Los Angeles, CA February 16th2014 – The Advances in Genome Biology and Technology (AGBT) ended yesterday evening. Here are some bullet points from the conference, based on candid discussions we’ve held with some attendees.Continued conference focus on NGS

• “… AGBT has felt more like A-NGS-BT lately; It’s been very NGS centric. I’m surprised that there is not more emphasis on validation this year …” – Director of Scientific Affair, Top 10 Research tools company
• NGS continues to run the show and revolutionize genomics, with a large number of talks and posters focused on targeted resequencing (e.g., MGH has sequenced 57K+ exomes to date) and RNA-Seq.
• As a result, a couple of participants that we talked to were surprised by the lack of emphasis on validation, using innovative tools such as digital dPCR (dPCR) that have taken off in the last couple of years. In the case of dPCR, “the technology has been demonstrated to offer greater sample throughput, statistical power and in some cases sensitivity”, a couple of attendees confirmed.
• When probed, attendees highlighted that other interesting applications and technologies were on display during the conference, such as 1) Hi-C (using light chromatine crosslinking and fusion to identify functional interactions in the genome; J. Burton), 2) sample preparation using molecular inversion probes (MIP; IDT) given their workflow simplicity and scalability (but still limited by their performance: ~95% capture vs. 99.9%+ using hybridization), 3) NABSys, positioning its platform as not competing with NGS, given the lack of single base pair resolution.
• NGS platforms are being adopted in the clinic at a pace that exceeds DeciBio’s previous estimates, despite continued concerns around workflow, actionability and reimbursements. “Only 5 - 8 genes are typically being reimbursed from panels that test ~20 genes (typically liquid tumors) or ~45 genes (typically solid tumors)”.

NGS platforms

• “…Competition among NGS platform providers is healthy. There is room for more than one platform depending on the application. I prefer to use our Illumina sequencers, but I continue to use our Ion platforms …” – Lab Core Director, Top 10 Cancer Sequencing Center
• Illumina remains dominant, and shows no sign of complacency. As opposed to 2012, Illumina did not use AGBT as a platform for material announcements. Rather, the company used the meeting to reiterate and clarify the benefits of its novel product line (now including NextSeq 500, HiSeq X ten, and NeoPrep). We talked to a couple of skeptics about Illumina’s NextSeq 500 2 channel chemistry (using essentially no signal as a signal), but ultimately, the data doesn’t lie, and the platform appears robust, offering the speed and throughput of the HiSeq 2500 in rapid run mode with the ease of use of a MiSeq. Among the 20+ customers that we talked to throughout the meeting, the vast majority were satisfied with current Illumina machines and services. A couple of customers, however, tried (in vain, to our knowledge) to convince Illumina senior executives to have access to the HiSeq X Ten on an individual basis, and ask for the company to waive the requirement to purchase a minimum of 10 instrument per site. While the $1,000 genome is here, it remains out of reach of many. Jesting customer in the Illumina lounge: “So you’re telling me that you built this amazing piece of machinery, but that I cannot put my hands on it unless I have $10M? This is unacceptable. Where is Jay? ...”
• Ion Torrent seems to finally be able see the light at the end of the tunnel, following a rough patch in 2013 due to technical challenges with the P2 chip and delays with Ion Chef. The P2 chip will be made available to early access customers in April / June 2014, and will be commercially available in H2 2014. The Ion Chef will be available in March 2014; the primarily feedback from early access customers is apparently very positive. Finally, the company also announced during its Ion Torrent Innovation Update that Hi-Q chemistry seems to be performing well (error rates: 0.2-0.4%). Finally, the Rapid Isothermal Amplification program (formerly known as Avalanche) is moving along.
• PacBio is back. Many attendees we spoke to were impressed with the headways that PacBio made with its long read technology, especially as describe in Richard McCombie’s talk using the P5-C3 chemistry (with ~10 kb mean read length, and reads of up to 36kb+). While only a couple of talks took advantage of the platform at AGBT 2013, this year, many talks (and publications throughout 2013) highlighted the benefits of the long read technology and its ability to effectively address niche applications for transcript variant analysis, hard to sequence / high repeat regions (e.g., Killer Cell Ig-like Receptor [KIR]) and metagenomics.
• In contrast, Oxford nanopore continues to face technical challenges. In a talk entitled “Assembly of Bacterial Genomes Using Long Nanopore Reads”, David Jaffe described the use of the oxford nanopore miniION instruments to disambiguate reads from a MiSeq instrument. While the platform enabled reads of 5kb+ (in contrast with longer expected reads), read quality seems to remain poor, or as put by the speaker “perfect reads were punctuated by artifacts”. 84% and 100% of long reads have at least a perfect 50-mer or 25-mer, respectively. Jaffe mentioned that going forward, accuracy could be improved by improving base calling algorithms, and by mixing pores with different propriety / accuracy profiles. Interestingly, the pores apparently measure the current from 6 base pairs simultaneously. In another talk using nanopores, the pores were reading 3-4-meres. Of note, Oxford Nanopore started to offer invitations to its MiniION Early Access program ($1,000 refundable deposit and $250 delivery charge) this week.
• Roche / 454 Life Sciences was essentially absent from the meeting. We’ll leave it at that. “I don’t want to see 454 go. This decrease in technology biodiversity doesn’t benefit the scientific community”
• Qiagen made a good first impression, with attendees apparently willing to give the company’s integrated ecosystem (starting with the QiaCube and ending with Ingenuity) the benefit of the doubt. The GeneReader is expected to first be commercialized with a throughput of 2 Gb (20M reads of ~110 bp each), which is relatively low compared to other instruments, but enough for the many clinical applications that the instruments is aimed for (initially primarily cancer panels). While the company did not disclose the instrument list price, they mentioned it would be competitive with the price of a MiSeq, with reagent costs of $400-500 per chip. Of note, the instrument will have random access capability, and will be able to accept chips (up to 20) every 15 minutes (although some chips might sit idle in the system if more than 6 are loaded).
• Genia's technology remains in a very early stage. With only 12 bp reads (up to 2x this length in best cases) and 250 pores in parallel, the platform demonstrated in a proof of concept study that the technology might be viable. “I understand one of Genia’s goal is to deliver a 1000x improvement in the number of pores by next year’s meeting”.
• Genapsys appears to be in marketing mode. We are waiting to see the data before passing any judgment.

Building of an ecosystem

• “… It is possible for companies to benefit from the emergence of NGS even without a sequencing instrument. Sequencing is actually the easy part right now. What’s hard, time consuming and expensive is the data analysis. […] We are going after the bioinformatics bottleneck, differentiating ourselves by the analysis speed of our software …” – Founder, NGS Start-up
• This year, both Illumina and Qiagen continued to emphasize and articulate the concept of an NGS ecosystem, a combination of multiple protocols and technologies. NGS providers do not hesitate to integrate vertically, as demonstrated by Qiagen’s acquisition of CLC Bio and Ingenuity. As we previously highlighted in our NGS report, it is possible for companies to participate in this space without a NGS platform, by offering front end (sample prep), back-end solutions (bioinformatics) or both (e.g., PerkinElmer with its Caliper and bioinformatics offerings, Agilent). The NGS opportunity has attracted a number of different players, including many start-ups (e.g., Bina, KapaBiosystems, GenCell). A number of these new players try to differentiate themselves by accelerating time to result, a parameter particularly important in clinical settings where NGS is expected to have the greatest impact in the near future. Another set of companies are offering NGS clinical kits, as demonstrated by Enzymatics Archer technology. " People were very interested [in Enzymatics' Archer targeted sequencing technology], as gene fusions have been notoriously difficult to measure using NGS, despite their profound effects on cancer biology."
• On the sample preparation end, people were impressed with the novel Agilent SureSelect QXT, a kit 3.5x faster than other solution (driven in part by a fast hybridization step of 90 minutes) with 30% less hands-on time. This kit is expected to compete with AmpliSeq, and have turnaround time competitive with rapid target protocols like Nextera. It will be interested to see how the Broad institute (that recently shifted from Agilent to Illumina) reacts to this new protocol. Other companies repeatedly came up in conversations, such as Kapa Biosystems (and its library prep) and Enzymatics (and its enzymatic DNA shearing technology). “That was a smart move from Enzymatics to set themselves up as gold sponsors for the event. They got lot of visibility from it and seem to be punching above their weight at the moment”.
• This year, bioinformatics companies had the opportunity to display their products and services during a software demonstration. Bioinformatics companies highlighted by attendees we talked to included Ingenuity (now part of Qiagen), GenomeOncology, DNA nexus, and BioTeam (in collaboration with Galaxy). In addition, attendees also highlighted that “BaseSpace is finally starting to look like something usable [and actually pretty good]”. The most optimists believe BaseSpace may act as a marketplace that one day may host thousands of apps.
• Fluidigm is continuing its push for single cell sequencing, enabled by the C1. While the technology is essentially functional, we expect to see a more robust adoption of this platform in the coming year. Of note, the company disclosed data demonstrating that just 250K reads (easily manageable for hundreds of cells on the MiSeq) were enough to pick up key genes that could differentiate cells within distinct sub-populations. Another presentation outlined that the C1 in combination with an ion proton can analyze transcriptomes of individual cells at a cost of $10/cell (to be decreased to $3/cell with the P2 chip).

Disclaimer: Many of the companies listed above are DeciBio clients and/or customers---

Authors: Stephane Budel, Partner at DeciBio, LLCConnect with Stephane Budel on Google+https://plus.google.com/+StephaneBudel