The main theme of this year's Advances in Genome Biology and Technology meeting should come as no surprise to regular readers: sequencing. Generating as many bases of DNA sequence as quickly, cheaply and accurately as possible is the goal of the moment, and the number of companies jostling to achieve that goal is growing rapidly.
The meeting saw impressive performances from established players in the field, especially Illumina: their new HiSeq 2000 instrument seems to have dug in as the platform of choice for generating vast amounts of high-quality short-read data. Life Technologies seem to be slowly abandoning the research genomics market (already dominated by Illumina) with their SOLiD platform, focusing instead on capturing the clinical sequencing market; they showed some impressive accuracy improvements for their technology.
As I mentioned in my previous post, PacBio largely underwhelmed the audience with their theatrical unveiling of a massive box with quite limited applications, although we'll have to wait and see how much its specifications improve over the next couple of years. Meanwhile, Complete Genomics gave an understated but seriously impressive series of presentations on their human genome sequencing service; I'll have more on them in a day or two.
Anyway, in this post I want to focus on the two brand new platforms announced in the emerging technologies session on the last day of the conference: the newcomer Ion Torrent, and Life Technologies' futuristic quantum dot technology.
Ion Torrent: quick, cheap, low-volume next-gen sequencing
One of the bigger surprises at this year's meeting was the emergence of yet another player in the next-generation sequencing market: Ion Torrent. The company seems to have been operating in stealth mode for some time, and hit the meeting with a functioning instrument (which was happily displayed to a steady stream of conference attendees during the meeting).
The fundamental idea behind Ion Torrent is pretty cool: it relies on the fact that as nucleotides are added to a growing DNA strand there is a release of hydrogen ions. The platform immobilises DNA strands in tiny wells within a semiconductor chip, and then washes As, Cs, Gs and Ts one by one over the wells. As each base is incorporated it releases hydrogen ions that can be detected as they pass through a pore at the base of each well.
The downside is that the technology is susceptible to the same kind of homopolymer stretch problem that plagued the 454 second-gen technology: if there are seven Gs in a row in your DNA strand, all will be incorporated at once - and distinguishing the signal corresponding to 6 Gs vs that produced by 7 Gs will be very challenging. Rothberg showed data suggesting that it was possible to accurately read across homopolymers up to 6 bases long, but it's unclear how well it will do for longer stretches.
The instrument itself is extremely compact compared to most other next-gen machines (especially compared to the utterly mammoth Pacific Biosciences instrument), and both the machine and runs will be cheap: $50,000 to get the box, and just $500 per run for reagents and sample prep. Initially, reads will be 100 bp long. The yield (150 Mb per run) is not yet competitive with other second-gen instruments.
While this is not yet a technology that will enable the production of cheap human genomes, it may well turn out to be a useful machine for quick-and-dirty assays - for example, checking out the quality of DNA preparations before investing in expensive sequencing on a more high-throughput technology.
Ion Torrent got an extremely positive buzz throughout the meeting, although this was dampened somewhat by a fairly unimpressive presentation by Rothberg. I got the impression that given the low cost of the machine and its potential applications there will be plenty of potential customers lining up.
Life Technologies' new platform: quantum dot single-molecule sequencing
Life Technologies' Joseph Beechem debuted a brand new single-molecule sequencing platform developed by the company that provides the second-gen SOLiD platform
The physics underlying the system is above my pay-grade, but basically involves tethering quantum dot nanocrystals to DNA polymerase molecules. These dots have the effect of amplifying the signal resulting from the addition of a fluorescent nucleotide to a growing DNA strand, allowing signals to be read from a single DNA molecule.
The most shocking thing about the technology is that in theory it can be used to generate reads of unlimited length. Beechem showed that the run could be interrupted mid-way through to wash off existing polymerase molecules and replace them with new ones, thus replacing any molecules that have been inactivated by chemical damage. After the replacement the strand continues to grow; that means that in theory the wash-and-replace process could be repeated over and over to continue reading each DNA strand until you run out of molecule. Beechem suggested that by instrument release each cycle would probably generate 1,000-1,500 bases, which is an impressive achievement in itself.
If that's true, and if it can be done at scale, it is extraordinarily cool: reads of unlimited length would profoundly transform genomics. However, we're yet to hear the hard numbers on error rates and throughput that are required to fully evaluate the promise of the system.
Instruments will apparently be available to early collaborators by the end of the year (far sooner than I would have expected given the science-fiction feel to the technology), so it sounds as though we'll be getting these numbers in the not-too-distant future.