GATC: in computing, it spells slow

18 July 2008

Datamonitor analyst Ruchi Mallya has taken a quick look at the production of the world's first strands of DNA of reasonable length that use artificial molecular groups in place of the guanine, adenine, thymine and cytosine groups found in the natural stuff. The piece asks: is artificial DNA the future of computers? Jack Schofield at the Guardian asks, naturally, is it going to be the case?

I have a short and simple answer. No. Not even close.

You can use DNA for computation but you wouldn't use it to replace any existing form of computer. It's just too darn slow. And there does not seem to be a realistic way of making logic circuits using DNA that even approach the complexity of today's silicon-based machines, let alone computers in 20 years' time.

The group that has arguably done the most work on DNA computing is at Caltech. I've seen Georg Seelig talk a couple of times on the topic and he is realistic about the potential uses for the technology.

"What is realistic is a few thousand components. We won't get to having millions of components in the same test tube," said Seelig at a recent meeting at the Royal Society.

The processes involved in coercing DNA to compute things are orders of magnitude slower than those in modern computers. If you look at Seelig's 2006 Science paper, you will notice that the graphs are marked in hours. It takes at least an hour for the DNA equivalent of one logic gate to fully switch from one state to another.

People tend to look at DNA as being just a code in which the only important aspect of the double helix is the order in which the four bases appear. And so, it's an easy leap from the concept of code to that of a computer. There is a growing body of evidence that the shapes that DNA adopts play a major role in the polymer's behaviour. This should not be a surprise: shape plays a vital role in reactions mediated by enzymes.

In chromosomes, DNA sequences that are rich in AT pairs are often the sites of 'promoters' – lengths of DNA that unzip more readily than others. This gives the enzymes that transcribe DNA a toehold from which they can crawl along the DNA that makes up each gene.

AT-rich sequences do not just unzip easily. They are bent in a way that makes it easier for the transcription enzyme to stick in some cases but prevent this process in others. Sometimes, the DNA is bent to prevent transcription taking place and it is only when other enzymes attach and alter the shape of the DNA that transcription can get under way. It is these subtle changes that seem to give DNA its versatility.

Seelig has found that by tuning the DNA sequences, it is possible to warp the DNA into complex shapes that react faster. But you are still talking about processes that are at least one thousand times slower than those in silicon.

So, if it's slow and nowhere near as scalable as the silicon transistor, what is DNA computing good for? Doing stuff in living cells. This is why people are investigating artificial DNA. They want something that won't poison a cell but which won't be treated by the cell in the same way as regular DNA. You don't want your smart drug becoming a kind of virus.

The idea is that you could introduce the artificial DNA into cells and have it work out what type of cell it is in. The DNA that forms the 'computer' might latch onto part of a gene or a piece of RNA - which can form double-stranded helices with matching DNA. That process might prevent a DNA logic gate from closing, or activate one and ultimately release a drug or a marker attached to another piece of the DNA computer.

Another reason for using artificial DNA is that it might stand a better chance of escaping destruction inside the cell. Every cell has enzymes designed to break down DNA – it's a good defence mechanism against viruses. They look for certain sequences of bases and then break the DNA at those points. Cells protect their own DNA by tacking on chemical groups to the vulnerable links.

If you want to put the DNA for a smart drug into a cell, you either need to protect the DNA in the same way, which complicates production of the artificial DNA. Or you can use DNA that the restriction enzymes do not recognise. That's why artificial DNA may turn out to be better.

Masahiko Inouye of the University of Toyama said the artificial DNA they have produced should resist attack by restriction enzymes but they have yet to test this. So far, what they have done is design bases analogous to the natural versions that have similar shapes and properties to see how sensitive DNA structure is to these kind of changes. The artificial DNA behaves very similarly to the normal stuff, which suggests that the sensitivity is quite low.

In the UK, Philipp Holliger at the Medical Research Council in Cambridge has made fluorescent green DNA using fluorophores tacked onto the natural bases. These molecular groups are pretty big but even with a fairly high ratio of modified bases to the normal ones, the DNA still formed a double helix, albeit a lot fatter than the one that Watson, Crick and Franklin uncovered.

Holliger wants to go further. "The hope is that we will not be limited to four nucleic acids in our alphabet in the near future," he said.

Artificial DNA has one potential use that is at the border between computation and materials science. Because it is possible to coax DNA into interesting shapes, it could be used to design new materials with complex internal structures. You might use unnatural bases in this structuring DNA to bind to the core material – naked DNA is surprisingly brittle.