Kevin Davies spoke with Jonathan Rothberg, the founder of 454 Life Sciences, on the eve of the publication of his company's landmark paper in Nature. He discussed the progress and potential impact of 454's revolutionary 'sequencing by synthesis' technology.

Rothberg founded 454 Life Sciences in 2000, while he was CEO, president, and chairman of CuraGen, a genomics-based pharmaceutical company. His inspiration had come a year earlier, when his newborn son was rushed into intensive care, leaving Rothberg to ponder ideas for accelerating genome sequencing and personalised medicine. He is also the founder of The Rothberg Institute for Childhood Diseases (TRI), a non-profit research institute dedicated to developing cures for orphan diseases. With more than 100,000 registered users in 101 countries, TRI's distributed computing effort is one of the largest dedicated computational chemistry efforts in the world.

How would you characterise the advance represented by 454's sequencing technology?

It's completely analogous to personal computers displacing mainframes. Now, anyone can have their own genome centre. If you can miniaturise something, then everything gets cheaper and faster.

Since Fred Sanger won the Nobel Prize in 1980, it's been a race to find a new way to sequence [DNA] cheaper and faster. There's never been any technology to pull that off. We've pushed Sanger electrophoresis as far as you can go, but it has two limitations: the amount of parallel sequencing you can do, and the amount of money and robotics to prepare the samples. Because of that, the technology is at a wall. The night my son was born [in July 1999], he was rushed into intensive care, and that night I was up all night. I said, "Why can't we just sequence his genome and know everything is fine or not fine?" I had just read that Intel had this new chip with 44 million transistors, so I recalled how vacuum tubes hit a wall. They'd burn out fast. In order to make things better, they have to take a different path -- they had to go to the transistor. They don't have the power of vacuum tubes, but they're cheaper and better. Everything's been trying to make sequencing better, so let's step back and make the transistor equivalent.

As for sequencing by synthesis -- they'd thrown it away, because it would only have enough power to sequence a few bases. What if you do lots in parallel? It was outside the box. Once I got it miniaturised, four of those wells will fit on the end of a single hair. Every time you make the wells smaller, you get more sequences for the same price.

What does the paper in Nature signify, in your own words?

In this paper, we describe sequencing 100 bases in each well, 400,000 wells at a time. We've also done 200 bases in a well. We've put more wells per square millimetre. We've sequenced the bacterial genome at 50X coverage. We didn't have to spend months preparing the sample, we can prepare any sample, even human, very quickly.

We take a nebuliser, and nebulise the DNA into little fragments, shake it in oil and water, so each DNA fragment goes into a separate water droplet. So instead of bacteria, we separate the DNA into drops, do PCR, so every drop has 10 million copies. Then we put in a bead, drive the DNA to the bead, so instead of the cloning and robots, one person can prepare any genome.

I think Nature published the paper because they will publish technology if they think it will transform science, if you can do something 100 times faster. People are using our instruments for bioterror, third-world drugs, etc. We've already had talks on sequencing HIV in a person, we can see if someone's resistant long before using standard sequencing. There's no more waiting in lines at genome centres. Anything you want, you can sequence pretty instantly.

Where is there most room for improvement? Accuracy? Read lengths?

We're achieving Bermuda standards (99.9 per cent accuracy). It works beautifully, we just needed a worldwide distribution network [454 just signed a licensing deal with Roche]. On the technology side, a priority is getting more sequences from each of the wells. We've done 200 bases at the same accuracy. And even 400 bases. We also want to get more sequences in a single run. We guarantee customers 20 million bases in a single run [on the instrument] -- we look to raise that to more than 100 million in the next few years.

What would it cost approximately to sequence a human genome using your current platform?

Everyone wants human genomes to be inexpensive. [Francis] Collins gave us $5 million to do a $100,000 genome. That's the next challenge. When someone goes to the hospital, you'll sequence the genome. Right now, it would cost under $1 million.

Where is the 454 platform being tested?

Eric Lander (the Broad Institute) has one, the Sanger Institute has one, JGI (Joint Genome Institute), Baylor, and the Venter Institute. They all paid list price -- US$500,000. We've sold a number of them. We're selling them as fast as we can physically make them. We just signed a deal with Roche for $60 million. We'll share the profits.

What's the feedback like from those institutes?

The major challenge is being able to integrate our data with their sequencing data, supplying software tools to make sure it's fully compatible. How do you span [DNA] repeats? The repeat problem is identical to Sanger capillary sequences. It's identical -- we have to have a kit that does paired-end sequencing.

How does 454's progress compare to other sequencing companies?

I want to say it's been a 25-year race, we're commercial, and we won the race... The reason we signed with Roche is that only the paranoid survive. Right now we won, nobody else's machine is really working. We're out there... my real worry is the same thing that [Andy] Grove and [Robert] Noyce worried about -- it's someone in their basement you don't know about. We want to get the instrument distributed as fast as possible. Over 50 per cent of our sales are to non-genome centres.

What is 454's relationship to its parent company CuraGen?

They operate as two separate companies. CuraGen is 100 per cent focused on drug development. We have three drugs in the clinic, and two more going into man next year. But CuraGen has made a successful transition from genomics to a genomic pharmaceutical company. As for 454, I'm the founder and chairman, and we've announced that it's cash-flow positive. It's run as an independent company, and some day it'll probably explore the public markets. We spent $50 million and 100 people to build this company and make sure this works.

It has been reported that you've offered to sequence James Watson's genome. Care to comment?

Hah! Jim is a visionary, and I think he's a bright guy. He's a huge believer and a strong supporter of our technology. That's all I can say...