Ehud Shapiro

 

My research, carried out in part at the Department of Biological Chemistry, explores the use of computer science concepts and tools in molecular biology. One research area is the development of a computer made of biological molecules. The research started with a conceptual model of a Turing machine, suitable for biomolecular implementation, and its mechanical realization. The next step (Nature 2001) was the development of an autonomous programmable molecular finite automaton. In this two-state, two symbol automaton the input and software were made of DNA, and the hardware consisted of DNA manipulating enzymes. A subsequent version of the automaton used only a single enzyme as hardware, and obtained all the necessary fuel for its operation from the cleavage of its input DNA molecule in the normal course of the computation (PNAS 2003). The automaton was then shown to be capable of stochastic computing. We controlled transition probabilities by varying the relative concentration of software molecules encoding for competing transitions (PNAS, in press). Recently, we augmented the stochastic molecular automaton with an input module, capable of sensing molecular disease symptoms, and an output module, capable of releasing a drug molecule in a controlled fashion. We programmed the computer to sense and diagnose molecular symptoms for prostate cancer and for small-cell lung cancer and to release, upon diagnosis, a molecule purported to be a drug for cancer.

 

Recent Publications

 

Personal Web Page