Why it’s dangerous to liken DNA to computer code

Watson and Crick with a model of DNA in 1953.

Watson and Crick with a model of DNA, the code of biological life, in 1953. Photograph: A Barrington Brown/Science Photo Library

Afew days ago, on my way to a discussion in the exquisite little McCrum theatre, which is hidden away in the centre of Cambridge, I had to pass through the courtyard of the Eagle pub in Bene’t Street. As I did so, I suddenly remembered that this is the hostelry where, on 28 February 1953, Francis Crick, rushing in from the nearby Cavendish Lab, announced to astonished lunchers that he and James Watson had discovered the secret of life. (They had just unveiled their double-helix model of the DNA molecule to colleagues in the laboratory; there’s now a blue plaque on the wall marking the moment.)

As a graduate student in the late 1960s, I knew the pub well because it was where some of my geeky friends from the Computer Lab, then located in the centre of town, used to gather. We talked about a lot of things then, but one thing that never really crossed our minds was that there might be a connection between what Crick and Watson had done in 1953 and the software that many of us were struggling to write for our experiments or dissertations.

With the 20/20 vision of hindsight, though, the connection should have been obvious. The implication of the Crick-Watson discovery was that the genes and genomes that underpin organic life are based on a code written in four letters – C (for cytosine), G (guanine), A (adenine) and T (thymine). Organisms are basically biological machines built by executing programs written in these letters. Just like computers, in other words, which are just machines that execute programs written in ones and zeroes.

The implications of this analogy are mind-blowing. Once you have figured out the sequence that programs an organism, then, in theory, you could replicate it. And if you can write biological code then you can edit it to change the organism or even create an entirely new one. You can, in other words, become a biological programmer and play God.

As it happens, people have been doing this for a while. In 2010, Craig Venter’s team wrote the genome of a bacterium from scratch and incorporated it into a cell to make what they called the world’s first synthetic life form. In 2012, researchers came up with Crispr-Cas9, a tool that appears to offer cut-and-paste facilities for precise gene editing. And earlier this year, researchers at the Medical Research Council’s Laboratory of Molecular Biology in Cambridge redesigned the DNA of the E coli bacterium and then created cells with a synthetic version of the altered genome.

It looks therefore as though we are on course for an explosion in bioengineering, the biological equivalent of what we now call software engineering. The “visionaries” (ie billionaire fantasists) of Silicon Valley are mad keen on it, perhaps because they think it will make death optional (for them). After all, if life itself is just code, well then, that’s what the Valley does. So let’s get on with it.

Before launching a toboggan down that slippery slope, though, it’s worth delving a bit further into what the analogy between genetic code and computer software really implies. This is what the computer security guru Bruce Schneier and the biotechnology expert Larisa Rudenko have done in a remarkable article published earlier this month. If genetic programming really is like computer programming, they argue, then that should ring alarm bells everywhere. “Programmers write software through trial and error. Because computer systems are so complex and there is no real theory of software, programmers repeatedly test the code they write until it works properly. This makes sense, because both the cost of getting it wrong and the ease of trying again is so low.” But all “finished” software is still riddled with bugs, which is why it requires continual updates and patches.

The trouble is that what is standard practice in software won’t work in bioengineering. Biological code, like computer code, will inevitably have programming bugs in it for the simple reason that we’ve never found a way of writing bug-free code of any complexity. As Schneier and Rudenko warn: “Unlike computer software, there’s no way so far to ‘patch’ biological systems once released to the wild, although researchers are trying to develop one. Nor are there ways to ‘patch’ the humans (or animals or crops) susceptible to such agents. Stringent biocontainment helps, but no containment system provides zero risk.”

In other words, choosing to play God might not be a smart career move for mere humans.

What I’m reading

Spy wary
Robert Heaton has a terrific blog post about what he discovered when installing a Hewlett-Packard printer. Headline: HP printers try to send back to HQ all kinds of data about what you do with them.

Near miss
A great Wired story about how wifi nearly didn’t happen. A useful refutation of technological determinism and an illustration of the importance of technical standards.

Going it alone
A fascinating Jalopnik website report about how Google concluded that full autonomy is safer than driver-assisted technology.

[“source=theguardian”]

Related posts