More powerful biomolecular computers

Biomolecular computers, along with quantum computers, may be a future alternative for traditional, silicon-based computers. Main advantages of biomolecular computers are massive parallel processing of data, expanded capacity of storing information and compatibility with living organisms (first attempts of using biomolecular computers in cancer therapy through blocking of improper genetic information are described in Benenson et al.). However, biomolecular computers have several drawbacks including time-consuming procedures of preparing of input, problems in detecting output signals and interference with by-products. Due to these obstacles, there are few laboratory implementations of theoretically designed DNA computers (like the Turing machine and pushdown automaton), but there are many implementations of DNA computers for particular problems . The first practical laboratory implementation of the general theoretical model of a machine performing DNA-based calculations was a simple twosymbol two-state finite automaton established by the Shapiro team. In the present work, we propose a new attitude, extending the capability of DNA-based finite automaton, by employing two or potentially more restriction enzymes instead of one used in other works. This creates an opportunity to implement in laboratories of more complex finite automata and other theoretical models of computers: pushdown automata, Turing machines. A finite automaton represents the simplest model of computer enabling solving plain problems (operations of addition and multiplication of integers cannot be performed by such machines). The finite automaton consists of a tape with cells containing an input word created from symbols of a certain finite alphabet and a control unit reading one after one symbols of the input word and changing its state according to transition rules. Each transition rule is of the form (s0, a) s1 which means: when an automaton is in the state s0 then after reading the