The Ion Bus on the beach at AGBT Marco Island FL
This week the annual Advances in Genome Biology and Technology (AGBT) meeting is being held, as it usually is, in Marco Island Florida. As a Gulf Coast resort area complete with white sand beaches and thatched hut shade, Marco Island every February has been one of the ‘must attend’ conferences for those who want to know what the current leading-edge research techniques and methods are using next-generation sequencing. And for the vendors, it is a non-stop week of activity. (One person on Twitter called it the ‘Detroit Auto Show of Genomics’.) Organized by none other than Eric Green (now Director of the NHGRI) and Elaine Mardis (Washington University School of Medicine) and in its 14th year, the approximately 800 attendees come from all over the world, and many of them represent some of the most productive and creative research laboratories in genomics. Due to its popularity (two years ago the online registration website filled up within a single day, and last year the registration website closed minutes within opening), this year there was an application process due to the limited facilities here in Marco Island.
Not that the conference site is insufficient in quality, it is just a matter of accommodations in terms of meeting spaces. Marco Island isn’t a major city such as San Francisco or Boston or even Pittsburgh with a major convention center; it has a Marriott resort hotel with enough room to handle a certain number of attendees, and about 800 is it. And for historical reasons, meeting in the same sunny location in the middle of winter is appreciated by many who come from much colder locales.
Every year this meeting showcases advances in technology around sequencing, every year the life sciences vendors who sell into this market view AGBT as a key tradeshow to make an impact on the market. Whatever the offering for the entire NGS workflow, you can likely find some representation here. As an example, target selection includes Agilent’s SureSelect™ or their HaloPlex™ or 454/Roche’s SeqCap EZ™ or RainDance Technologies, and newer offerings such as Kailos Genetics TargetRich™. The venue is different in that the limited space is used up by a poster session area, and a main conference room that is broken up into smaller rooms for the parallel-track sessions; there is no dedicated large exhibit area. However there are a number of small private rooms just off and around these two larger spaces, where the vendors set up areas to market their offerings.
Inside these suites it is a different atmosphere than the traditional trade-show format; couches and other seating is not unusual, along with catering and lighting to create an atmosphere. And vendors take advantage of the scheduled breaks to offer scientific presentations of their own, showcasing their latest technologies. (Of course there are presentation sessions in the conference for this technology should they be selected, but that may not always be the case for a particular vendor or new product.)
This year Life Technologies lined up four speakers: Joe Boland from Stephen Chanock’s group at the National Cancer Institute’s Advanced Technology Center in Gaithersburg MD, talking about their exome and RNA-Seq results using the Ion Proton; Dr. Marcel Nelen from the Nijmegen Medical Center, Netherlands about BRCA1/2 validation using AmpliSeq; Dr. Audrey Papp from the Ohio State University Center for Pharmacogenomics about using AmpliSeq™ RNA to measure allele-specific gene expression; and Dr. Jonathan Rothberg founder of Ion Torrent talking about the latest Ion Proton advances.
The first three topics may be for another day, but for the Ion Torrent Proton sequencer and series of Proton chip (PI, the upcoming PII and PIII), a lot of news.
For an overview, it includes: 150K runs on PGM to-date, due to the fast run-times and ability to iterate quickly; AmpliSeq multiplexing now extended to over 6K amplicons as a commercial product; 91 publications; 36 Torrent Suite pipeline plugins.
Dr. Rothberg showed a PI chip run with a mode on the order of 230 bases or so, which with a great loading percentage (the entire chip looked red/orange) equated to about a 15GB run. (This is internal data, as existing customers are obtaining from 7-10GB, although Joe Boland also referenced a 15GB run in his talk; it may well be an experimental pilot protocol and/or other changes.) There was also a comparative electron micrograph of the relative well sizes, from the existing 314/316/318 chips with the same 3+ um well size, to the PI well size (1.68 um pitch), to the PII well size (0.84 um).
About the PII, a perfect 190+ bp read was shown. (For those not familiar with the idea, the longest read with zero errors against the reference, even if it is only one read, is the ‘leading edge’ indicator of the quality of the platform. Since the average mode readlength will lag, along with the lag-time to get higher quality R&D reagents and processes down to the kit level, the expectation is that the mode readlength will be on the order of about 1/2 the perfect read length as a guide.)
The PI chip, now shipping (and over 12K Proton runs to-date), will be supplemented with a PII chip in mid-2013, with a goal of 32GB and 100bp readlengths. At a ‘to be announced’ date in the future, the PIII chip will launch at 64GB, with the PII increasing in throughput and readlength to yield 64GB at that point. He showed a progressive chart that had PI throughput increasing when the PII chip comes out, and the PII chip increasing when the PIII chip comes out, but alas I wasn’t able to capture the entire slide.
He also mentioned that the ‘cross-over’ point for Ion Torrent would be in 2013, in terms of throughput per day. He charted out the 454 trajectory from 2005 onward, with Solexa crossing over the 454 throughput in 2008, and then overlaying the Ion Torrent throughput with the cross-over point in 2013.
Lastly, a brief update on Avalanche was shared. (Avalanche was announced last September at Ion World, which was an emulsion PCR-free method of isothermal amplification.) Using beads, a 597 perfect read was obtained on a PGM chip in about 1.5 hours, and Avalanche can now be performed without beads. Then showed an almost 200bp perfect read with this new bead-free method.
(By ‘bead-free’, he did not elaborate, but it might well be the case to do direct amplification on the chip itself, or some other method other than beads to fit the small feature size that PIII is expected to have.)
So I’m off to get going on another busy day in Marco Island. Every year I’m here, there is no time for lying around on the beach, but that doesn’t matter too much for me. (Perhaps it was growing up in Santa Monica that has has already given me my lifetime’s worth of lying around on a beach, not sure about that!)