Graph replacement chemistry for DNA processing

Abstract

The processing of nucleic acids is abstracted using operators on directed and labeled graphs. This provides a computational framework for predicting complex libraries of DNA/RNA arising from sequences of reactions involving hybridisation intermediates with significant combinatorial complexity. It also provides a detailed functional classification scheme for the reactions and side-reactions of DNA processing enzymes. It is complementary to the conventional string-based DNA Computing grammars such as splicing systems, in that the graph-based structure of enzyme-nucleic acid complexes is the fundamental object of combinatorial manipulation and in that the allowed reactions are specified by local graph replacement operators (i.e. catalysts for structural transitions) associated with enzymes. The focus of the work is to present a calculus for the compact specification and evaluation of the combined action of multiple DNA-processing reactions. Each enzyme and its sidereactions may be classified by a small set of small graph replacement operators. Complex replication and computation schemes may be computed with the formalism.