AGBT 2017: A Big Small Meeting
Article Mar 03, 2017 | By Josh P. Roberts
Advances in Genome Biology and Technology (AGBT, known as the “Marco Island meeting”) is not the typical present-your-latest-scientific-findings-to-your-peers meeting. While the science is certainly there, it’s more about the technology behind that science that draws researchers down to south Florida every winter (that, and the climate). February’s AGBT 2017 – held across the peninsula in Hollywood Beach while its traditional home is being renovated – is “where people come to evaluate what’s new and happening in next generation sequencing (NGS)” and the like, explains Elaine Mardis, Co-Director, Institute for Genomic Medicine at Nationwide Children's Hospital, and AGBT 2017 co-chair. “It’s focused on the technology and how people are using it across the spectrum of research.”
Mardis describes AGBT as a private “big small meeting” – capped at 850 people (“we routinely turn away about 100 or so people because we don’t have enough room”). Oral and poster session abstracts are not available online, and authors can opt out of allowing their presentations to be tweeted about – which lets unpublished work and work-in-progress be freely discussed. Many authors themselves tweeted out their slide decks, though.
Attendees encounter what’s on the bleeding edge or in the pipeline, and what can be done with it. Small, pilot-like studies “to show what can be done” may be presented by sponsored scientists using tools introduced at a previous AGBT. And what’s being discussed here may likely be the standard, routine technology behind the large studies seen at general society meetings in a year or two, remarks Michael Talkowski, Director of the Genomics and Technology Core of the Massachusetts General Hospital. At the same time, there will be ready-to-go kits “for you to buy and do your science. It’s the whole spectrum.”
There is a hotel ballroom with large monitors set up, where researchers can demonstrate their software and its capabilities, Mardis explains. “That way they can address more than one or two people at a time.”
Rather than a traditional exhibit hall, vendors at AGBT hold court in private spaces – hotel suites or conference rooms -- where they can demonstrate products, hold workshops, and allow their R&D teams to interact with attendees one-on-one.
“It’s a unique opportunity to both get a very good swath of what the current technology in the genomics space is, and then also to have all these discussions around novel ideas you might have and the applications of the technology. And of course most vendors are delighted to have those conversations as a way of learning what their technology can be used for,” notes Martin Hirst, Associate Professor of Microbiology and Immunology at the University of British Columbia.
As an NGS meeting, of course Illumina loomed large at AGBT. Having debuted their NovaSeq platform to investors the previous month at the J.P. Morgan 35th Annual Healthcare Conference, the sequencing giant introduced NovaSeq to the people who will be using it, and using it to develop algorithms, reagents, peripherals, and other adjuncts. But NovaSeq, which essentially marries Illumina’s NextSeq to two-color chemistry, isn’t really revolutionary in that it just does more of what the company has been doing: generating huge amounts of sequence, in relatively short reads, at low cost and high quality – “they just generate a ton more of it,” notes Talkowski.
“They’ve engineered their approaches to give them more headroom to keep growing in the amount of capacity that they have,” added Paul Flicek, Head of Genes, Genomes and Variation Resources at European Bioinformatics Institute. “They’re nowhere near the end of what they can do.” That this has been building, and continues to build, over time is a hallmark of a maturing technology.
That a lot of the technology is maturing is a theme Flicek took away from the meeting. Things are getting more refined, with the ability to do larger datasets, faster workflows, and more accurate experimental preparations. We’re starting to see the impacts of very large datasets – “not only is it having all of these data together that you can see patterns that you haven’t been able to see before, but it’s the computational firepower as well that’s really required to do that.”
“PacBio, with their new chemistry, and now the Sequel machine, has been moving progressively over time. … 10X’s tech is solid – they’re continuing on a nice trajectory. … NanoString also – it’s clear that their tech is now working,” Flicek says. “10X basically requires the Illumina machines to do their stuff, and so it’s something to say that Illumina is stable enough that there are companies that exist around it.”
Another sign of a maturing field was the well-received comparison, by Cold Spring Harbor Laboratory’s Mike Schatz and his colleagues, of the various capabilities of long-read (PacBio and Oxford Nanopore) and synthetic long read (Illumina and 10X Genomics) technologies. “What his lab and other labs like it do is bioinformatics and computational biology – so the analytic methods, which are really of vital need for all of the new techs,” Talkowski points out. “That’s a major component of [AGBT], and what a lot of people presented there as well.”
Illumina, 10X, and PacBio were well-represented both as vendors and in scientific presentations. Oxford Nanopore was conspicuous in its absence as a sponsor, but there were reports about how the portable, diminutive sequencer was used on a bus and on the space station.
Hitachi kept their hat in the ring with proof-of-concept of its solid state nanopore sequencer. “This is one of the first new sequencing platforms or concepts that I’ve seen in a while that has preliminary data,” Hirst remarks. “The technology required to actually make this happen is quite a leap forward. That was a notable talk about a forward-looking tech.”
Make it a Single
Single cell RNA-seq and some other single cell analyses that have shown up in past meetings “probably are going to be substantially more important this meeting,” Robert Lyons, Director of the University of Michigan DNA Sequencing Core, predicted with great foresight prior to the meeting.
Combinatorial barcodes, single cell amplification, and other techniques were used to get readouts from single cells. “There were quite a few people applying single cell analyses at the RNA level, at the DNA level, and at the kind of epigenetic level, that shows there are going to be some real new insights into biology coming down the pipe from single cell work,” says Talkowski. “I think it’s going to be much more routine in the future.”
Much single cell work was scattered throughout the meeting -- just one of the technologies that people are bringing to bear on the questions they have. To cite just a few presentations: John Marioni talked about single cell transcriptomics and cell fate decisions. Will Greeleaf described using ATAC [Assay for Transposase-Accessible Chromatin]-seq on single cells on a Fluidigm platform to look at regulatory regions by inference of open versus closed chromatin. And there was a poster on single cell methylation.
Vendors are getting in on the act as well. “BD Biosciences are proposing a new single cell app; Bio-Rad has their app; WaferGen. So there are a number of companies working in the single cell space, looking at different ways of compartmentalizing the cell,” says Hirst. “Droplet-based formats and solid stat formats (like BD) remain the two main approaches.”
AGBT generally tends to be dominated by new methods for sequencing, new algorithms for analyzing it, and prognosticating on the sequencing market, pointed out Lyons. In fact, the more clinical applications of sequencing and its role in personalized medicine have recently been given their own forum (the AGBT Precision Health Meeting, held each fall in Scottsdale, Arizona), “where a lot of the cancer-specific discussion will take place.” May the (self-styled) tech geeks rejoice.
Josh P. Roberts is a freelance writer living in Minneapolis, USA
Encouraging Step Towards a Potential Therapeutic Strategy for AsthmaArticle
Researchers have discovered a compound with the ability to induce long-lasting bronchorelaxation and prevent hyperreactivity in murine models of asthma.READ MORE
LipidFinder: An Open-Source Python Workflow for Novel Lipid DiscoveryArticle
Researchers at the University of Cardiff, UK have developed a Python based computational workflow that is tailored to the analysis of large volumes of data, with the ability to identify novel lipids.READ MORE
Transitioning to Cloud-based Solutions Within the Clinical Research IndustryArticle
Many in the clinical research industry are transitioning to cloud-based solutions for improved medical research and drug development efficiency. Big data analytics and distributed computing are two cloud-based techniques which can centralize computational and research tools.READ MORE