The ultimate goal of DNA computing is to store information at higher density and solve complex problems with less computational time and minimal error. Most algorithmic DNA lattices have been constructed using the free-solution growth (FSG) annealing method, and hairpin-embedded DNA rule tiles have been introduced in most algorithmic implementations to differentiate 0- and 1-bit information. Here, we developed streptavidin (SA)-decorated algorithmic COPY (produced line-like patterns with biotinylated 1-bit rule tiles) and XOR (triangle-like patterns) lattices constructed by a substrate-assisted growth (SAG) method and FSG. SA decoration in algorithmic lattices provides an efficient platform for visualizing bit information, and the SAG method in algorithmic assembly offers full coverage of algorithmic lattices on a substrate with a relatively lower DNA concentration than previous methods. The algorithmic COPY and XOR lattices assembled with various ratios of 0- and 1-bit rule tiles were verified by atomic force microscopy. We found that even asymmetric DNA patterns produced by certain algorithmic logic gates could be easily constructed by SAG. Finally, we evaluated sorting factors and error rates of algorithmic COPY and XOR lattices to determine the bit population and quality of the algorithmic assembly. Because of the catalytic effect of the substrate, the sorting factor of algorithmic DX-DNA lattices did not greatly influence the specific rules (i.e., COPY and XOR logic gates) annealed by SAG. Additionally, we found that the overall error rates of algorithmic DX-DNA lattices prepared by the FSG and SAG methods were low, within the range of 1-3%. Hence, the self-assembled algorithmic patterns generated with DNA molecules may serve as a scaffold for molecular demultiplexing circuits and computing.
Keywords: DNA self-assembly; algorithm; logic implementation; streptavidin; substrate-assisted growth.