Research‎ > ‎


DNA has certain unique properties such as self-assembly and self-complementary in hybridization, which are important in many DNA-based technologies. DNA computing, for example, uses these properties to realize a computation, in vitro, which consists of several chemical reactions [1]. Other DNA-based technologies such as DNA chips for mutational analysis and sequencing [2] also depend on hybridization to assemble nanostructure and to amplify DNA templates, respectively. Hybridization of DNA can be controlled by properly designing DNA sequences.

Various DNA sequence design approaches have been proposed to date [3-10] but most of them assumed that the length of the DNA sequences required is fixed. However, in a DNA computing model called length-based DNA computing [11] for example, the encoding by length is realized whereby the cost of each path is encoded by the length of the DNA in a proportional way. Even though a generate-and-test DNA sequence design for length-based DNA computing has been proposed [12], no fitness function is used for the generation of DNA sequences. Note that if an optimization approach, such as ant colony optimization (ACO) is employed for DNA sequence design of arbitrary length, the original fitness functions: similarity, H-measure, continuity, and hairpin, can not be used since those original fitness functions are formulated to evaluated DNA sequence of fixed length.

Hence, there is a need to re-formulate the original fitness functions to cater the evaluation of DNA sequences of arbitrary length. After similarity, H-measure, continuity, and hairpin, are re-formulated, the DNA sequence design is implemented using ACO [10]. The sequences generated using ACO with modified fitness functions will be compared with DNA sequences generated based on generate-and-test approach. It is expected that the modified fitness functions can be further employed in DNA sequence design, especially if a set of an arbitrary length of DNA sequences is required.