Digital information encoding on DNA

Novel approaches to information encoding with DNA are explored using a new Watson-Crick structure for binary strings more appropriate to model DNA hybridization. First, a Gibbs energy analysis of codeword sets is obtained by using a template and extant error-correcting codes. Template-based codes have too low Gibbs energies that allow cross-hybridization. Second, a new technique is presented to construct arbitrarily large sets of noncrosshybridizing codewords of high quality by two major criteria. They have a large minimum number of mismatches between arbitrary pairs of words and alignments; moreover, their pairwise Gibbs energies of hybridization remain bounded within a safe region according to a modified nearest-neighbor model that has been verified in vitro. The technique is scalable to long strands of up to 150-mers, is in principle implementable in vitro, and may be useful in further combinatorial analysis of DNA structures. Finally, a novel method to encode abiotic information in DNA arrays is defined and some preliminary experimental results are discussed. These new methods can be regarded as a different implementation of Tom Head's idea of writing on DNA molecules [22], although only through hybridization.