The 2016 Next-Generation Sequencing Field Guide Preview: Zombie Systems and New Hope
After a year of minimal activity, we finally have some significant changes in Next Gen Land. In the 2016 update of the NGS Field Guide, I will continue to give my overall interpretation about the various instruments, but with less snark. In the Overall Instrument Purchase Table (Table 0), I continue to use the Green, Yellow, and Red light analogies, and again, these are from the point of view of someone at a medium to large university or research facility focused on molecular ecology (not human or crop genomics). So, even though the Illumina HiSeq X’s are awesome instruments & an obvious Green Light for human genomics applications, and use by molecular ecologists who can get access to them [they are now authorized for ≥30x genome sequencing of non-humans (only ≥30x)]; HiSeq X’s are still simply not practical for purchase (≥$6M for minimum of 5) or ownership (≥$1M/month to feed ≥5) for molecular ecologists.
If you are in the market for a next generation DNA sequencer in early 2016, Illumina is still the easiest choice to make, but we do now have some reasonable alternatives and one can hope even more will be officially launched this coming year. In addition the purchase overview table, I am providing updates to purchase & maintenance costs table, the major advantages and disadvantages table, as well as an updated spreadsheet for the expendable supplies. Thus, readers can make direct apples-to-apples comparisons on purchasing these instruments and obtaining data from them. I am setting aside updates to the other tables for the time being, but hope/plan to get to them later this year, likely with some help. Now, let’s start with the Illumina instruments and go from there.
Illumina has come out with a new instrument (the MiniSeq) and several modest upgrades to their existing line-up. The MiniSeq is the lowest-cost instrument in the Illumina stable, coming in at $49,500. The MiniSeq will achieve roughly the same number of clusters as the MiSeq, but is limited to PE150 reads (vs. PE300 on the MiSeq). The MiniSeq has a variety of kits, but the Mid-output (8 M reads) PE150 kit @ $550 will be especially attractive. Alternatively, the High-output (25M reads) PE150 kit @$1500 does not compete well vs. the MiSeq (15M PE250 reads @ $1135; and 25M PE300 reads @ $1530). Given the obvious benefits of PE250 or PE300 for amplicon sequencing, I think the MiSeq will remain the better instrument for molecular ecology applications (even though I generally advocate sequence capture rather than amplicons).
The Illumina MiSeq and NextSeq are clear bread and butter instruments for molecular ecology applications. The MiSeq offers the longest possible reads of all Illumina instruments, with the HiSeq rapid-run being only slightly shorter. The quality of bases continues to be low at the end of read2 on MiSeq PE300 and NextSeq PE150. HiSeqs continue to provide higher quality data, even out to the end of read2, at lower cost per base, but their economics make it harder to realize those savings without multiple instruments running 24×7. Thus, ownership of MiSeqs and NextSeqs is a Green Light, while the various HiSeqs range from Yellow to Red Lights for ownership by non-human-genomics labs.
Molecular ecologists who feel that they need to own an Illumina that uses the HiSeq X’s patterned flow cells, should check out the HiSeq 3000/4000’s. Unfortunately, the HiSeq 3000/4000 does not have the flexibility in multiple types of flow cells and read lengths that are available for the HiSeq 1500/2500.
It seems quite likely that Illumina has more upgrades in the pipeline than just the Firefly, but with such a dominant market position, it just hasn’t made any sense for them to bring them to market. Why would they undercut themselves and suffer the headaches associated with major upgrades? Without some significant competition, I would expect Illumina to continue hauling in the cash from their current instruments. When something else big is actually released to the market, then we might see what else Illumina has.
Ion Torrent was acquired by Thermo Fisher and now has a shiny new instrument, the S5 (and S5 XL). Ion Torrent is mostly focused on Ampliseq and the S5 helps on that front. AmpliSeq could be used in molecular ecology applications, but the high upfront cost relative to sequence capture will generally limit its use to very large-scale projects or projects where ultra-high sensitivity is necessary. The Ion Torrent instruments can be used for sequence capture, but they are used much less than Illumina for this application. It is not clear when or if the long-promised PII chip for Proton will be released. It seems equally or perhaps more likely that the suite of chips for the S5 will be expanded and the Proton will become a zombie. There are several applications for which the S5 is a reasonable choice, and it is poised to be a good alternative to Illumina’s three lowest throughput instruments. However, the S5 is not always the most economical choice, it has much less available software relative to Illumina, and there aren’t yet enough labs using them for me to be convinced that they are ready for the green light. So, I’m not sure why most current users would switch from Illumina to the S5. If, however, your goal is to have an orthogonal dataset to validate Illumina data, you need ampliseq or a few other specific applications, or you just don’t want to give your money to Illumina, then check out the new Ion Torrent S5 – it is a reasonable alternative for any of those situations.
New for 2016, PacBio is introducing the Sequel. This instrument is much smaller, lighter, half the purchase price, less than one-third the service contract cost, & poised to deliver 7x more data per run relative to the RS II. The Sequel is now within the purchasing ability of many labs and will deliver a good bit more data at less than half the cost per base (run cost is 3.5x higher for 7x more data). PacBio has finally figured out some of the best uses of their data and has been making good progress in developing software to take advantage of their data. If the read lengths remain at least as good as the RS II, then the Sequel will be an excellent choice for de novo sequencing, especially of smallish genomes or where budgets are not tightly constrained. Unfortunately, the early marketing materials don’t show actual data from the Sequel – they show data from the RS II and imply that the Sequel is just as good. The Sequel is still many times more expensive per base of sequence obtained relative to Illumina or Ion Torrent, but the cost advantages of this model brings many more projects within reach. Given the current unknowns of the Sequel, it is also a Yellow Light at this time, but is poised to go green if all marketing claims are met in the real world of imperfect users.
454 is supposed to stop supporting their instruments in 2016. I was surprised to find that their website still exists and is touting the benefits of 454. I can’t imagine why anyone would still be using a 454, but perhaps my imagination is stuck on zombies.
Oxford Nanopore is in a strange semi-released but still under development mode. A good bit of information is now available for the MinION, especially here, but the major knowledgebase for the MinION still lives more within the user community (i.e., constrained and controlled by user agreements with Oxford Nanopore) rather than in the published or otherwise freely available world. It is hard to guess if a true commercial launch is forthcoming, or if they intend to keep things in this strange semi-open place where the Oxford Nanopore big brother maintains control. It is clear that the MinION has made quite significant leaps forward, but one of the biggest constraints remains the high error rate, which started well north of 10x higher than the rate initially stated at AGBT in 2012, and remains high (and biased) for most users. If Oxford Nanopore can continue to decrease the error-rate and bias, as now seems to be likely, then their instruments will find many applications within molecular ecology. If the PromethION does well, then they would be poised to take over some of the Sequel’s market in the future.
GenapSys is now teamed with Sigma-Aldrich, and they are still recruiting early access (Genius Club) members, but they have yet to release any relevant details about their system, such as the target number of reads or read length, let alone accuracy. One has to wonder who would sign up to spend $2500 when all you really know is the size & color of the instrument? It is nice that they will give a full refund if it’s returned within 30 days, but they have obviously never worked with the purchasing folks at a state university… or maybe they have…and know that returning things is nearly impossible [I said less snark, not an absence of it.]
Finally, there is Genia, which was acquired by Roche. Genia has released some information about their technology, and still claim to be able to deliver a human genome sequence for $100, but they have not yet released any specifics such as number of reads, read length, or accuracy. Given Roche’s very poor track record with NGS acquisitions, I am skeptical that Genia will turn out differently, but I do hope for a better outcome.
So, although Illumina remains the easy choice for NGS sequencing in 2016, it will be a year in which NGS instrument choices finally return to having at least a few reasonable choices from companies not named Illumina. Nanopore sequencers remain more hope than reality, but if Oxford Nanopore can actually deliver on the prelimary data they have recently shown, and GenapSys and/or Genia make it out of the gate, we can hope that their force shall awaken…