It is believed that DNA will save us from computers. With advances in silicon transistor replacement, DNA computers promise to provide us with massive parallel computing architectures not currently possible. But here's the catch: the molecular microcircuits that have been invented until now have had absolutely no flexibility. Today, using DNA to compute is like “building a new computer out of new hardware to run a single program,” says scientist David Doty.
Doty, a professor at the University of California, Davis, and his colleagues decided to find out what it would take to build a DNA computer that could actually be reprogrammed.
DNA computer
In a paper published this week in the journal Nature, Doty and colleagues at the University of California and Maynooth demonstrated just that. They showed that a simple trigger can be used to force the same basic set of DNA molecules to implement many different algorithms. While this research is still research in nature, reprogrammable molecular algorithms could be used in the future to program DNA robots that have already successfully delivered drugs to cancer cells.
In electronic computers like the one you use to read this article, bits are binary units of information that tell the computer what to do. They represent the discrete physical state of the underlying equipment, usually in the presence or absence of electrical current. These bits - or even the electrical signals that implement them - are transmitted through circuits made up of gates that perform an operation on one or more input bits and provide one bit as an output.
By combining these simple building blocks over and over again, computers can run surprisingly complex programs. The idea behind DNA computing is to replace electrical signals with nucleic acids - silicon - with chemical bonds, and create biomolecular software. According to Eric Winfrey, a computer scientist at Caltech and co-author of the work, molecular algorithms use the natural information processing ability embedded in DNA, but instead of giving control to nature, "the growth process is controlled by computers."
Promotional video:
Over the past 20 years, several experiments have used molecular algorithms for things like playing tic-tac-toe or assembling various shapes. In each of these cases, the DNA sequences had to be carefully designed to create one particular algorithm that would generate the DNA structure. What is different in this case is that researchers have developed a system in which the same basic DNA fragments can be ordered to create completely different algorithms and, therefore, completely different end products.
This process begins with DNA origami, a method of folding a long piece of DNA into a desired shape. This rolled-up piece of DNA serves as a "seed" (seed), which starts an algorithmic conveyor, just as caramel gradually grows on a string dipped in sugar water. The seed remains largely the same regardless of the algorithm, and changes are made in only a few small sequences for each new experiment.
After the scientists created the seed, they added it to a solution of 100 other DNA strands, DNA fragments. These fragments, each of which consists of a unique arrangement of 42 nucleic bases (the four main biological compounds that make up DNA), are taken from a large collection of 355 DNA fragments created by scientists. To create a different algorithm, scientists must choose a different set of starting fragments. A molecular algorithm involving random walk requires different sets of DNA fragments that the algorithm uses to count. As these pieces of DNA join together during assembly, they form a circuit that implements the chosen molecular algorithm on the input bits provided by the seed.
Using this system, scientists created 21 different algorithms that can perform tasks such as recognizing multiples of three, choosing a leader, generating patterns, and counting to 63. All of these algorithms were implemented using different combinations of the same 355 DNA fragments.
Of course, writing code by dropping DNA fragments into a test tube will not work yet, but this whole idea represents a model for future iterations of flexible computers based on DNA. If Doty, Winfrey, and Woods get their way, the molecular programmers of tomorrow will not even think about the biomechanics underlying their programs in the same way that modern programmers do not need to understand the physics of transistors to write good software.
The potential uses for this nanoscale assembly technique are staggering, but these predictions are based on our relatively limited understanding of the nanoscale world. Alan Turing could not predict the emergence of the Internet, so there may be some incomprehensible applications of molecular informatics.
Ilya Khel
