Analog Computation by DNA Strand Displacement Circuits

DNA circuits have
been widely used to develop biological computing
devices because of their high programmability and versatility. Here,
we propose an architecture for the systematic construction of DNA
circuits for analog computation based on DNA strand displacement.
The elementary gates in our architecture include addition, subtraction,
and multiplication gates. The input and output of these gates are
analog, which means that they are directly represented by the concentrations
of the input and output DNA strands, respectively, without requiring
a threshold for converting to Boolean signals. We provide detailed
domain designs and kinetic simulations of the gates to demonstrate
their expected performance. On the basis of these gates, we describe
how DNA circuits to compute polynomial functions of inputs can be
built. Using Taylor Series and Newton Iteration methods, functions
beyond the scope of polynomials can also be computed by DNA circuits
built upon our architecture.