When you think about it, computer science and biology have a lot in common.
Evolutionary algorithms in artificial intelligence take inspiration from evolution; from a ‘generation’ of potential solutions to a problem, for example, the best potential solutions are given the highest chance of survival and ‘reproduction’—the creation of a new solution made up of parts of its parent solutions. ‘Point mutations’ occur when small changes are made to a potential solution, and there is even ‘crossing-over’ and ‘recombination’ of these potential solutions.
Similarly, neural networks borrow from the structure and function of brains, where nodes are neurons and connections between nodes the synapses between neurons.
It works the other way too. In a genetics lecture last week, our lecturer invited the class to “consider the ‘computer analogy’” where genes, encoded in DNA by its four base (A, C, G, T) code, are the inputs to cellular machines which output RNA and proteins.*
There’s even a whole field where biology meets computer science, called bioinformatics— the computational analysis of DNA sequences to find genes in DNA, compare how similar different sequences are, and predict the structure and function of gene products.
And so if biology and computer science have so much in common, perhaps it’s not so surprising that scientists have created basic computers from biochemicals, and are even planning to make robots from biological organisms.
Maybe not surprising, but undeniably, and indisputably Pretty. Damn. Cool.
Over the past 10 to 15 years, biologists, chemists, and computer scientists have combined their powers to make something even cooler than Captain Planet.
In an early experiment in 1994, University of Southern California scientist Leonard Adleman solved instances of the Hamiltonian path problem (a graph problem in discrete maths) by encoding all possible solutions in DNA molecules, and isolating the correct solution using a series of reactions.
More recently, Serbian scientists Darko Stefanovic and Milan Stojanovic built a basic computer out of DNA to play tic-tac-toe (or noughts and crosses as we call it in Her Majesty’s Commonwealth).
Stefanovic and Stojanovic built their computer from DNA logic gates.
Logic gates are the fundamental building blocks of computers, arranged in more and more complex ways to allow more and more complex operations. Logic gates take 1s and 0s as inputs and give 1s and 0s as outputs, according to Boolean logic. So, an AND gate, for example, will output 1 given two 1s as input, and output 0 given anything else (“A and B” is true if and only if “A” is true and “B” is true).
Regular computers take electronic signals as inputs and output electronic signals. Scientists have realised that DNA can do something similar.
Stefanovic and Stojanovic’s used a single strand of DNA (an “acceptor”) with a “substrate binding region” to which another strand of DNA can bind. The scientists added a chemical to this binding strand so that when it attached to the acceptor, resulting complex would glow.
To make the acceptor into a logic gate, they attached another piece of DNA which they called a “sensor gate”, which closed off the acceptor to binding strands. This sensor gate can be opened in the presence of particular chemicals. So, to make an AND gate, for example, the scientists added two sensor gates to an acceptor. In the presence of chemical A AND chemical B, the gates open up, the binding strand can attach to the acceptor, and the DNA glows.
Stefanovic and Stojanovic used DNA logic gates to play a game of noughts and crosses by filling nine wells, representing the nine squares of the noughts-and-crosses board, with a bunch of these logic gates. To play, a human opponent would drip a chemical representing the particular square they were choosing (say, square 7) into each of the wells. Square 7 glows green. Say the player has already chosen squares 4 and 9, and the computer squares 1, 3, and 5. In the second well the computer’s 7-AND-9-AND-NOT-1 logic gate would open up, a binding strand would bind and the DNA would glow red—the computer wins with squares 1, 2 and 3!
It’s not just computers made of DNA that have been keeping scientists busy. Scientists at the University of the West of England recently won funding to engineer robots from single-celled organisms Physarum polycephalum—slime mould.
As reported in the New Scientist magazine, the organisms will be “‘programmed’ using light and electromagnetic stimuli to trigger chemical reactions similar to a complex piece of chemistry called the Belousov-Zhabotinsky reaction, which [lead researcher Adam] Adamatzky previously used to build liquid logic gates for a synthetic brain”.
In a world in which many of the fundamental limits of our current computing technology are being reached (and the fossil fuels that power them are dwindling), advances that have the potential to fundamentally change what we think a computer is, are not only fascinating, but extremely important.
And hey, if we are eventually overrun by sentient colonies of robotic slime mould, at least we know we’ll be leaving the Earth to a life-form that’s Pretty. Freaking. Amazing.
*Lest I be accused of plagiarism, the lecturer was Geoff Chambers, and the paper Biol 241 (which is awesome, you should take it).