Implementation of k-shortest paths computation on a molecular computer via hybridization-ligation and parallel overlap assembly

Abstract

In this paper, an in vitro implementation of DNA computing for finding k-shortest paths of a weighted graph is reported. After the initial solution is subjected to amplification by polymerase chain reaction, k-shortest paths are visualized by polyacrylamide gel electrophoresis and the selection can be done. For initial pool generation, hybridization-ligation and parallel overlap assembly are employed and compared. The experimental results show the effectiveness of the proposed in vitro computation, where both initial pool generation methods are able to generate the initial pool candidates for the k-shortest paths computation on a DNA computer. However, parallel overlap assembly is more preferable than that of hybridization-ligation in terms of population size and generation time.