Published online 20 November 2008 | Nature | doi:10.1038/news.2008.1245


Human genomes in minutes?

Not yet, but biotechnology company is on track for 2013.

A much-watched Silicon Valley company has published a proof of concept for its DNA-sequencing technology, which it hopes will one day be going through entire human genomes in minutes.

Pacific Biosciences of Menlo Park, California, reports today that it can generate continuous stretches of DNA sequence up to thousands of base pairs long. It also reports sequencing at an average speed of 4.7 bases per second1. 'Second-generation' sequencing machines can deliver anywhere from 2 to 30,000 bases per second.

Sequencing using polymeraseDNA polymerase incorporates labelled nucleotides, while a digital camera records the sequence through the bottom of the well.Stephen W. Turner

But the paper, published in Science, also shows the company still has a long way to go until it can ship its first machines to customers. That isn't expected until 2010, and genomes in minutes will have to wait until about 2013. For instance, the company reports a relatively high error rate — in one experiment, misidentifying 27 bases of a 150-base-pair piece of DNA, an error rate of about 20%. And the company has not reported how reliably it can generate the long reads that are a major selling point of its technology.

Rubbing out errors

"This is a tour de force of technology," says Tim Harris, who was a senior research director at Helicos Biosciences, one of Pacific Bioscience's rivals, based in Cambridge, Massachusetts. However, Harris says, "the main issues of this paper are [that] they failed to demonstrate long read length, and the error rates are pretty high". Harris is now at the Howard Hughes Medical Institute's Janelia Farm Research Campus in Ashburn, Virgina.

But, Steven Turner, cofounder and chief technology officer of Pacific Biosciences, says that the company has already surpassed the performance it is publishing today. At a conference in Cold Spring Harbor, New York, last month, for instance, Turner reported that the company had sequenced the genome of a bacteriophage — a virus that infects bacteria — and had improved its error rates (see Accessible genomes move closer). Turner says the error rates will continue to drop as the company perfects its chemistry and detection technologies.


The paper establishes that the company is capable of reading out thousands of bases of sequence in one go, says Turner. This is good news for genome scientists, because assembling shorter fragments of DNA read out by current second-generation technologies takes time and computing power and introduces possible sources of error.

"We are unaware of any sequencing technology in existence today that can go out to as long as we've demonstrated in this paper," says Turner. Second-generation machines, by comparison, are delivering read lengths of at most hundreds of base pairs long.

Less than an hour

Pacific Biosciences's technology is based on DNA polymerase – the enzyme that helps builds the DNA double helix in the body. The company has devised a way to tether individual DNA polymerase enzymes to the bottom of tiny wells, which it calls zero-mode waveguides. DNA templates are loaded into the wells, which are then washed with labelled DNA bases. A digital camera records the sequence of these labelled bases as they are incorporated into the growing DNA strand by the DNA polymerase.

By packing as many as a million zero-mode waveguides on to a chip and making other improvements, Turner says that the company will be able to minimize errors and match the throughput of current second-generation technologies when it begins shipping instruments in 2010, while delivering longer sequence reads.

And the machine will only get better after that, Turner says: "It will be at the very beginning of its Moore's law trajectory of improvement [the exponential improvement of technology over time], and that means within the five-year time frame, we'll be able to sequence the entire human genome in less than an hour." 

  • References

    1. Eid, J. et al. Science advance online publication, doi:10.1126/science.1162986 (20 November 2008).
Commenting is now closed.