05-05-2011, 12:31 PM
Abstract
Biotechnological methods can be used for cryptography. Here two different cryptographic approaches based onDNA binary strands are shown. The first approach shows how DNA binary strands can be used for steganography,a technique of encryption by information hiding, to provide rapid encryption and decryption. It is shown that DNAsteganography based on DNA binary strands is secure under the assumption that an interceptor has the sametechnological capabilities as sender and receiver of encrypted messages. The second approach shown here is based onsteganography and a method of graphical subtraction of binary gel-images. It can be used to constitute a molecularchecksum and can be combined with the first approach to support encryption. DNA cryptography might become ofpractical relevance in the context of labelling organic and inorganic materials with DNA ‘barcodes’. © 2000 ElsevierScience Ireland Ltd. All rights reserved.Keywords: DNA computing; Cryptography; Steganography; Graphical decryption; DNA binary strands; Molecular checksum; DNAbarcodeselsevier.com:locate:biosystems
1. Introduction
As a medium with high information density,DNA was proposed for computational purposes(Adleman, 1994). Since then several approacheshave been investigated like implementations ofcombinatorial (Adleman, 1994; Lipton, 1995;Ouyang et al., 1997) and functional (Guarnieri etal., 1996) algorithms and approaches based onself-assembly (Winfree et al., 1996, 1998; Rauhe etal., 1999). Theoretical considerations dealt withTuring machines, associative memory andcryptanalysis.Cryptography has been shown recently as anew application of DNA Computing: Clelland etal. (1999) have demonstrated an approach tosteganography by hiding secret messages encodedas DNA strands among a multitude of randomDNA. Steganography means hiding of secret messagesamong other information to conceal theirexistence (Kahn, 1967; Schneier, 1996) and isknown as a simple cryptographic method. Clel land et al. (1999) have used a substitution cipherfor plaintext encoding where a unique base tripletis assigned to each letter of the alphabet, eachnumeral and some special characters.Instead, as digital messages usually correspondto 0–1-series, a binary DNA representation hasbeen used here (see Fig. 1). The binary encoding isin particular suitable for the construction ofdatastructures and for simple and rapid decryption.Decryption can be done by an adaptedmethod of digital DNA typing originally developedfor minisatellite analysis (Jeffreys et al.,1991) (see Fig. 2). Using this method the informationcontent can be decrypted and read directly byPCR and subsequent gel-electrophoresis, requiringno additional work such as subcloning orsequencing.DNA binary strands were assembled by concatenationof short double stranded DNAmolecules representing 0 (0-DNA bit), 1 (1-DNAbit), start or end as described earlier (Rauhe et al.,1999). The DNA molecules contain overlappingsequences (‘sticky ends’) and were polymerized toDNA binary strands by annealing and ligation(see Fig. 1). In order to isolate single moleculesfrom the pool of all generated DNA strands, themolecules were ligated into plasmids (see Fig. 2b)and cloned in bacteria. Then the informationalcontent of every cloned strand could be readindividually by PCR and subsequent gel-electrophoresis.For the readout PCR a strand’s startterminator and its bits were used as priming sites(see Fig. 2).
2. DNA steganography — method I
As the readout procedure is based on the knowledgeof the primer sequences the primers areessential if there is no other way of reading thebinary strand. Thus mixing a certain binary strandwith other DNA becomes a steganographic approachto encryption as it prevents reading thebinary strand by sequencing.For encryption, the message strand that correspondsto the binary encoded plaintext was mixedwith other DNA, so-called dummy strands, inequimolar amounts (see Fig. 2). To achieve bettersecurity, the dummy strands should have the samebinary format as the message strand. For decryption,a unique identification sequence (key sequence)attached to the message strand is required.This can be any of the terminator sequences,normally the start sequence (see Fig. 2). Thusdecryption was done by readout of the messagestrand using the appropriate key sequence as oneprimer of a PCR reaction (see Fig. 2). The otherprimer is either the corresponding 0-DNA bit or thecorresponding 1-DNA bit. Performing both PCRreactions separately and visualizing the results bygel-electrophoresis yielded complementary patternsof bands that were read from the gel
Download full report
http://elmu.umm.ac.id/file.php/1/jurnal/...0/1971.pdf