Year: 2026
Type: article
Abstract: Introduction Chemical reaction networks (CRNs) are widely used as a design and modeling language for molecular computing and can be compiled into DNA strand-displacement systems. However, designing efficient CRNs for nontrivial mathematical functions remains difficult. Here we present FUNDNA, a software tool that automates the design of CRNs for computing mathematical functions from a high-level user specification. Methods FUNDNA accepts a target single-variable function, computes a Maclaurin-series approximation, rearranges the resulting polynomial using either Horner’s rule or double-NAND expansion, maps the rearranged expression to cascaded molecular AND/NAND units, and outputs the corresponding CRN. The generated CRNs can then be translated to DNA reactions using existing tools. Results We tested the FUNDNA workflow using the 14 function-entry templates available through the software interface and performed DNA-level simulations on four representative eligible examples. The generated rearranged expressions matched their theoretical targets, and the simulated outputs closely tracked the exact function values. Across the reported test cases, the mean square error was below 3.52 × 10 −4 . Discussion FUNDNA provides an accessible workflow for automated design of function-computing CRNs and can help accelerate the development of more complex biomolecular computing circuits.
Authors: Jackson Huse, Sayed Ahmad Salehi, Mahdi Ahmadi, Mohammadreza Tavasoli Naeini
Institution: University of Kentucky
Language: English
Topic: DNA and Biological Computing
Subfield: Theoretical Computer Science
Field: Mathematics
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,801
Funder: National Science Foundation
OpenAlex ID: W4387906939
