space increases microorganisms’
lethality by four orders of magnitude in relation to Earth’s conditions [8]. UVB and
UVC altogether cover the 200–315 nm light spectrum; these are the most hazardous to
microorganisms and are responsible for their high lethality in space. This is caused by
high irradiance absorption by DNA and proteins in such spectral ranges. In vegetative
cells, this UV irradiance leads to DNA mutations, such as cyclobutane pyrimidine dimers
and pyrimidine–pyrimidone photo products [9]. Meanwhile, in bacterial spores, thymine
dimer photoproducts, so-called spore photoproducts (SP), are formed due to UV radiation.
Despite this fact, all these dimers can be repaired by the direct reversal mechanism. Spores
possess an additional SP-specific repair pathway that makes spores significantly more
resistant to UV radiance than vegetative cells [10].
Regardless of such hostile conditions, it has been proven that spores of Bacillus subtilis
shielded against UV solar radiation are able to survive in outer space for nearly 6 years.
Although only 1–2% of the population recovered, the outcome was significantly increased
(even to 90% of population recovered) if 5% glucose was added to the spore multilayer.
It was suggested that glucose binds additional water molecules, preventing the cell from
becoming completely desiccated. It also replaces water molecules, thereby stabilizing the
macromolecular structure [8]. Furthermore, some microorganisms can even cope with a
full space environment. For example, the lichens Rhizocarpon geographicum and Xanthoria
elegans survived a 2-week exposure to outer space. After that time, the lichens completely
restored their photosynthetic activity and no ultrastructural changes were revealed in most
of the fungal and algal cells of lichens [11]. It is supposed that their thick cortex with
UV-screening pigments (rhizocarpic and parietin phenolic acids) are responsible for their
survival [12].

2. Coding Files in DNA

Encoding information in DNA is based on binary code. A specific nucleotide corre-
sponds to a code, for example, 00 → A, 01 → C, 10 → G, and 11 → T. While binary data are
“translated” into a DNA sequence, it is important to avoid long homopolymers (more than
three same nucleotides in a row) and unreasonable GC content, as both might generate mis-
takes during the synthesis and sequencing of DNA strings. In fact, encoding a file requires
converting text into a code such as ASCII (Figure 1) or Base64, and then converting the
coded file into a binary system. The encoding field uses different coding algorithms, such
as Huffman, to condense messages and balance code, preventing homopolymer sequences.
Two examples of coding systems, their modifications, and other algorithms of a similar
kind generate proper DNA strings [13,14], which are capable of long data storage.
BioTech 2023, 4, x FOR PEER REVIEW 3
BioTech 2023, 12, 44 3 of 17

Figure [Link]
Figure Anexample of coding
example the message
of coding “ramy”“ramy”
the message into an ASCII
into code. Converting
an ASCII code. binary data binar
Converting
into nucleotide sequences is made by computer algorithms.
into nucleotide sequences is made by computer algorithms.
Church et al. (2012), for the first time, encoded a draft of a book, eleven JPG images
Church
and one et [Link]
JavaScript (2012),infor DNA the[15].
firstFortime,
this encoded a draft
purpose, they usedof a book,
a simple eleven JPG im
encoding
and
methodoneinvolving
JavaScript theprogram
translationin ofDNA [15].AFor
zeros into or Cthis
andpurpose,
ones into Ttheyor [Link] a aresult,
simple enc
the authors received 54,898 oligonucleotides, each containing three
method involving the translation of zeros into A or C and ones into T or G. As a r parts: 96 bases of data,
22-bases-long sequences at both ends, allowing those oligonucleotides to be parallelly
the authors received 54,898 oligonucleotides, each containing three parts: 96 bases of
amplified by PCR, and the 19-bases-long index sequence, pointing out the segment position
22-bases-long
in the original filesequences
[15]. Encodingat both
one ends,
bit perallowing
base allowedthose
the oligonucleotides
authors to avoid sequences to be parallell
plified
that werebypotentially
PCR, and theto 19-bases-long
hard indexinformation
write or read. Splitting sequence,into pointing
blocks ofoutdatathe segment po
allowed
in
thethe original
authors file [15].the
to circumvent Encoding
problemsone bit perwith
associated basetheallowed
synthesisthe authors
of long DNA to avoid sequ
strings.
This pioneering work demonstrated the real possibility of using
that were potentially hard to write or read. Splitting information into blocks of da DNA as a data storage
material, and also showed the enormous capacity of this method. An important element of
lowed the authors to circumvent the problems associated with the synthesis of long
the works of that time was to show the limitations of the method used. Through this work,
strings. This
it was noted pioneering
that the informationwork demonstrated
encoded the real
in DNA is prone possibility
to sequencing of using
errors, mainlyDNA in as a
storage material,
homopolymer regions. and also showed the enormous capacity of this method. An imp
elementOne of
yearthe works
later, of that
Goldman time
et al. was
(2013) to to
tried show the limitations
overcome the sequencing of the method
errors occurring used. Thr
by encoding data with redundancy [16]. The authors encoded
this work, it was noted that the information encoded in DNA is prone to sequenci all 154 of Shakespeare’s
sonnets, a scientific article, a medium-resolution color photograph of the European Bioin-
rors, mainly in homopolymer regions.
formatics Institute, and a 26 s long excerpt from Martin Luther King’s 1963 “I have a
dream”One year using
speech later, the
Goldman
Huffmanetalgorithm
al. (2013)totried
coverttonumeric
overcome datatheintosequencing
a nucleotide errors o
ring by encoding data with redundancy [16]. The authors
sequence [16]. In summary, bytes of binary sequences were converted into base-3 encoded all 154digits
of Shakesp
(or ternary)
sonnets, from 0 to 2,article,
a scientific which were then associated with three
a medium-resolution colornucleotides,
photograph A, T,
ofand
theCEuropean
(or
G if C has been used for the encoding of the previous ternary
formatics Institute, and a 26 s long excerpt from Martin Luther King’s 1963 “I h digit). DNA strings were
divided into 100-nucleotide-long oligos with an overlap of 75 residues between adjacent
dream” speech using the Huffman algorithm to covert numeric data into a nucleoti
fragments, creating four-fold redundancy (Figure 2). Alternate fragments were converted
quence
to their reverseIncomplement,
[16]. summary, which bytesreduces
of binary the sequences
probability ofwere converted
systematic failure,into
suchbase-3
as dig
ternary)
issues withfromDNA 0 to 2, which Indexing
sequencing. were then associated
sequences with three
comprising nucleotides,
17 nucleotides wereA, T, and C
also
encoded at the beginning and end of each fragment.
if C has been used for the encoding of the previous ternary digit). DNA strings we
vided into 100-nucleotide-long oligos with an overlap of 75 residues between ad
fragments, creating four-fold redundancy (Figure 2). Alternate fragments were conv
to their reverse complement, which reduces the probability of systematic failure, su
issues with DNA sequencing. Indexing sequences comprising 17 nucleotides wer
encoded at the beginning and end of each fragment.
BioTech 2023, 12, 44 BioTech 2023, 4, x FOR PEER REVIEW 4 of 17 4 of 18

Figure 2. The coding scheme

Figure 2. Theimplemented
coding schemebyimplemented
Goldman [Link]
Digital information (a) is converted
et al. Digital information to
(a) is converted to
base-3 (b) using a Huffman
base-3 (b) code
usingand is subsequently
a Huffman code and isissubsequently
converted to
is DNA strings
converted (c). Dividing
to DNA strings (c).DNA
Dividing DNA
strings as shown
strings as shown generated generated
four-fold four-fold
redundancy (d).redundancy (d).

Ailenberg(2009)
Ailenberg and Rotstein and Rotstein
encoded (2009)
text,encoded
music, text,
and music,
images and
in images
DNA by in DNA
usingby using
modified Huffman coding (Figure 3) [17]. In their work, they
modified Huffman coding (Figure 3) [17]. In their work, they constructed a plasmids constructed a plasmids li-
brary each containing 10,000 bp of information and an index plasmid that contains basic
library each containing 10,000 bp of information and an index plasmid that contains
information, such as the title, author, plasmid number, and primer assignments used to
basic information, such as the title, author, plasmid number, and primer assignments
read coded information [17]. The authors also constructed a separate encoding table for
used to read coded information [17]. The authors also constructed a separate encoding
each type of file, which allowed the authors to encode each character from the keyboard.
table for each type of file, which
The authors allowed the
also indicated the possibility
authors toofencode each
extending character
their from the
code according to the de-
BioTech 2023, 4, x FOR PEER REVIEW
keyboard. The authors also indicated
scribed rules. the possibility of extending their code according to the 5 of 18
described rules.

Figure 3. An example
[Link] musicofin
An example [Link]
coding Fragment of Fragment
in DNA. “Mary Had a Little
of “Mary Lamb”
Had encoded
a Little Lamb” encoded
using Huffman code. A nucleotide sequence corresponding to the music code is shown in (a) and thein (a) and
using Huffman code. A nucleotide sequence corresponding to the music code is shown
the encryption part in (b). Adapted from Ailenberg and Rotstein [17].
encryption part in (b). Adapted from Ailenberg and Rotstein [17].
The first example of the graphical file recoded in DNA was a simplified lamb draw-
ing (Figure 4). Although this image consists of simple geometric figures, the simplicity
and geometry of the image are not general requirements. Yazdi et al. (2017) managed to
encode The Citizen Kane poster photograph and Smiley Face emoji (Figure 5) [18]. For
 BioTech 2023, 12, 44 5 of 17

The first example of the graphical file recoded in DNA was a simplified lamb drawing
(Figure 4). Although this image consists of simple geometric figures, the simplicity and
geometry of the image are not general requirements. Yazdi et al. (2017) managed to encode
The Citizen Kane poster photograph and Smiley Face emoji (Figure 5) [18]. For this purpose,
they used Base64 encoding to convert files into binary format. The DNA string length used
by the authors was 1000 bp, containing 984 bp of information and 16 bp of address sequence.
The purpose of the addressing method was to enable random access to codewords via
highly selective PCR reactions. This approach allows the specific amplification of a pool of
oligos without amplifying and reading all sequences from a given pool. This work also
presented a new deletion-correcting method called homopolymer check codes. This method
of correction divides DNA sequences into strings of homopolymers, e.g., {AATCCCCGA}
into strings {AA, T, CCC, G, A}, which gives a homopolymer sequence of length {2,1,3,1,1}.
The homopolymer length sequence contains special redundancy that protects against
asymmetric substitution errors. Hypothetically, when two deletions occur in the sequence
resulting in {ATCCGA}, the length of the homopolymer fragments is {1,1,2,1,1}. Recovering
the original sequence is possible by correcting two bounded magnitude errors. Combining
Tech 2023, 4, x FOR PEER REVIEW this with GC content balancing, the subsequent alignment of DNA oligonucleotides, and 6 of
post-sequencing sequence sorting based on the correctness of the index sequence resulted
in a new coding method.

Figure 4. Indication
Figure of elements
4. Indication of elementsofofthethenucleotide sequence
nucleotide sequence in in which
which a Little
a Little LambLamb was encoded a
was encoded
an example image presenting
and an example a lamb
image presenting fromfrom
a lamb thethe“Mary
“Mary Had
HadaaLittle Lamb”rhyme
Little Lamb” rhyme encoded
encoded by by Ail
Ailenberg and Rotstein [17]. The sequence of a file type defines it as an image.
berg and Rotstein [17]. The sequence of a file type defines it as an image. The geometric shape ofThe geometric shape
lamb of the lambthe
enables enables
use the use of 238
of only only bp
238 of
bp DNA
of DNAfor for encoding.
encoding. Encoding
Encoding has been
has performed using
been performed usin
a template of signs indicating the type of shape and its spatial
template of signs indicating the type of shape and its spatial coordinates. coordinates.
 Figure 4. Indication of elements of the nucleotide sequence in which a Little Lamb was encoded a
an example image presenting a lamb from the “Mary Had a Little Lamb” rhyme encoded by Ail
berg and Rotstein [17]. The sequence of a file type defines it as an image. The geometric shape of
BioTech 2023, 12, 44 6 of 17
lamb enables the use of only 238 bp of DNA for encoding. Encoding has been performed usin
template of signs indicating the type of shape and its spatial coordinates.

Figure 5. Smiling emoji and original Citizen Kane poster photograph encoded and decoded by Yazdi
Figure 5. Smiling emoji and original Citizen Kane poster photograph encoded and decoded
et al. [18]. The raw images were encoded and synthesized in the form of DNA strings (a,b). Images
Yazdi et al. [18]. The raw images were encoded and synthesized in the form of DNA strings (a
received after decoding
Imageswithout
receivedhomopolymer
after decodingcheck codes
without during processing
homopolymer (c,d).during
check codes Images received (c,d). Ima
processing
after sequencing received
DNA strings when homopolymer error correction was made in
after sequencing DNA strings when homopolymer error correction was order to reduce
made in order
the number of errors thatthe
reduce occurred
numberduring each
of errors thatencoding and decoding
occurred during step (e,f).
each encoding andTwo errorsstep
decoding in the
(e,f). Two err
in thesufficient
Citizen Kane file were Citizen Kane file were
to make sufficient
the recovery ofto make
the theimpossible.
image recovery ofOne
the image impossible.
error in the emoji One erro
did not influence the
the emoji
imagedid not influence the image quality.
quality.

Coding motionCoding
picturemotion picture
as motion GIFs as and
motion GIFs has
movies andalso
movies
beenhas also been
achieved in achieved
the in
DNA data storage field. In 2017, Shipman et al. encoded five
DNA data storage field. In 2017, Shipman et al. encoded five frames of a galloping mareframes of a galloping m
from Eadweard Muybridge’s “The Human and Animal
from Eadweard Muybridge’s “The Human and Animal Locomotion Photographs” [19]. Locomotion Photographs” [1
In their experiment, CRISPR-Cas was used to integrate an encoded
In their experiment, CRISPR-Cas was used to integrate an encoded short movie into the short movie into
genomes of
genomes of a population ofaliving
population of living
bacteria. bacteria.
The usage The method
of this usage of does
this method does the
not change not change
overall encoding protocol. Strings of DNA are integrated into the CRISPR array thanks
to appropriate integrases. Spacer sequences in the CRISPR array were used to encode
barcodes defining which set of pixels was encoded in a specific part. The use of the
CRISPR method for GIF encoding was of great importance because it allows the encoding
of subsequent sequences without the need to additionally index them. This is because
newly added sequences are almost always integrated in such a way that they push the
previously integrated sequences away from the leader region. Therefore, the order of the
sequence was conditioned by successive transformations in which DNA with encoded
movie frames was introduced to bacterial cells.
A number of other works referring to information encoding in DNA are summarized
in Table 1 below.

Table 1. Works regarding the coding of information on DNA. In “redundancy or error correction”
column, “n.d.” indicates that there is no information in the original work.

Redundancy or
Length of
Authors Data Size Encoding Method Error Modification Reference
Strings
Correction
DNA string
Bornholt et al. 51 KB 120 Huffman code – [20]
exclusive-or
Blawat et al. 22 MB 230 Own bit mapping BCH code – [21]
Organick et al. 200 MB ~150 Base-4 Reed–Solomon – [22]
Choi et al. 854 B 85 Own bit mapping Reed–Solomon Degenerate bases [23]
Enzymatic DNA
Lee et al. 96 B ~50 ASCII codec [24]
synthesis
Enzymatic
Tabatabaei et al. 2 KB; 392 KB 450 Own bit mapping Not needed [25]
nicking (Pf Ago)
BioTech 2023, 12, 44 7 of 17

Table 1. Cont.

Redundancy or
Length of
Authors Data Size Encoding Method Error Modification Reference
Strings
Correction
Yang et al. 23 KB 83 A, C = 0; G, T = 1 n.d. TNA [26]
682 B; 39 KB; 28 Artificial
Ren et al. ~100 RABR; RALR Reed–Solomon [27]
MB nucleotides
ASCII; Elias Epigenetic
Mayer et al. 24,5–33,6 KB ~40 n.d. [28]
gamma encoding

3. Synthesis of DNA Strings

Chemical DNA synthesis has made tremendous progress since the 1970s, when frag-
ments of about 20 nucleotides could be synthesized, to the present, when fragments of up
to 500 nucleotides can be easily made. The technology commonly used for the synthesis of
DNA strands enables only short 200–300 nucleotides sequences to be synthesized, which is
a limitation when coding a large amount of data. Nevertheless, the technology used for
DNA synthesis on microarrays seems to be more suitable for this purpose. It allows the
synthesis of parallel oligonucleotides containing different sequences (Figure 6). By using it,
the time and cost needed for the synthesis of large-scale DNA libraries might be greatly
reduced [29]. Microarrays have enabled the high-fidelity synthesis of oligo pools of about
300 nucleotides in length [30]. Regardless of the synthesis method, long DNA fragments
must be assembled from oligos. It is also necessary to add indexes to each fragment, or
sequence overlapping in successive DNA fragments [3], unless—as discussed above—the
BioTech 2023, 4, x FOR PEER REVIEWCRISPR method is used to record information in the bacterial genome. In 2017, Heckel 8 of 18
et al. considered the storage capacity using both assembly methods and have shown that
an index-based coding system is optimal for data storage purposes [31].

Figure 6.
Figure A solid-phase
6. A solid-phase method
method forfor the
the synthesis
synthesisof ofoligonucleotides
oligonucleotidesusing
usingphotolabile
photolabilecompounds.
compounds.
A spacer containing the photolabile group is covalently joined to the surface. Once Once spots
spots on
on the
surface are exposed to UV light through slits in the physical mask, the photolabile protecting
exposed to UV light through slits in the physical mask, the photolabile protecting group group is
is removed and the synthesis of oligonucleotide begins. The subsequent appropriate
removed and the synthesis of oligonucleotide begins. The subsequent appropriate phosphoramidite phospho-
ramidite with the photolabile
with the photolabile groupapplied
group is then is thento applied to the
the entire entireofsurface
surface of the
the plate. plate.
It can It can
form form
covalent
covalent bonds only in the absence of the preceding photolabile group. In the subsequent
bonds only in the absence of the preceding photolabile group. In the subsequent steps, additional steps,
additional spots are exposed to radiation, and another phosphoramidite is applied where necessary.
spots are exposed to radiation, and another phosphoramidite is applied where necessary. Until the
Until the final oligonucleotide is completely synthesized, the chain-extending processes are re-
final oligonucleotide is completely synthesized, the chain-extending processes are repeated [29].
peated [29].
4. New Storage Medium, Old Problems, and Solutions
4. New Storage Medium, Old Problems, and Solutions
A serious problem with the usage of DNA for data storage purposes is that long-term
storage,serious
A problem
synthesis, with the usage
and sequencing of introduce
might DNA for data
somestorage purposes
errors (such is that long-term
as deletion, insertion,
storage, synthesis, and sequencing might introduce some errors (such as deletion, inser-
tion, or substitution). It should be stressed that errors are not the only issue when DNA is
used as the data storage medium, but this is a problem of all information storage technol-
ogies. This is why there is a solution to it in the form of error-correcting codes (ECCs), in
which a minimal amount of special data is added for error-correction purposes. In classi-
BioTech 2023, 12, 44 8 of 17

or substitution). It should be stressed that errors are not the only issue when DNA is used
as the data storage medium, but this is a problem of all information storage technologies.
This is why there is a solution to it in the form of error-correcting codes (ECCs), in which a
minimal amount of special data is added for error-correction purposes. In classical data-
storage devices, the use of ECCs adds redundancy and allows the correction of essentially
all errors that occur during use. ECCs such as fountain code, rapid tornado code, HEDGES
(Hash Encoded, Decoded by Greedy Exhaustive Search), or the Reed–Solomon code [32] are
used in DNA data storage. In general, ECCs introduce sequence redundancy, which enables
the subsequent recovery of complete data even in the case that some oligonucleotides used
for data storage are physically damaged. The implementation of ECCs slightly diminishes
the storage capacity (because ECCs are often based on adding external fragments to the
sequences encoding data), but its advantages—namely the possibility of error correction—
outweigh this limitation. ECCs enable insertions and deletions to be corrected, as well as
the loss of some parts of the DNA strings. An alternative to ECCs was the previously used
high-depth sequencing, which, for obvious reasons, only corrected sequencing errors.
One of the most frequently mentioned ECCs in the literature is a Reed–Solomon code
(Figure 7). In general, the Reed–Solomon code is based on the transformation of the original
data set to a symbol set. The symbols are then converted to coefficients in a system of linear
BioTech 2023, 4, x FOR PEER REVIEW 9 of 18
equations and their solutions enable the original data set to be accessed. Meiser et al. (2020)
have used a Reed–Solomon code for storing a full album of music in DNA [33].

Figure
Figure 7. Principle
7. Principle of Reed–Solomon
of Reed–Solomon correction:
correction: first,is the
first, the data datainto
divided is divided
parts, andinto
eachparts,
part isand each part is
assigned
assigned x and y values
x and that determine
y values its location.
that determine Based on the
its location. Basedcoordinates, the points are matched
on the coordinates, the points are matched
to the polynomial function P(x), which is used to determine the parity symbols. Parity symbols are
to the polynomial function P(x), which is used to determine the parity symbols. Parity symbols
extra data points that match the original DNA sequence and are stored with the original data. When
are extra
some of the data points
original data that match
are lost, the original
the remaining dataDNApointssequence
and parityand are stored
symbols can be with
used tothe original data.
recreate the original
When some of thepolynomial function
original data are and
lost,receive original data.
the remaining data points and parity symbols can be used to
recreate the original polynomial function and receive original data.
Recently, Xie et al. (2023) conducted an analysis showing the value of the sequencing
depth for retrieving the right string of data [34]. Sufficiently deep sequencing allows the
Recently, Xie et al. (2023) conducted an analysis showing the value of the sequencing
use of MSA (multiple sequence alignment) methods to establish a consensus sequence and
depth for retrieving the right string of data [34]. Sufficiently deep sequencing allows the
correct errors that may appear on the DNA strands. The MAFFT algorithm was chosen
for the analysis, which has been shown to be able to correct more than 95% of errors at a
sequencing depth reaching 100× when the error rate is lower than 15%. The authors
showed that adequately deep sequencing combined with MSA is able to correct errors
when their frequency is less than 20%. Above this value, error correction based on MSA
is possible with the simultaneous use of ECC. This method enables the cost and time re-
BioTech 2023, 12, 44 9 of 17

use of MSA (multiple sequence alignment) methods to establish a consensus sequence and
correct errors that may appear on the DNA strands. The MAFFT algorithm was chosen
for the analysis, which has been shown to be able to correct more than 95% of errors at
a sequencing depth reaching 100× when the error rate is lower than 15%. The authors
showed that adequately deep sequencing combined with MSA is able to correct errors
when their frequency is less than 20%. Above this value, error correction based on MSA
is possible with the simultaneous use of ECC. This method enables the cost and time
reduction needed for the DNA data storage procedure.
Erlich and Zielinski (2017) used the fountain algorithm to encode 2.14 × 106 bytes
of data [35]. The fountain encoding algorithm works in three steps: preprocessing, the
Luby transform, and screening (Figure 8). Overall, it aims to convert the input file into a
BioTech 2023, 4, x FOR PEER REVIEWcollection of DNA strings that pass synthesis and reading constraints. 10 of 18

Preprocessing—In this step, the input file is compressed using a lossless algorithm. Then,
the algorithm partitions the file into non-overlapping K segments, in which each segment
Luby transformation—This
is L bits long. L is defined stepbyconsists of many substeps. Briefly, a pseudo-random num-
the user.
ber generator determines the number
Luby transformation—This step consists of segments
of many that will be packed
substeps. Briefly,into a single packet. number
a pseudo-random
Encoded segments become packets known as droplets.
generator determines the number of segments that will be packed intoFor this, the algorithm uses a ro- packet.
a single
bust solution probability distribution, which assumes that most of the droplets will be
Encoded segments become packets known as droplets. For this, the algorithm uses a robust
created with a small number of input segments. On the segments of one droplet, the algo-
solution probability distribution, which assumes that most of the droplets will be created
rithm performs a bitwise exclusive or XOR operation. For example, consider that the al-
with
gorithm a small number
randomly of input
selected segments.
three input On 0100,
fragments: the segments
1100, [Link] droplet, the algorithm
case, the drop-
performs a bitwise exclusive or XOR operation. For example,
let is 0100 ⊕1100 ⊕1001 = 0001. In the end, the algorithm adds an index that specifiesconsider that thethealgorithm
binary representation of the seed, which, in turn, corresponds to the state of the random is 0100
randomly selected three input fragments: 0100, 1100, 1001. In this case, the droplet
⊕1100 ⊕
number 1001 = 0001.
generator In the end,
of the transform the the
during algorithm
generation adds
of theandroplet.
index Finally,
that specifies
it enablesthe binary
representation
the of thetoseed,
decoder algorithm inferwhich, in turn,
the identities corresponds
of the segments in tothe
thedroplet.
state of the random number
Screening—In
generator ofthe thelast step, the algorithm
transform during the excludes those strings
generation of the that do notFinally,
droplet. pass theitbio-
enables the
chemical
decoder constraints.
algorithm to Firstly,
inferbinary data are translated
the identities into a nucleotide
of the segments sequence: {00, 01,
in the droplet.
10, 11} to {A, C, G,
Screening—In theT}.last
Then, DNAthe
step, strings are screened
algorithm for GCthose
excludes content and homopolymers.
strings that do not pass the
The
biochemical constraints. Firstly, binary data are translated into a and
sequences that do not pass the screen are removed and the formation screeningsequence:
nucleotide
of the oligonucleotides are repeated until the desired conditions are obtained. In practice,
{00, 01, 10, 11} to {A, C, G, T}. Then, DNA strings are screened for GC content and
the authors recommend synthesizing 5–10% more oligonucleotides than the input seg-
homopolymers. The sequences that do not pass the screen are removed and the formation
ments.
and screening of the oligonucleotides are repeated until the desired conditions are obtained.
The idea for the decoding algorithm is to start with single-segment droplets and
In practice, the authors recommend synthesizing 5–10% more oligonucleotides than the
propagate that information through the other droplets until all the segments are recov-
input segments.
ered.

Figure
[Link]
Depictionof of
DNA fountain
DNA strategy.
fountain strategy.

5. DNA Preservation
Although the theoretical density of DNA data storage reaches petabytes per gram,
usually this value is unreachable. Due to the necessity of adding protective substances to
the DNA, the loading efficiency (DNA weight/total weight) ranks below 100%. Moreover,
the presence of indexes, such as Reed–Solomon codes, in long strands of DNA cause the
BioTech 2023, 12, 44 10 of 17

The idea for the decoding algorithm is to start with single-segment droplets and
propagate that information through the other droplets until all the segments are recovered.

5. DNA Preservation
Although the theoretical density of DNA data storage reaches petabytes per gram,
usually this value is unreachable. Due to the necessity of adding protective substances to
the DNA, the loading efficiency (DNA weight/total weight) ranks below 100%. Moreover,
the presence of indexes, such as Reed–Solomon codes, in long strands of DNA cause the
loss of data storage density. It was estimated that the index ratio of 200 bp DNA reaches
6.5%. Furthermore, DNA without protection is liable to degradation due to physical and
chemical factors, such as temperature, water, UV irradiation, oxidation, or extreme pH
values [36]. Therefore, current research focuses on increasing the DNA data storage density
and the time of its preservation by protecting DNA from the influence of high humidity
and the presence of oxygen [37].
The methods used for DNA preservation can be divided into two essential categories:
in vitro preservation, where DNA is usually stored in a single physical DNA pool, or
in vivo preservation, which uses living cells as DNA carrier systems [32].

5.1. In Vitro Preservation

The most common way to store data within DNA in vitro is solution storage. At
first, DNA was preserved in ethanol, however, over time the ammonium-based ionic
liquids gained popularity. Due to hydrogen bonding between ionic liquid and DNA, those
solutions improve DNA stability. However, the solution storage allows DNA to be stored
for only a year, which is insufficient to fulfill the aims of DNA data preservation (>1000
years).
On the contrary, solid-state DNA appears to be more stable due to its reduced molec-
ular mobility and lack of water, which causes hydrolytic damage [35]. The successful
amplification of DNA from ancient specimens, such as the Pleistocene cave bear, addi-
tionally indicates the effectiveness of the method [37]. Based on this discovery, Grass and
co-workers proposed DNA silica fossilization technology, through which they obtained
stable DNA after 35 days in 65 ◦ C (equivalent to two years at room temperature) [38].
Furthermore, Newman et al. (2019) developed a method for the preservation of dehydrated
DNA spots on glass cartridges, which can subsequently be recovered by a water droplet.
Multiple DNA spots on one cartridge additionally increase the storage density of 50 TB
of data per glass cartridge [39]. Choi et al. (2020) created a DNA micro-disc, which al-
lows easy access to data-encoded DNA and write-once-read-many memory. Firstly, the
encoded DNA’s primer sequences and data description were included in the QR code,
which facilitates easy access to the data. Secondly, due to the immobilization of DNA on
the micro-disc, after DNA enrichment using PCR, the original and amplified DNA are
separated. The sequence of the amplified DNA is subsequently converted into binary data
and the immobilized DNA can be read out in the future. Eventually, Choi et al. (2020)
reached a density of up to 1012 bit/mm3 for a single micro-disc and assessed the durability
of dehydrated DNA over 100 years at a temperature below 10 ◦ C [40].
DNA can also be easily stored via freeze drying or the addition of additives. In fact,
the lower the temperature, the longer the possible preservation. However, lyophilization
may cause cytolysis due to the formation of ice cracks [36]. Moreover, the estimated annual
cost of maintaining frozen samples around the globe likely surpasses USD 100 million
each year [41]. Therefore, due to the high cost currently, scientists are trying to develop an
effective method of DNA preservation at room temperature. For instance, the addition of
additives such as trehalose or PVA enables the DNA to be preserved at room temperature.
Both stabilizers create hydrogen bonds with negatively charged phosphate groups in DNA,
which has a protective effect on its stability [36]. However, Ivanowa and Kuzmina (2013)
indicate that, generally, the additives are insufficient for long-term DNA storage. Diluted
DNA in trehalose solution stored for a month at room temperature granted only 46% PCR
BioTech 2023, 12, 44 11 of 17

success, and 2-year preservation in Tris-buffered PVA granted 50% PCR success, where
PCR success was calculated as a percentage of positive wells per plate (96 samples) [42].
In Table 2, we summarize the storage methods used and the PCR success after storage
for a specified period at a specified temperature.

Table 2. Storage methods of DNA and PCR success after the storage.

Storage Method Time Temperature PCR Success Reference

Chemical
encapsulation
Silica nanoparticles 9 months RT x [43]
DNA-layered
1 month x x [44]
titanate nanohybrid
Solution
Preservation
”DNA stable” 4 years RT 98% [42]
DMSO salt solution 4 months RT 42% [45]
DMSO salt solution 2 years RT x [46]
70% ethanol 4 months RT 27% [45]
70% ethanol 2 years RT x [46]
90% ethanol 6 months RT 96% [47]
Formalin-fixed 30 years RT 30% [48]
Formalin-fixed 2–6 years RT x [49]
Paraffin-embedded
2–6 years RT x [49]
tissues
DETs buffer 6 months RT 92% [47]
TE buffer 1 night −20 ◦ C 100% [50]
TE buffer 3 years −20 ◦ C x [51]
Dehydratation
Ancient bone 521 years 13 ◦ C x [52]
Filter Paper 4 years RT 82.5% [53]
Dried DNA 4 months RT 35% [45]
FTA cards up to 128 days RT 95% [54]
Silica Gel 6 months RT 50% [47]
Oven-dried 6 months RT 72% [47]
Oven-dried 6 months −20 ◦ C 86% [47]
Freeze drying
DNA 4 years 4 ◦C 49% [42]
RT is abbreviation for “room temperature”. X indicates that the information was not specified in the reference.

In Table 3, we present the durability of DNA in various accelerated aging tests. Such
tests are performed to simulate the long-term behavior of DNA molecules in a much shorter
time by applying harsh conditions. The results of those experiments are presented as C/C0
(%), which is the percentage of the initial amount of DNA present in the sample after the
accelerated aging test.

Table 3. The durability of DNA in accelerated aging tests.

Relative
Storage Method Time Temperature Half-Life Temperature C/C0 Reference
Humidity
Parameters in
Experimental Conditions
Non-Experimental Conditions
Chemical
encapsulation
Silica nanoparticles 2 weeks 70 ◦ C 50% 20–90 years 20 ◦ C 90% [43]
Silica nanoparticles 10 days 60 ◦ C 50% 5 months RT 65% [55]
BioTech 2023, 12, 44 12 of 17

Table 3. Cont.

Relative
Storage Method Time Temperature Half-Life Temperature C/C0 Reference
Humidity
Parameters in
Experimental Conditions
Non-Experimental Conditions
Calcium
6 days 70 ◦ C 50% 1 year 10 ◦ C 0.1% [56]
phosphate crystals
1 million
”DNAshell” 2 days 100 ◦ C 50% 25 ◦ C x [57]
years
”DNAshell” 30 h 76 ◦ C 50% 100 years 25 ◦ C x [58]
”DNAshell” + ◦C ◦C
1 month 76 50% 2000 years 25 x [58]
trehalose
In silica 1 week 70 ◦C 50% 200 years 10 ◦C 10% [4]
Solution
Preservation
”DNA stable” 1 week 65 ◦ C 50% 4 years 25 ◦ C 10% [4]
”GenTra” 1 week 65 ◦ C 50% 2 years 25 ◦ C 50% [59]
TE buffer 20 days 65 ◦ C 50% 20 years −20 ◦ C x [51]
Dehydratation
DNA 6 weeks 50 ◦ C 50% x x 10% [60]
DNA silica
35 days 65 ◦ C 50% 2 years RT 15% [61]
fossilization
Dehydration with
6 days 70 ◦ C 50% 750 years 10 ◦ C 10% [62]
earth alkaline salts
DNA micro-disc 2 weeks 70 ◦C 50% >700 years 0 ◦C x [40]
DNA with ◦C ◦C
10 days 70 75% 17 years 10 x [63]
trehalose
Filter card 1 week 70 ◦C 50% 3.7 years 25 ◦C 1% [4]
Freeze drying
Polymer-plasmid
10 months 40 ◦ C 50% 3 years RT x [64]
complexes
Trehalose 2 months 60 ◦ C 50% 2 years RT x [65]
Cryosilicified
4 weeks 70 ◦ C 60% 1200 years 20 ◦ C 31% [66]
samples
Additives
Trehalose 2 years 56 ◦ C 50% 20 years RT 50% [42]
Trehalose 1 week 65 ◦ C 50% 160 years 10 ◦ C 20% [59]
PVA 2 years 56 ◦ C 50% 20 years RT 15% [42]
”Sugar mix” 1 week 65 ◦ C 50% 1 year 20 ◦ C 30% [59]
RT is abbreviation for “room temperature”. x indicates that the information was not specified in the reference.

5.2. In Vivo Preservation

Recently, in vivo preservation has been intensively developed. Preservation within a
living cell allows the DNA to be replicated with a few orders of magnitude, much faster
than by PCR, during the cell’s proliferation processes [67].
Bacteria are the most intuitive way to preserve DNA within a living organism. How-
ever, during bacterial replication, the spontaneous mutation rate is 2.2 × 10−10 mutations
per nucleotide per generation, or 1.0 × 10−3 mutations per genome per generation [68].
A generation time of about 20–30 min for E. coli means that after a few years of culti-
vation, mutations might represent a significant problem. Furthermore, the size of the
introduced plasmid is a serious limitation of in vivo preservation methods. So far, the
greatest amount of information in vivo has been encoded by Hao et al. (2020) thanks
to the mixed-circle method developed by them. The procedure involves the cloning of
data-encoded DNA oligonucleotides into plasmids and transforming E. coli cells with
recombinant, data-containing plasmids. During data recovery, plasmids are sequenced,
and oligonucleotides are assembled into original sequence. Eventually, 2304 kbp synthetic
BioTech 2023, 12, 44 13 of 17

oligonucleotides (encoding 455 KB of digital files) were used to create the mixed culture of
bacterial cells [67].
The solution to the problem of the limited size of the introduced plasmid appears to
be in vivo preservation on a yeast artificial chromosome. In 2021, Chen et al. created a
circular 255 kbp yeast artificial chromosome (a data-carrying chromosome; dChr) encoding
a total of 38 KB of digital data (two pictures and a video) [69]. Moreover, the dChr was
replicated with high fidelity, no mutation appeared after the 100th generation of replication,
while the encoding method used in this setup was tolerant toward a comparatively low
accuracy of Nanopore sequencing, enabling the fast retrieval of reliable data [69]. The high
fidelity of dChr replication could be achieved due to its chromatin-like structure formed
in vivo [70]. As it is known that nucleosomes regulate DNA repair mechanisms [71,72], the
utilization of eukaryotic organisms, such as Saccharomyces cerevisiae, carrying dChr is one of
the promising approaches for DNA data storage.
Another approach to in vivo storage is the preservation of data in endogenous DNA,
such as genomic DNA. This can be achieved using DNA-modifying enzymes such as nucle-
ases, integrases, or recombinases, although recently, the CRISPR-Cas9 system has gained
much popularity [73]. At the beginning of 2022, Liu et al. used a dual-plasmid system based
on a single crRNA-guided endonuclease (CRISPR-Cas12a) to encode a codebook (56 bytes)
and a picture (376 bytes) [74]. The authors used two plasmids, one with data-encoded
(target) DNA and the second with templates for the expression of Cas protein and crRNA,
which after bacteria transformation, enabled the introduction of target DNA to the E. coli
genome. Ultimately, the rewriting reliability reached 94% and the information sequenced
from the 252nd generation was 100% correct [74].
Studies on antimutator phenotypes have provided valuable insights into the sources
and mechanisms of spontaneous mutations. Research on carbon-starved E. coli populations
has shown that stress responses are required for the mutagenic repair of DNA breaks [75].
In the growing E. coli population, mutants of the α subunit of replicative DNA polymerase
III have been well characterized as antimutator alleles, suggesting that DNA replication
errors are a major source of spontaneous mutagenesis under optimal growth conditions [76].
However, these alleles also reduce specific transition mutations, making it unclear whether
replication errors in wild-type cells stem from the intrinsic fidelity of DNA polymerase III
or specific subpopulations with unique properties [77].
Despite the understanding of the molecular mechanisms controlling mutagenesis, the
process of spontaneous mutation in cells with functional mutation-prevention systems
remains unknown. To investigate this, a mutation assay on isogenic E. coli cells grow-
ing optimally without external stress was performed. It was revealed that spontaneous
DNA replication errors occurred more frequently in subpopulations experiencing internal
stresses, such as issues with proteostasis, genome maintenance, and reactive oxidative
species production. These mutator subpopulations do not significantly impact the aver-
age mutation frequency or the overall fitness of the population in a stable environment.
However, they play a crucial role in enhancing population adaptability in fluctuating
environments by providing a reservoir of increased genetic variability [78].
In turn, such mutator subpopulations may be responsible for introducing spontaneous
mutations in the E. coli population used for DNA data storage. Further understanding
the molecular background of spontaneous mutations may be helpful in minimizing the
occurrence of errors in the DNA used as a data storage medium in in vivo preservation
methods.

6. DNA Sequencing
To convert the DNA sequence back to its digital code, DNA has to be sequenced
and decoded to digital data using computer algorithms. Currently, the most commonly
used platforms for the sequencing of data-encoding DNA are Next-Generation Sequencing
by Illumina sequencing and Third Generation Sequencing by Oxford Nanopore Technol-
ogy [37].
BioTech 2023, 12, 44 14 of 17

One of the biggest advantages of Nanopore over Illumina for data output purposes is
its single-molecule sequencing of the extended alphabet, or its ability to sequence not only
natural nucleotides, but also chemically modified nucleotides. The applicability of such an
extended alphabet could significantly improve data storage in DNA by increasing storage
density and, possibly, writing speed [79]. However, Nanopore also has some limitations,
for instance, lower accuracy compared to Illumina. In fact, a direct comparison of the
error rates of Nanopore (∼10% per nucleotide in single read-out) and of Illumina (∼0.5%
per nucleotide) shows that Nanopore technology is approximately 20 times less accurate.
Therefore, at the moment, for DNA data storage purposes, the most commonly used is
Illumina sequencing [37].

7. Conclusions
Modern societies generate huge amounts of data and the rate of their growth has mul-
tiplied in recent years. The need to store both currently generated data and those generated
in the past using classical data storage methods are consuming huge financial outlay and
physical space. It also entails high costs for the environment, with the introduction of new
methods of data storage thus urgently required.
For a long time, people have paid attention to the high storage density and longevity
of DNA. In this article, we have provided a brief overview of how information is encoded
and stored in DNA. The continuous development of these methods leads to a reduction
in the number of errors appearing in the encoding and decoding processes, extending the
durability of DNA as a data carrier, and reducing the cost of its storage.
Despite the continued growth in the field of information storage on DNA, some
challenges still remain. There is a need to refine the methods used for the fast and error-free
synthesis of oligonucleotides, and in the long run, also of long DNA chains. The method
used to read nucleotide sequences also must evolve towards greater credibility.
Despite the current obstacles, the prospects for implementing data storage on DNA
are very promising. There are even new ideas related to the use of chemical analogues of
DNA, such as TNA, with even higher possible storage densities [26].

Author Contributions: Conceptualization, T.B., N.T. and T.I.; supervision, T.I.; writing—original
draft, T.B. and N.T.; writing—review and editing, T.B., N.T. and T.I. All authors have read and agreed
to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Acknowledgments: Figure 5 originates from: Yazdi et al., Portable and Error-Free DNA-Based Data
Storage. Sci. Rep. 2017, 7, 5011, Springer Nature, distributed under the Creative Commons Attribution
4.0 International License. We changed emoji pictures in panels: b, d, and f. Figure 6 is based on the
Figure 1 from: Sinyakov et al., Application of Array-Based Oligonucleotides for Synthesis of Genetic
Designs. Mol. Biol. 2021, 55, 487–500, Springer Nature. It has been reproduced with permission from
Springer Nature.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. De Silva, P.Y.; Ganegoda, G.U. New Trends of Digital Data Storage in DNA. BioMed Res. Int. 2016, 2016, 8072463. [CrossRef]
[PubMed]
2. Rydning, J.; Reinsel, D.; Gantz, J. The Digitization of the World from Edge to Core; IDC: Framingham, MA, USA, 2018.
3. Ceze, L.; Nivala, J.; Strauss, K. Molecular Digital Data Storage Using DNA. Nat. Rev. Genet. 2019, 20, 456–466. [CrossRef]
[PubMed]
4. Grass, R.N.; Heckel, R.; Puddu, M.; Paunescu, D.; Stark, W.J. Robust Chemical Preservation of Digital Information on DNA in
Silica with Error-Correcting Codes. Angew. Chem. Int. Ed. Engl. 2015, 54, 2552–2555. [CrossRef] [PubMed]
BioTech 2023, 12, 44 15 of 17

5. Zhirnov, V.; Zadegan, R.M.; Sandhu, G.S.; Church, G.M.; Hughes, W.L. Nucleic Acid Memory. Nat. Mater. 2016, 15, 366–370.
[CrossRef] [PubMed]
6. Van der Valk, T.; Pečnerová, P.; Díez-Del-Molino, D.; Bergström, A.; Oppenheimer, J.; Hartmann, S.; Xenikoudakis, G.; Thomas,
J.A.; Dehasque, M.; Sağlıcan, E.; et al. Million-Year-Old DNA Sheds Light on the Genomic History of Mammoths. Nature 2021,
591, 265–269. [CrossRef]
7. Horneck, G.; Klaus, D.M.; Mancinelli, R.L. Space Microbiology. Microbiol. Mol. Biol. Rev. 2010, 74, 121–156. [CrossRef]
8. Horneck, G.; Bücker, H.; Reitz, G. Long-Term Survival of Bacterial Spores in Space. Adv. Space Res. 1994, 14, 41–45. [CrossRef]
9. Cadet, J.; Sage, E.; Douki, T. Ultraviolet Radiation-Mediated Damage to Cellular DNA. Mutat. Res. 2005, 571, 3–17. [CrossRef]
10. Xue, Y.; Nicholson, W.L. The Two Major Spore DNA Repair Pathways, Nucleotide Excision Repair and Spore Photoproduct Lyase,
Are Sufficient for the Resistance of Bacillus Subtilis Spores to Artificial UV-C and UV-B but Not to Solar Radiation. Appl. Environ.
Microbiol. 1996, 62, 2221–2227. [CrossRef]
11. Sancho, L.G.; de la Torre, R.; Horneck, G.; Ascaso, C.; de Los Rios, A.; Pintado, A.; Wierzchos, J.; Schuster, M. Lichens Survive in
Space: Results from the 2005 LICHENS Experiment. Astrobiology 2007, 7, 443–454. [CrossRef]
12. Gauslaa, Y.; Solhaug, K.A. Photoinhibition in Lichens Depends on Cortical Characteristics and Hydration. Lichenologist 2004, 36,
133–143. [CrossRef]
13. Ahmed, R.K.; Mohammed, I.J. Developing a New Hybrid Cipher Algorithm Using DNA and RC4. Int. J. Adv. Comput. Sci. Appl.
2017, 8, 71.
14. Zhang, Y.; Kong, L.; Wang, F.; Li, B.; Ma, C.; Chen, D.; Liu, K.; Fan, C.; Zhang, H. Information Stored in Nanoscale: Encoding Data
in a Single DNA Strand with Base64. Nano Today 2020, 33, 100871. [CrossRef]
15. Church, G.M.; Gao, Y.; Kosuri, S. Next-Generation Digital Information Storage in DNA. Science 2012, 337, 1628. [CrossRef]
16. Goldman, N.; Bertone, P.; Chen, S.; Dessimoz, C.; LeProust, E.M.; Sipos, B.; Birney, E. Towards Practical, High-Capacity,
Low-Maintenance Information Storage in Synthesized DNA. Nature 2013, 494, 77–80. [CrossRef]
17. Ailenberg, M.; Rotstein, O.D. An Improved Huffman Coding Method for Archiving Text, Images, and Music Characters in DNA.
BioTechniques 2009, 47, 747–754. [CrossRef]
18. Yazdi, S.M.H.T.; Gabrys, R.; Milenkovic, O. Portable and Error-Free DNA-Based Data Storage. Sci. Rep. 2017, 7, 5011. [CrossRef]
19. Shipman, S.L.; Nivala, J.; Macklis, J.D.; Church, G.M. CRISPR-Cas Encoding of a Digital Movie into the Genomes of a Population
of Living Bacteria. Nature 2017, 547, 345–349. [CrossRef]
20. Bornholt, J.; Lopez, R.; Carmean, D.M.; Ceze, L.; Seelig, G.; Strauss, K. A DNA-Based Archival Storage System. In Proceedings
of the Twenty-First International Conference on Architectural Support for Programming Languages and Operating Systems,
Atlanta, GA, USA, 2–6 April 2016; ACM: Atlanta, GA, USA, 2016; pp. 637–649.
21. Blawat, M.; Gaedke, K.; Hütter, I.; Chen, X.-M.; Turczyk, B.; Inverso, S.; Pruitt, B.W.; Church, G.M. Forward Error Correction for
DNA Data Storage. Procedia Comput. Sci. 2016, 80, 1011–1022. [CrossRef]
22. Organick, L.; Ang, S.D.; Chen, Y.-J.; Lopez, R.; Yekhanin, S.; Makarychev, K.; Racz, M.Z.; Kamath, G.; Gopalan, P.; Nguyen, B.;
et al. Random Access in Large-Scale DNA Data Storage. Nat. Biotechnol. 2018, 36, 242–248. [CrossRef]
23. Choi, Y.; Ryu, T.; Lee, A.C.; Choi, H.; Lee, H.; Park, J.; Song, S.-H.; Kim, S.; Kim, H.; Park, W.; et al. High Information Capacity
DNA-based Data Storage with Augmented Encoding Characters Using Degenerate Bases. Sci. Rep. 2019, 9, 6582. [CrossRef]
[PubMed]
24. Lee, H.H.; Kalhor, R.; Goela, N.; Bolot, J.; Church, G.M. Terminator-Free Template-Independent Enzymatic DNA Synthesis for
Digital Information Storage. Nat. Commun. 2019, 10, 2383. [CrossRef] [PubMed]
25. Tabatabaei, S.K.; Wang, B.; Athreya, N.B.M.; Enghiad, B.; Hernandez, A.G.; Fields, C.J.; Leburton, J.-P.; Soloveichik, D.; Zhao, H.;
Milenkovic, O. DNA Punch Cards for Storing Data on Native DNA Sequences via Enzymatic Nicking. Nat. Commun. 2020, 11,
1742. [CrossRef] [PubMed]
26. Yang, K.; McCloskey, C.M.; Chaput, J.C. Reading and Writing Digital Information in TNA. ACS Synth. Biol. 2020, 9, 2936–2942.
[CrossRef] [PubMed]
27. Ren, Y.; Zhang, Y.; Liu, Y.; Wu, Q.; Su, J.; Wang, F.; Chen, D.; Fan, C.; Liu, K.; Zhang, H. DNA-Based Concatenated Encoding
System for High-Reliability and High-Density Data Storage. Small Methods 2022, 6, e2101335. [CrossRef]
28. Mayer, C.; McInroy, G.R.; Murat, P.; Van Delft, P.; Balasubramanian, S. An Epigenetics-Inspired DNA-Based Data Storage System.
Angew. Chem. Int. Ed. 2016, 55, 11144–11148. [CrossRef]
29. Sinyakov, A.N.; Ryabinin, V.A.; Kostina, E.V. Application of Array-Based Oligonucleotides for Synthesis of Genetic Designs. Mol.
Biol. 2021, 55, 487–500. [CrossRef]
30. Song, L.-F.; Deng, Z.-H.; Gong, Z.-Y.; Li, L.-L.; Li, B.-Z. Large-Scale de Novo Oligonucleotide Synthesis for Whole-Genome
Synthesis and Data Storage: Challenges and Opportunities. Front. Bioeng. Biotechnol. 2021, 9, 689797. [CrossRef]
31. Heckel, R.; Shomorony, I.; Ramchandran, K.; Tse, D.N.C. Fundamental Limits of DNA Storage Systems. In Proceedings of the
2017 IEEE International Symposium on Information Theory (ISIT), Aachen, Germany, 25–30 June 2017; pp. 3130–3134.
32. Zhang, Y.; Ren, Y.; Liu, Y.; Wang, F.; Zhang, H.; Liu, K. Preservation and Encryption in DNA Digital Data Storage. Chempluschem
2022, 87, e202200183. [CrossRef]
33. Meiser, L.C.; Antkowiak, P.L.; Koch, J.; Chen, W.D.; Kohll, A.X.; Stark, W.J.; Heckel, R.; Grass, R.N. Reading and Writing Digital
Data in DNA. Nat. Protoc. 2020, 15, 86–101. [CrossRef]
BioTech 2023, 12, 44 16 of 17

34. Xie, R.; Zan, X.; Chu, L.; Su, Y.; Xu, P.; Liu, W. Study of the Error Correction Capability of Multiple Sequence Alignment
Algorithm(MAFFT) in DNA Storage. BMC Bioinform. 2023, 24, 111. [CrossRef]
35. Erlich, Y.; Zielinski, D. DNA Fountain Enables a Robust and Efficient Storage Architecture. Science 2017, 355, 950–954. [CrossRef]
36. Tan, X.; Ge, L.; Zhang, T.; Lu, Z. Preservation of DNA for Data Storage. Russ. Chem. Rev. 2021, 90, 280–291. [CrossRef]
37. Doricchi, A.; Platnich, C.M.; Gimpel, A.; Horn, F.; Earle, M.; Lanzavecchia, G.; Cortajarena, A.L.; Liz-Marzán, L.M.; Liu, N.;
Heckel, R.; et al. Emerging Approaches to DNA Data Storage: Challenges and Prospects. ACS Nano 2022, 16, 17552–17571.
[CrossRef]
38. Paunescu, D.; Puddu, M.; Soellner, J.O.B.; Stoessel, P.R.; Grass, R.N. Reversible DNA Encapsulation in Silica to Produce
ROS-Resistant and Heat-Resistant Synthetic DNA “Fossils”. Nat. Protoc. 2013, 8, 2440–2448. [CrossRef]
39. Newman, S.; Stephenson, A.P.; Willsey, M.; Nguyen, B.H.; Takahashi, C.N.; Strauss, K.; Ceze, L. High Density DNA Data Storage
Library via Dehydration with Digital Microfluidic Retrieval. Nat. Commun. 2019, 10, 1706. [CrossRef]
40. Choi, Y.; Bae, H.J.; Lee, A.C.; Choi, H.; Lee, D.; Ryu, T.; Hyun, J.; Kim, S.; Kim, H.; Song, S.-H.; et al. DNA Micro-Disks for the
Management of DNA-Based Data Storage with Index and Write-Once-Read-Many(WORM) Memory Features. Adv. Mater. 2020,
32, e2001249. [CrossRef]
41. Anchordoquy, T.J.; Molina, M.C. Preservation of DNA. Cell Preserv. Technol. 2007, 5, 180–188. [CrossRef]
42. Ivanova, N.V.; Kuzmina, M.L. Protocols for Dry DNA Storage and Shipment at Room Temperature. Mol. Ecol. Resour. 2013, 13,
890–898. [CrossRef]
43. Chen, W.D.; Kohll, A.X.; Nguyen, B.H.; Koch, J.; Heckel, R.; Stark, W.J.; Ceze, L.; Strauss, K.; Grass, R.N. Combining Data
Longevity with High Storage Capacity—Layer-by-Layer DNA Encapsulated in Magnetic Nanoparticles. Adv. Funct. Mater. 2019,
29, 1901672. [CrossRef]
44. Kim, T.W.; Kim, I.Y.; Park, D.-H.; Choy, J.-H.; Hwang, S.-J. Highly Stable Nanocontainer of APTES-Anchored Layered Titanate
Nanosheet for Reliable Protection/Recovery of Nucleic Acid. Sci. Rep. 2016, 6, 21993. [CrossRef] [PubMed]
45. Frantzen, M.a.J.; Silk, J.B.; Ferguson, J.W.H.; Wayne, R.K.; Kohn, M.H. Empirical Evaluation of Preservation Methods for Faecal
DNA. Mol. Ecol. 1998, 7, 1423–1428. [CrossRef] [PubMed]
46. Kilpatrick, C.W. Noncryogenic Preservation of Mammalian Tissues for DNA Extraction: An Assessment of Storage Methods.
Biochem. Genet. 2002, 40, 53–62. [CrossRef] [PubMed]
47. Murphy, M.A.; Waits, L.P.; Kendall, K.C.; Wasser, S.K.; Higbee, J.A.; Bogden, R. An Evaluation of Long-Term Preservation
Methods for Brown Bear(Ursus Arctos) Faecal DNA Samples. Conserv. Genet. 2002, 3, 435–440. [CrossRef]
48. Vitošević, K.; Todorović, M.; Slović, Ž.; Varljen, T.; Matić, S.; Todorović, D. DNA Isolated from Formalin-Fixed Paraffin-Embedded
Healthy Tissue after 30 Years of Storage Can Be Used for Forensic Studies. Forensic. Sci. Med. Pathol. 2021, 17, 47–57. [CrossRef]
49. Ferrer, I.; Armstrong, J.; Capellari, S.; Parchi, P.; Arzberger, T.; Bell, J.; Budka, H.; Ströbel, T.; Giaccone, G.; Rossi, G.; et al. Effects
of Formalin Fixation, Paraffin Embedding, and Time of Storage on DNA Preservation in Brain Tissue: A BrainNet Europe Study.
Brain Pathol. 2007, 17, 297–303. [CrossRef]
50. Smith, S.; Morin, P.A. Optimal Storage Conditions for Highly Dilute DNA Samples: A Role for Trehalose as a Preserving Agent. J.
Forensic. Sci. 2005, 50, 1101–1108. [CrossRef]
51. Nguyen, H.H.; Park, J.; Park, S.J.; Lee, C.-S.; Hwang, S.; Shin, Y.-B.; Ha, T.H.; Kim, M. Long-Term Stability and Integrity of
Plasmid-Based DNA Data Storage. Polymers 2018, 10, 28. [CrossRef]
52. Allentoft, M.E.; Collins, M.; Harker, D.; Haile, J.; Oskam, C.L.; Hale, M.L.; Campos, P.F.; Samaniego, J.A.; Gilbert, M.T.P.; Willerslev,
E.; et al. The Half-Life of DNA in Bone: Measuring Decay Kinetics in 158 Dated Fossils. Proc. Biol. Sci. 2012, 279, 4724–4733.
[CrossRef]
53. Chaorattanakawee, S.; Natalang, O.; Hananantachai, H.; Nacher, M.; Brockman, A.; Krudsood, S.; Looareesuwan, S.; Patarapotikul,
J. Storage Duration and Polymerase Chain Reaction Detection of Plasmodium Falciparum from Blood Spots on Filter Paper. Am.
J. Trop. Med. Hyg. 2003, 69, 42–44. [CrossRef]
54. Saieg, M.A.; Geddie, W.R.; Boerner, S.L.; Liu, N.; Tsao, M.; Zhang, T.; Kamel-Reid, S.; da Cunha Santos, G. The Use of FTA Cards
for Preserving Unfixed Cytological Material for High-Throughput Molecular Analysis. Cancer Cytopathol. 2012, 120, 206–214.
[CrossRef]
55. Koch, J.; Gantenbein, S.; Masania, K.; Stark, W.J.; Erlich, Y.; Grass, R.N. A DNA-of-Things Storage Architecture to Create Materials
with Embedded Memory. Nat. Biotechnol. 2020, 38, 39–43. [CrossRef]
56. Antkowiak, P.L.; Koch, J.; Rzepka, P.; Nguyen, B.H.; Strauss, K.; Stark, W.J.; Grass, R.N. Anhydrous Calcium Phosphate Crystals
Stabilize DNA for Dry Storage. Chem. Commun. 2022, 58, 3174–3177. [CrossRef]
57. Coudy, D.; Colotte, M.; Luis, A.; Tuffet, S.; Bonnet, J. Long Term Conservation of DNA at Ambient Temperature. Implications for
DNA Data Storage. PLoS ONE 2021, 16, e0259868. [CrossRef]
58. Clermont, D.; Santoni, S.; Saker, S.; Gomard, M.; Gardais, E.; Bizet, C. Assessment of DNA Encapsulation, a New Room-
Temperature DNA Storage Method. Biopreserv. Biobank. 2014, 12, 176–183. [CrossRef]
59. Organick, L.; Nguyen, B.H.; McAmis, R.; Chen, W.D.; Kohll, A.X.; Ang, S.D.; Grass, R.N.; Ceze, L.; Strauss, K. An Empirical
Comparison of Preservation Methods for Synthetic DNA Data Storage. Small Methods 2021, 5, 2001094. [CrossRef]
60. Evans, R.K.; Xu, Z.; Bohannon, K.E.; Wang, B.; Bruner, M.W.; Volkin, D.B. Evaluation of Degradation Pathways for Plasmid Dna
in Pharmaceutical Formulations via Accelerated Stability Studies. J. Pharm. Sci. 2000, 89, 76–87. [CrossRef]
BioTech 2023, 12, 44 17 of 17

61. Puddu, M.; Paunescu, D.; Stark, W.J.; Grass, R.N. Magnetically Recoverable, Thermostable, Hydrophobic DNA/Silica Encapsu-
lates and Their Application as Invisible Oil Tags. ACS Nano 2014, 8, 2677–2685. [CrossRef]
62. Kohll, A.X.; Antkowiak, P.L.; Chen, W.D.; Nguyen, B.H.; Stark, W.J.; Ceze, L.; Strauss, K.; Grass, R.N. Stabilizing Synthetic DNA
for Long-Term Data Storage with Earth Alkaline Salts. Chem. Commun. 2020, 56, 3613–3616. [CrossRef]
63. Bonnet, J.; Colotte, M.; Coudy, D.; Couallier, V.; Portier, J.; Morin, B.; Tuffet, S. Chain and Conformation Stability of Solid-State
DNA: Implications for Room Temperature Storage. Nucleic Acids Res. 2010, 38, 1531–1546. [CrossRef]
64. Cherng, J.-Y.; Talsma, H.; Crommelin, D.J.A.; Hennink, W.E. Long Term Stability of Poly((2-Dimethylamino)Ethyl Methacrylate)-
Based Gene Delivery Systems. Pharm. Res. 1999, 16, 1417–1423. [CrossRef] [PubMed]
65. Molina, M.D.C.; Anchordoquy, T.J. Degradation of Lyophilized Lipid/DNA Complexes during Storage: The Role of Lipid and
Reactive Oxygen Species. Biochim. Biophys. Acta Biomembr. 2008, 1778, 2119–2126. [CrossRef] [PubMed]
66. Zhou, L.; Lei, Q.; Guo, J.; Gao, Y.; Shi, J.; Yu, H.; Yin, W.; Cao, J.; Xiao, B.; Andreo, J.; et al. Long-Term Whole Blood DNA
Preservation by Cost-Efficient Cryosilicification. Nat. Commun. 2022, 13, 6265. [CrossRef] [PubMed]
67. Hao, M.; Qiao, H.; Gao, Y.; Wang, Z.; Qiao, X.; Chen, X.; Qi, H. A Mixed Culture of Bacterial Cells Enables an Economic DNA
Storage on a Large Scale. Commun. Biol. 2020, 3, 416. [CrossRef]
68. Lee, H.; Popodi, E.; Tang, H.; Foster, P.L. Rate and Molecular Spectrum of Spontaneous Mutations in the Bacterium Escherichia
Coli as Determined by Whole-Genome Sequencing. Proc. Natl. Acad. Sci. USA 2012, 109, E2774–E2783. [CrossRef]
69. Chen, W.; Han, M.; Zhou, J.; Ge, Q.; Wang, P.; Zhang, X.; Zhu, S.; Song, L.; Yuan, Y. An Artificial Chromosome for Data Storage.
Natl. Sci. Rev. 2021, 8, nwab028. [CrossRef]
70. Zhou, J.; Zhang, C.; Wei, R.; Han, M.; Wang, S.; Yang, K.; Zhang, L.; Chen, W.; Wen, M.; Li, C.; et al. Exogenous artificial DNA
forms chromatin structure with active transcription in yeast. Sci. China Life Sci. 2021, 65, 851–860. [CrossRef]
71. Meas, R.; Wyrick, J.J.; Smerdon, M.J. Nucleosomes regulate base excision repair in chromatin. Mutat. Res.-Rev. Mutat. Res. 2019,
780, 29–36. [CrossRef]
72. Sun, Z.; Zhang, Y.; Jia, J.; Fang, Y.; Tang, Y.; Wu, H.; Fang, D. H3K36me3, message from chromatin to DNA damage repair. Cell
Biosci. 2020, 10, 9. [CrossRef]
73. Hao, Y.; Li, Q.; Fan, C.; Wang, F. Data Storage Based on DNA. Small Struct. 2021, 2, 2000046. [CrossRef]
74. Liu, Y.; Ren, Y.; Li, J.; Wang, F.; Wang, F.; Ma, C.; Chen, D.; Jiang, X.; Fan, C.; Zhang, H.; et al. In Vivo Processing of Digital
Information Molecularly with Targeted Specificity and Robust Reliability. Sci. Adv. 2022, 8, eabo7415. [CrossRef]
75. Al Mamun, A.A.M.; Lombardo, M.-J.; Shee, C.; Lisewski, A.M.; Gonzalez, C.; Lin, D.; Nehring, R.B.; Saint-Ruf, C.; Gibson, J.L.;
Frisch, R.L.; et al. Identity and function of a large gene network underlying mutagenic repair of DNA breaks. Science 2012, 338,
1344–1348. [CrossRef]
76. Oller, A.R.; Schaaper, R.M. Spontaneous mutation in Escherichia coli containing the dnaE911 DNA polymerase antimutator allele.
Genetics 1994, 138, 263–270. [CrossRef]
77. Schaaper, R.M. Suppressors of Escherichia coli mutT: Anitimutators for DNA replication errors. Mutat. Res. 1996, 350, 17–23.
[CrossRef]
78. Woo, A.C.; Faure, L.; Dapa, T.; Matic, I. Heterogeneity of spontaneous DNA replication errors in single isogenic Escherichia coli
cells. Sci. Adv. 2018, 4, eaat1608. [CrossRef]
79. Tabatabaei, S.K.; Pham, B.; Pan, C.; Liu, J.; Chandak, S.; Shorkey, S.A.; Hernandez, A.G.; Aksimentiev, A.; Chen, M.; Schroeder,
C.M.; et al. Expanding the Molecular Alphabet of DNA-Based Data Storage Systems with Neural Network Nanopore Readout
Processing. Nano Lett. 2022, 22, 1905–1914. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.