A symmetric DNA encryption process with a biotechnical hardware

1 Introduction

The growing rate of network technologies day by day dramatically increases the formation of data in the world. As the flow of information increases on the network, time security threats are also increasing for users. In order to protect the data, cryptography and steganography methods have been used from the past to the present. While cryptography codes the data, steganography hides it (Kaundal and Verma, 2014).

Cryptography is one of the most remarkable solutions for security issues. In cryptography, data is transferred between the sender and recipient over an untrusted medium. Cryptographic algorithms are classified by depending on the use of keys during encryption and decryption. These categories are symmetric cryptography and asymmetric cryptography. In the first one, encryption and decryption processes are implemented by using the same key. In the second one, encryption and decryption processes are implemented by using different keys. Here, each party has a pair of keys called private key and public key. The private key is always kept secret, while the public key can be shared (Namasudra et al., 2020).

Recently, DNA computing has been applied to cryptography. In DNA cryptography, DNA is an information carrier instead of any other medium like text, image, video, etc. Thus, this approach takes advantage of biological technology to achieve encryption. In other words, DNA is used as the carrier medium, while modern biological techniques are being used as application tools. However, it has disadvantages like expensive experimental equipment, complicated operations, and complex biotechnology. Hence, it still cannot be widely applied in the field of cryptography. In order to overcome these problems, some operations of DNA computing are employed to confuse information (Wang et al., 2015). Nowadays, the DNA concept is popular in the field of information security since it has a complex structure. Here, unlike the traditional computer computation concepts like 0 and 1, data are encrypted and stored by using DNA bases, Adenine (A), Cytosine (C), Guanine (G), and Thymine (T). So, the sender can choose any combination of these bases during the encryption process by improving the security of the data (Namasudra et al., 2020).

The motive of this area is to employ helpful metaphors from natural biological models to generate efficient security alternatives in the field of cryptography. In this paper, depending on the Central Molecular Biology principle (CMB) for encryption and decryption of systems, a novel encryption approach by combining DNA carrier medium, DNA encoding, and DNA XOR operation with Feistel network structure has been derived. Besides, the proposed DNA encryption process was both digitally and biologically integrated into designed biotechnical hardware. In order to provide confusion of the data, DNA encoding and DNA XOR have been integrated into Feistel Network. Another significant finding about DNA is that the capacity of electrical and optical media was exceeded by DNA. A gram of DNA may contain 10²¹ bases or about 10⁸ terabytes of data (Gehani et al., 2003). Finally, we aim to increase the number of hidden bits (capacity) by employing DNA as the carrier medium instead of traditional multimedia carriers (image, text, video and etc.) in cryptography.

The proposed study has been organized as six sections. In Section 2, similar studies in the literature have been mentioned. In Section 3, the used methodologies have been mentioned, and the proposed approach has been explained step by step by providing short samples. In Section 4, the simulation and design of biotechnical hardware have been explained. The results of the performed experiments have been shared and discussed in Section 5. A general outcome has been expressed in Section 6. Finally, the benefitted supports have been mentioned in Section 7.

2 Related work

The concept of DNA computing was firstly discussed in 1994 by Adleman with the goal of solving the Hamiltonian path problem. In 1995, Lipton solved the scheduling problem for 2-bit numbers, which is in another NP (non-deterministic Polynomial) complexity class, by using DNA molecules in a test tube (Kaundal and Verma, 2014). In 2003 by Chen, DNA cryptography algorithms based on molecular computing and single-used cipher were designed. Encryption/decryption steps were carried out on 2-dimensional images (Chen, 2003). In 2005, Tanaka et al. proposed a DNA cryptography algorithm based on asymmetric encryption. In the study, public keys were synthesized by helping non-identical mixtures. After key generation, the message was encoded with the first public key, and it was tied with the second public key into a DNA sequence. Asymmetric encryption is more secure than symmetric encryption, but it is slower than symmetric encryption when it is compared with symmetric encryption in terms of speed (Tanaka et al., 2005). In 2011, Kumar and Singh proposed a new method based on writing secret data in DNA sequences. In the proposed algorithm, they encrypted the word “HELLO” as plaintext by 350-bit of a single-used key. First, the word “HELLO” was converted to ASCII code equivalent. Then, they had the equivalent of binary pattern in 35 bits and 35 × 10 = 350 bits One Time Pad sequence. By using this produced nucleotide sequence and ssDNA key, they sent and received the data through the encryption/decryption perfectly (Kumar and Vijayaraghavan, 2011). A method based on RSA encryption algorithm that is encrypted via DNA sequences was proposed by Monika and Upadhyaya In 2015. The concerning method was adapted to SSL (Secure Socket Layer) technology in order to provide a higher level of security during the communication (Monika and Upadhyaya, 2015).

In 2016, Mousa proposed a DNA-Genetic Encryption Technique called D-GET. The primary purpose here is to make the technique more secure and less predictable. In this technique, the binary form of any digital data is converted to DNA sequencing, reshaped, encrypted, crossovered, mutated, and then reshaped. These main steps of D-GET are repeated three times or more. Experimental results demonstrated that his proposed technique had multilayer protection stages against different attacks and a higher level of security based on the multi-stages and genetic operations (Mousa, 2016).

In 2016 Goyat and Jain proposed a DNA-based cryptographic algorithm to secure the entire communication and storage of cloud servers for implementation and design. Their DNA-based cryptographic technique was basically developed using the substitution and other basic operator’s implementation. The experimental performance of their study in terms of time and space complexity provided effective and low resource-consuming techniques for data security in cloud-based systems. (Goyat and Jain, 2016).

In 2017 Ahmed and Mohammed developed a novel hybrid security algorithm called RC4-DNA-Alg. This algorithm uses both the symmetric stream cipher RC4 and DNA-indexing algorithms to provide secure data hiding with high complexity in the scope of steganography framework. The performance evaluation of their proposed scheme was measured by considering these three parameters; conditional entropy, randomness tests, and encryption time. Their result showed outperformance in security and distorted in hybrid cipher when compared to the native RC4 (Ahmed and Mohammed, 2017).

In 2018 Thangavel and Varalakshmi proposed a DNA cryptosystem to secure the original data within the DNA. They called their method Enhanced ElGamal cryptosystem. Their method is an asymmetric cryptosystem that has the purpose of solving key management issues in the cloud. They aim to overcome this problem by transferring the key file securely between the Data Owner and the Data User. They mentioned that Enhanced ElGamal cryptosystem provided better user authentication results. Besides, they added that it had better performance against security attacks (Thangavel and Varalakshmi, 2018).

In 2018 Narendren et al. proposed DNA based algorithm which acts as an additional layer to the RSA algorithm. In their method, a round function was employed that is based on the process of protein synthesis from DNA. They mentioned that the security analysis of their proposed scheme has proven to give good results (Narendren. et al., 2018). In 2018, Sharma and Sohal presented a novel cryptographic technique for cloud systems. Their approach uses client-side data encryption for encrypting the data before uploading it onto the cloud. Besides, it can also be considered a multifold symmetric-key cryptography technique that is based upon DNA cryptography. They compared their study with the existing symmetric-key algorithms; DNA, AES, DES, and Blowfish. They mentioned that their experimental results illustrated better outcomes when compared with the traditional algorithms in terms of ciphertext size, encryption time, and throughput (Sohal and Sharma, 2018).

In 2019 Basu et al. developed a cryptography algorithm inspired by the Central Dogma of Molecular Biology (CDMB). They implemented encryption and decryption stages via Genetic Coding simulation, which implies conversion from binary to DNA bases, Transcription, which means conversion from DNA to mRNA, and Translation which means conversion from mRNA to protein in a reversible way. Here, input blocks have the length of 16-bits. The final form of the blocks was concatenated to form the final ciphertext in the form of protein bases. Besides, they trained A Bidirectional Associative Memory Neural Network (BAMNN) for key generation. They mentioned that their method showed efficient encryption and decryption times when compared with the other existing systems in the literature (Basu et al., 2019).

In 2020 Tahir et al. developed a new model based on a genetic algorithm (GA) called CryptoGA to cope with data integrity and privacy issues on cloud data. In their method, GA was used to generate the keys for encryption and decryption, which are integrated with a cryptographic algorithm to ensure the privacy and integrity of cloud data. In experiments, execution time, throughput, key size, and avalanche effect were considered for evaluation and comparison. They expressed that experimental results analysis ensured the integrity and preserved the privacy of the user’s data against unauthorized parties. Moreover, they mentioned that CryptoGA was robust and provided a better performance on selected parameters when compared with the other state-of-the-art cryptographic algorithms like DES, 3DES, RSA, Blowfish, and AES (Tahir et al., 2021).

In 2020, Indrasena et al. proposed a bio-inspired cryptographic DNA system. Their method consists of three phases: encryption, key generation, and decryption. In their method, transcription and translation and some of the inverse procedures were used from reproducing the normal procedures in the genetic encoding for encryption and decryption of the data. Moreover, Central Dogma of Molecular Biology (CMDB) was employed. They compared their method with the traditional cryptographic techniques and reported 67% increased processing time for the encryption process and decryption process, respectively (Indrasena Reddy et al., 2020).

In 2021, Thabit et al. designed a cryptography algorithm that has two layers to improve cloud computing security. In their study, the first layer was inspired by Shannon’s diffusion and confusion theory. The second layer was inspired by the structures of genetic coding for cryptographic purpose. They employed simulation of natural processes of genetic cryptography (translation from binary to DNA bases), transcription (regeneration from DNA to mRNA), and translation (regeneration from mRNA to protein). They obtained remarkable experimental results which can be used to secure applications on cloud computing in terms of size and execution time. (Thabit et al., 2021).

In this study, a DNA cryptography technique has been proposed by integrating DNA encoding and DNA operators into the Feistel network structure. Here, DNA itself was used as a carrier instead of traditional digital media such as image, text, or video, while modern biological tools were being used as implementation tools. Besides, the developed simulation software and the synthesized DNA sequence were both digitally and biologically integrated into specifically created biotechnical hardware. Since storing data in DNA environment was also aimed, the amount of data (bits) that can be embedded in a DNA base (A, C, G, and T) becomes an important problem here. When the studies in the literature were examined, it was seen that the number of bits that can be encoded per nucleotide did not exceed two. In the proposed study, this rate has been increased with the original compression algorithm that is carried out before the encryption process. When the biologically synthesized DNA sequence was compared with the simulated DNA sequence obtained by the decoding process in the simulation, 100% match success was obtained. The simulation software was also integrated into the hardware that can execute in a plug-and-play manner. These are the main features that make the proposed study versatile in terms of DNA-based data storage.

3 The proposed method

In this section, first, the main principle of DNA cryptography has been mentioned by briefly expressing the biological structure of DNA. Then the proposed algorithm has been explained with the details of each substep by providing the concrete samples.

3.4

3.4 Encryption process

In the proposed study, DNA encoding (Kumar and Singh, 2011) and DNA XOR (Mondal and Mandal, 2017) are integrated into the Feistel network. The Feistel network has been depicted in Fig. 1. This integration has been performed by using traditional DES S-boxes. In Feistel networks, the design of F function is crucial since it measures the originality and complexity of the approach. Here, a novel F function has been developed by benefitting DNA XOR and DNA shifting, as depicted in Fig. 2(b). Fig. 2(a) depicts the usage of DES S-box. As a result, the proposed scheme is an example of symmetric encryption.

Close

Fig. 1

Block diagram of the Feistel Network Coding (a) Encryption process (b) Decryption process (Knudsen and Robshaw, 2011).

Export to PPT

Close

Fig. 2

(a) The use of DES S_Box_S1 (b) The designed F function.

Export to PPT

Mathematically, the formulation of Feistel network encryption can be shown in Eqs. (4) and (5). $$L_{i+1}=R_i$$ $$R_{i+1}=L_i\oplus F(R_i,K_i)$$

Step 1: In this step, the obtained base sequence after compression is converted to binary via DNA Base-Bit transformation, expressed as follows by Eq. (6):

(6)

$${B\text{'}}_{1,l\times 2}=DNAEncode{(D}_{1,l})$$

Step 2: In this step, the obtained bit sequence is divided into groups of 12 bits blocks via Eq. (7).

(7)

$${G\text{'}}_{1,j=}\{w^{\text{'}}:\left|w^{\text{'}}\right|=12\&w^{\text{'}}\in B\text{'}\}$$

Here, each group is again divided into two equal parts in itself. These 6-bits groups are called left, L_i and right R_i. Thus, L_i and R_i are ready to submit to Feistel network structure as described in Table 1.

Table 1 The process of adapting the compressed data to Feistel Network structure.

D	TGTAGATTTTTTTGTGTTTATGGGTCCTTGT…
B′	111011001011111111111111101110111111001110101011010111111011…
G′	111,011,001,011		111,111,111,111		101,110,111,111		…
6 bits groups	111,011	001,011	111,111	11,111	101,110	111,111	…	…
Part of Feistel	Lİ	Rİ	Lİ	Rİ	Lİ	Rİ	Lİ	Rİ

Step 3: In Feistel networks, F function makes the structure original according to the designer. In the proposed study, a novel F function has been formed by using DES S-Boxes, DNA XOR, and DNA shifting operators. Pseudo codes of this process have been provided below for ease of understanding:

1. Truncate w’ into two equal pieces (R_i, L_i)

2. Exchange left and right pieces

3. (Li) ↔ (Ri)

4. For x  = 1 to 8

5.   {

6.    s = S_Box_Sx (Ri)

7.    For y = 1 to 5

8.    e’ = DNA_F_Function(s) (Refer to Figure 2.b)

9.   }

In Fig. 2(a), R_i from Table 1 enters S-Box_S1. Hence the output becomes in the length of 4 bits. The bases corresponding to the obtained 4 bits are subjected to DNA XOR with Matrix_1 in Fig. 2(b). As a result, Matrix_2 is obtained. Matrix_2 is subjected to DNA XOR operation again, but this time as depicted in a way marked as 3 in Fig. 2(b). The result still consists of 4 bases and, accordingly, 8 bits. Also, this is the output of our F function. Here, 2 LSB bits of each 8 bit are the round keys, as shown in Fig. 2(b). The remaining 6 bits, namely 3 bases, are ready to enter XOR process with L_i in the Feistel network. In total, 8 loops are made for 8 S-Boxes (S1,…,S8), and Matrix_1 is shifted 5 times for each S-Box loop. Thus, L_i+1 and R_i+1 are formed after 40 loops. Eventually, the encoded DNA chain is obtained. In Fig. 2(a) and (b), only one round of the developed F function has been presented. Pseudo codes of this process have been provided below for ease of understanding:

Due to Feistel structure, notice that

• Li + 1 = Ri

• Ri + 1 = e’ XOR Li

• e = Li + 1 // Ri + 1

• E = {e: e = Li + 1 // Ri + 1})

In Fig. 3, a flowchart of the encryption process has been provided by separating and marking each step.

Close

Fig. 3

Flowchart of the encryption process.

Export to PPT

3.5

3.5 Decryption process

Owing to the symmetric structure of encryption process, the decoding process can be easily obtained by inverting the encryption process. The plain data can be obtained back by adapting the encrypted DNA sequence to the Feistel network structure, as seen in Fig. 1(b).

The decoding phase of the Feistel network can be mathematically shown as follows via Eqs. (8) and (9): $$R_i=L_{i+1}$$ $$L_i=R_{i+1}\oplus F\left(L_{i+1},K_i\right)$$

Step 1: In this step, the encrypted DNA chain is divided into groups of 6 bases for presenting the Feistel network, as depicted in Fig. 1(b). These obtained 6-base groups are expressed as L_i+1 and R_i+1. They are in the form of two half in order to be adapted to the Feistel structure again, but this time, inversely. This process is shown via Eq. (10):

(10)

$${B\text{'}\text{'}}_{1,j\times 2}={\mathit{Base}}_2{(E}_{1,j})$$

Step 2: In this step, B″ is separated into words, each of which length is 12 bits. We call the resulting array T. In the decryption process, L_i+1 is the entry of F function. It takes L_i, directly and here F function runs in reverse order of the encryption process. The 6-bit groups formed in R_i+1 and the result of F function enters DNA XOR processing. Hence, the obtained result is R_i. This process is shown by Eq. (11):

(11)

$$T_{1,j\times 2}=\{w^{\text{'}\text{'}}:\left|w^{\text{'}\text{'}}\right|=12\&w^{\text{'}\text{'}}\in ({B\text{'}\text{'}}_{1,j\times 2})\}$$

Pseudo codes of this process have been provided below for ease of understanding:

1. Truncate w″ into two equal pieces (Ri′, Li′)

2. Notice that: Li′ = Li + 1 and Ri′ = Ri + 1 since the decryption process is the reverse function of encryption.

3. For y = 1 to 5

4.  {

5.   e′ = reverse (DNA_F_Function (Li′)

6.   For x  = 1 to 8

7.   s = reverse (S_Box_Sx (e′))

8.  }

Step 3: Due to the structure of Feistel, L_i+1 becomes the new L_i. The resultant L_i and R_i groups are subjected to Bit-Base transformation, and thus, the plain DNA sequence is obtained.

• Ri + 1 = s XOR Li′

• Li + 1 = Ri′

• $b\text{'}\in$ [B′] = Li + 1 // Ri + 1

Notice that D corresponds to DNA encoding of B. Here, after the decompression process that is the reverse function of compression process, B is obtained. Since B is the binary form of input data I, the initial input file has been obtained. In Fig. 4, flowchart of the decryption process has been provided by separating and marking each step.

Close

Fig. 4

Flowchart of the decryption process.

Export to PPT

4 Simulation and the design of biotechnical hardware

Simulation of the proposed method has been carried out by using C# language. Software testing has been performed by a computer with i7, 2.2 GHz processor, 6 GB RAM, and Win7 64bit operating system. The simulation output, including encryption and decryption phases, has been presented in Fig. 5.

Close

Fig. 5

Encryption and decryption phases interface of the simulation software.

Export to PPT

In Fig. 5, flow of the simulation software can be followed in “Process Information” part. The concerning output after encryption and decryption processes can be obtained via the given note pad file on the right side of the interface.

Biotechnical hardware has a unique design that is drawn in CATIA V5 program. The created design was manufactured with the help of a 3D printer. For a detailed view, Fig. 6 needs to be considered. DNA cells in Fig. 6 in biotechnical hardware provide the possibility for storing the synthesized DNA. On the other hand, the simulation software that produces the concerning digital DNA sequence resulting from encryption or decryption is ready to be run on any Windows-based system.

Close

Fig. 6

The internal structure of the developed biotechnical hardware.

Export to PPT

When the general specifications of the studies in the literature were examined, it was seen that the processes of DNA computing were gradually increasing. However, it was noted that this process always linearly depended on the speed of the biological process. Furthermore, most of the performed studies stayed limited by the simulation. In the proposed study, this lack is addressed. First of all, a novel DNA cryptography algorithm that is suitable with the nature of DNA was improved. Then, the limitation of the simulation was solved by integrating the simulation software into the produced plug-and-play biotechnical hardware. The internal structure of the developed biotechnical hardware is illustrated in Fig. 6. Here, the integrated DNA cells were designed to store the synthesized DNA sequence in the form of DNA drops. These DNA drops are needed to be diluted by a special solution before sequencing. Hence, we designed each DNA cell to store 1 mg of DNA. It is widely known that 1 g of DNA can store 10⁸ Terabytes of data. Since we have 6 DNA cells in the designed biotechnical hardware, we are able to store 6 × 10⁵ TBytes of data.

5 Experimental results

In this section, the performed biological experiments, performance, and security analyses and their results have been presented by providing concrete samples as much as possible.

5.1

5.1 In vitro experiments

In the proposed study, 9588 bp of encrypted DNA sequence, which is obtained via the simulation software, has been synthesized with the support of Düzce University Scientific Research Projects Coordination ship. After synthesizing DNA chain, it is an essential requirement to obtain the original DNA sequence by re-reading it. This process is called DNA sequencing. Fig. 7 shows the sequence result of the synthesized DNA chain. Here, after sequencing of 9588 bp DNA chain, it has been seen that the synthesized and the sequenced DNA chains match without loss of information. The synthesized 9588 bp DNA chain has been encapsulated in a plasmid.

Close

Fig. 7

DNA chain sequence results.

Export to PPT

6 Conclusion

In this paper, depending on the Central molecular biology principle for encryption and decryption of systems, a novel encryption approach by combining DNA carrier medium, DNA encoding, and DNA XOR operation with Feistel network structure has been derived. Besides, the proposed DNA encryption process was both digitally and biologically integrated into designed biotechnical hardware.

The performed experiments showed that DNA cryptology has a high potential to be a new method in data security. Experimental results demonstrated that the proposed study has efficient outcomes in terms of capacity, brute force attack, key space, and entropy analyses. Besides, the implementation of the proposed method has been verified by vitro experiments.

However, in the scope of this study, there are still some obstacles. First of all, this study bases on the symmetric encryption process. For a more secure scheme, asymmetric encryption-based structures have to be examined. In DNA structure, random access to data is another problem since this technology is a brand new field. Here, we need extra techniques to speed up the whole process when we consider the processing time and the speed of the synthesized DNA instead of the silicon medium. As mentioned above, dynamic encoding of each character is another requirement that increases the robustness of an algorithm. However, the design has to be performed without occupying the memory and causing any speed loss in this process. All of these issues are planned to be tackled in future studies.

Funding

The experimental stages of the proposed study were supported by Düzce University Scientific Research Projects Coordinationship (project number 2016.06.01.502) in the project called “Data Storage in the DNA Chain.” Furthermore, the compression algorithm in the proposed study is in the national patent process with the Patent Pending ID number of 2017/00459.