February 11, 2011
Kevin Davies, Editor-in-Chief, Bio-IT World : MARCO ISLAND, FL - The annual Advances in Genome Biology and Technology (AGBT) conference remains the premier destination to witness the latest technologies and applications of next-gen sequencing (NGS). This year’s meeting, expanded by some 200 people to accommodate some of the overwhelming demand (several registrants from Boston and elsewhere didn’t make it), didn’t quite live up to the razzamatazz of recent years. But the absence of any brand new third- (or fourth-?) generation NGS platforms allowed attendees to focus more on some excellent science and range of applications.
Admirably comprehensive, real-time blog coverage (where allowed) was provided by an army of Twitter aficionados and the tireless Anthony Fejes at Fejes.ca. There’s also a very good two-part review from Prithwish Pal at Omically Speaking.
What follows are a few of my own observations and takeaways:
- Ion Torrent (now part of Life Technologies), which debuted its $50,000 Personal Genome Machine at AGBT last year, received some timely validation from the Broad Institute’s Chad Nusbaum. Over past 12 months, the Broad has pushed the machine’s run output up to Megabases (Mb), with one use being DNA quality control for Illumina runs (probably not the market niche Ion is looking for). Ion says it is currently getting 300,000 reads of more than 100 bases/run, while the Broad is typically about two weeks behind them. Improving the quality of the reads while halving the sample prep time are two key priorities for Ion this year, for which it is looking to the NGS community for help willing to reward handsomely.
- Congrats also to Baylor’s Adam English for overhauling teams from Liverpool and Tufts to win a free PGM: http://twitpic.com/3wm67t# It’s (relatively) easy to give away a machine when it costs a tenth of the competition. Ion also offered yoga classes at 7 each morning, a nifty and original marketing ploy that I happily slept through.
- Life Technology’s 2nd-gen SOLiD platform was overshadowed by Illumina in a procession of talks on opening day from some of the major NGS labs, but is pushing cancer genome applications as one of its most attractive applications. That is evident from its latest promotional campaign and the validation from speakers such as TGen’s David Craig, who spoke about SOLiD’s “sweet spot” in analyzing cancer genomes because of its accuracy, approaching 99.9%.
- PacBio screened its documentary, “The New Biology,” which first aired at the Air & Space Museum in Washington DC last November. It’s an impressive film, though I did miss Charlie Rose and the roundtable. Opening PacBio’s commercial workshop, co-founder Steve Turner creditably addressed the accuracy rate, which is hovering at 85% (c.f. 99%+ for more established platforms). While there is clearly room for improvement, some of which may come from incorporating data from the kinetics of nucleotide incorporation, there is impressive uniformity of sequence accuracy across a wide spectrum of GC content, in stark contrast to some other platforms. In the closing talk, CSO Eric Schadt discussed PacBio’s recent cholera genome paper in the New England Journal and “connecting molecular biology to clinical medicine.” There were also early signs of the value of integrating PacBio’s long reads with the more abundant short-read coverage of an Illumina for de novo genome assemblies.
- Illumina representatives brought along the new MiSeq machine, which will be officially released later this year. CSO David Bentley said the HiSeq 2000 (which had a launch spec of 200 Gb/run) has surpassed 1 Terabase/run in house. So much for Moore’s Law... There was also some promising early data on the genetic changes in the evolution of chronic lymphocytic leukemia (CLL). “This technology had the potential to direct future clinical trials and therapeutic decisions,” said Bentley.
- I thought one of the best talks was from Harvard Med School’s Tim Yu, who discussed the analysis of 40 genomes from Complete Genomics, selected from consanguineous families from the Middle East. Yu’s team was able to identify variations in five suspected autism genes, including potentially mild variants in a pair of known metabolic disease genes.
- My previously published interview with Complete CEO Cliff Reid made a bit of news. Reid says Complete will be producing somewhere between 800-1,200 human genomes by the end of 2011. The company is currently generating 400 genomes/month. Reid said Complete could expand its service to include transcriptome sequencing later this year, and hinted that expansion into Asia is another real possibility. I’m also struck that Lee Hood’s Institute of Systems Biology recently placed its third order with Complete, this one for 600 genome sequences.
- Oxford Nanopore unveiled its GridION “box” just days before AGBT, but none of the firm’s contingent attending the show spoke on the program, preferring it seems to lounge by the hotel pool. ONT’s scientific founder, Oxford chemistry professor Hagan Bayley, was a late scratch, but collaborator Mark Akeson (University California Santa Cruz) demonstrated some recent significant progress in single-strand DNA sequencing using nanopores coupled to DNA polymerase, which ratchets the strand through the pore. Akeson said that the system should be working for short reads, at least in an academic setting, later this year.
- Steve Salzberg (U Maryland) presented impressive benchmarking of Ben Langmead’s Bowtie DNA alignment software, which seems to outperform BGI’s SOAPdenovo and other established tools by a considerable margin. Bowtie 2 is on tap for release this spring, Salzberg said.
- Rob Knight (HHMI/U Colorado), with a quaint kiwi accent and sense of humor, presented an excellent talk on ‘the human microbiome’ – the results of metagenomic profiling of bacteria living in – and on – people, including detailed analyses of the different species living on different parts of our face, hands etc.
- Two notable absences from AGBT this year were representation from Helicos, the first single-molecule sequencing company, and BGI, the Chinese organization that last year launched a subsidiary called BGI Americas to offer access to its enormous sequencing resource in Hong Kong.
As usual, ABGT concluded with the organizers polling the audience for suggestions on the future size and scope of the conference. There was broad consensus that the size of the meeting was just about right. Suggestions for future topics can be sent to NHGRI’s Eric Green, Washington University’s Elaine Mardis, or any of the other organizers. I’d like to see more talks on genomic medicine and the data management aspects of NGS, but I doubt the organizers will have too much trouble attracting talks on those themes for next year.
In the meantime, you can get your next NGS fix at conferences including CHI’s X-GEN Congress in San Diego (March 14-18), Human Genome Meeting 2011 in Dubai (March 14-17), and Bio-IT World Conference & Expo in Boston (April 12-14).