Optimal Codeword Construction for DNA-based Finite Automata

TL;DR

This paper develops optimal codeword construction methods for DNA-based finite automata, providing theoretical bounds, automatic encoding solutions, and simulation validation to advance biomolecular computation applications.

Contribution

It introduces automated encoding techniques using heuristic and ILP methods, along with theoretical bounds and simulation validation for DNA finite automata construction.

Findings

Derived exact bounds on symbols and states

Proposed heuristic and ILP encoding solutions

Validated approach through simulation experiments

Abstract

Biomolecular computation has emerged as an important area of computer science research due to its high information density, immense parallelism opportunity along with potential applications in cryptography, genetic engineering and bioinformatics. Computational frameworks using DNA molecules have been proposed in the literature to accomplish varied tasks such as simulating logical operations, performing matrix multiplication, and encoding instances of NP-hard problems. In one of the key applications, several studies have proposed construction of finite automata using DNA hybridisation and ligation. The state and symbol encoding of these finite automata are done manually. In this manuscript, we study the codeword construction problem for this approach. We derive exact theoretical bounds on the number of symbols and states in the finite automata and also obtain the complete set of symbols in a specific case. For automatic encoding, two different solutions, based on a heuristic and on Integer Linear Programming (ILP), are proposed. Furthermore, we propose an early simulation-based validation of laboratory experiments. Our proposed flow accepts a finite automaton, automatically encodes the symbols for the actual experiments and executes the simulation step-by-step.