Here’s my take on what we might see in 2020 for the sequencing platform market. This isn’t specifically about JPM, but I wanted to get this out ahead of the conference since it’s obviously one of the main venues for “big announcements”.
Illumina
1) Merger/long reads
I was all prepared to talk about the ILMN-PACB merger, but the merger cancellation came out quicker than I could predict it – let’s just pretend that I was going to nail the exact details (that way I’ll get at least one of these predictions correct). So now what for Illumina? While I’m disappointed that the merger didn’t go through, I don’t think it will impact Illumina too much – PacBio needed Illumina much more than the other way around. In terms of “long read” platforms, look for Illumina to talk up internal R&D efforts again (something they haven’t done in a couple of years). However, any success in this area will likely involve acquiring a pre-revenue startup or two.
2) Short read platform
I think it’s still too early for Illumina to start talking about any significant NovaSeq upgrades. Maybe a new flow cell? But I’m not sure what they can squeeze out of the current platform without a significant upgrade. Same for the rest of their platforms. Except for one. The MiSeq is getting really long in the tooth and is the only one not using upgraded (downgraded?) 2-color or 1-color chemistry. Look for Illumina to slot something new between the MiniSeq and the NextSeq.
What about pricing? Will that famous NIH graph remain flat, or will it start ticking lower? I think we’ll start seeing movement this year. Read on to see where the pressure will come from.
3) Clinical
Since they have at least $1.2B burning a hole in their pocket, I’m going to predict Illumina acquires someone with a significant clinical presence. Qiagen might have made sense, but it looks like they’ve taken themselves off the table. This is a long shot prediction that would involve a lot of cash and big changes, but Illumina needs something like this to truly break into the clinical space. FYI, I keep making this prediction year after year on the theory that I’ll eventually be right.
PacBio
1) Merger
See the Illumina entry for how I already perfectly predicted exactly what was going to happen. Since this is a separate entry, I’m now two for two 😉 So what’s next for PacBio? Lots of stress and hard work.
2) Staying it alone?
They’re getting a nice cash cushion from Illumina in the form of keeping the previous merger cash payments and getting a $98M breakup fee, but I fear it just delays the inevitable. They have a great system and great people, but they keep burning cash. It’s possible the Sequel II can turn that around, but they better do it quickly. Watch the platform sales over the next two quarters carefully to get an idea of where this is going.
3) Acquisition?
If that doesn’t work they could try again with a different suitor. This has been talked about a lot, but I don’t think there are any great ones out there. The three most obvious would be:
Thermo
Makes sense because they already have the Ion Torrent platform. But Thermo doesn’t have a history of really supporting platforms with a lot of new R&D which is what PacBio will need if they’re going to survive. If Thermo tries to “milk the cow” like they are more or less doing with the Ion Torrent platform, PacBio won’t last long.
Roche
Roche liked PacBio so much they gave them millions of dollars to build a clinical version of their platform. The end result was PacBio getting paid all the money, the Sequel being born, and Roche deciding they didn’t like what they saw enough to move forward. Instead they went with the Genia nanopore platform (see below for more on that).
BGI/MGI (in theory these are two separate companies, but given the cross talk and management structure, they can be considered one entity that I’ll call “BGI”)
This makes a lot of sense for BGI – sort of a Complete Genomics part two. They could even throw lots of development money at it if they felt like it. I’m just not sure how many of the PacBio people would actually stay. Also, given the current environment, I don’t see this passing CFIUS review.
If not these three, then who? Some other large company with genome connections like Agilent or Danaher? Maybe someone even bigger who decides they want to break into the space like GE Healthcare or Siemens? These all seem like long shots to me.
Regardless, they’re going to have to start talking about their future again. While I don’t think we’ll see a 48M chip or a Sequel III in 2020, I think they’ll start talking about their roadmap. The need to in order to raise money or attract an acquirer.
Oxford Nanopore
I’m a little confused about ONT. Are they the amazing technology company that has made rapid and consistent improvements over the past several years, with >7000 instruments placed around the world, and an incredibly bright future as spelled out by their management? (Already beating Illumina’s $/Gb metric, able to match their $/read metric and soon-to-be-solved systematic error issue.) Or are they the sad, tiny little “also ran” that’s going nowhere and can’t compete without special protection, as spelled out by both the CMA and FTC?
Anyway, I digress. What will the year bring for ONT? I confess that I struggle keeping up with all of the future product announcements coming from ONT. I wouldn’t dare try to predict something on their roadmap that hasn’t already talked about at London Calling. It’s more about which of the announcements will live to see the light of day and how close does reality match the announcement. But here’s my best guess: they will continue to come out with amazing technology and platforms and they will steadily improve their error rate. But I predict they won’t solve the error issues in 2020 – they won’t be able to match Illumina let alone PacBio. I also predict that Illumina won’t really notice any pricing pressure from them in 2020 (but maybe 2021 will be a different story).
It won’t happen in 2020, but the prediction about ONT that I’m most interested in is seeing truly “distributed” sequencing. Sequencing that is so small, cheap and simple that it can be deployed inconspicuously in the home. One for the toilet for monitoring your gut microbiome, one in your toothbrush for monitoring your oral health, one embedded in your kitchen counter, cutting board, or dinner plate to monitor your food. And probably lots of others I haven’t thought of. ONT isn’t close to doing this, but they’re the only company with an even semi-plausible path to this future (think flongle + VolTRAX but smaller and completely integrated and automated). I’m not sure if ONT can ever truly displace Illumina, but I think maybe they can open up new, almost unrecognizable, markets. But not in 2020. Maybe by 2030? That would be worth waiting for.
Thermo
I predict Thermo will continue to sell the Ion Torrent platform. They will likely announce new partnerships and maybe give the machine a new name (with a slightly different compute configuration or color scheme). But nothing significant will happen here. Ever.
BGI/MGI
BGI is one of the hardest to pin down. They talk such a big game, make such big announcements, and claim to have a really large install base. The problem is, even after having made several trips to China over the past year, talking to 50+ customers across a broad spectrum of segments, I’ve only seen one BGI instrument (a BGISEQ 500). I think they’re out there, but they’re highly concentrated where BGI can keep an eye on them and baby them to make sure they run as smoothly as possible. But here goes:
Prediction 1: They will have a much more widely distributed install base in China – lots of placements at more institutions.
Prediction 2: They won’t see broad placements outside of China, at least not in 2020. Lots of fights over IP and just general difficulty transitioning from a services-based business to a true platform-based business will be hard and will prevent them from making headway outside of China.
Prediction 3: They will continue to announce new systems. Even though the T7 is barely out in the wild they’ll announce something even bigger – the T14, T20, T80…, before the year ends.
Prediction 4: They will make enough noise with their progress in China and new, large, cheap (per Gb) platforms that it will start forcing Illumina to drop their prices. BGI won the Abu Dhabi deal. If they win a similar deal in Europe (or, dare I say, North America), it would freak Illumina out.
Genia
I predict one of two things: either we continue to hear absolutely nothing or we hear the Roche has quietly shut down the program. They have simply taken too long to bring something to the market. Unless they have an amazing ace up their sleeve, ONT has long ago passed them by. This might be the year I add them to the NGS Necropolis.
GenapSys
I predict that the GenapSys system will do moderately well in limited applications with their very low capital cost machine. The first two chips (1M and 16M) are suited for small organisms and QC-type runs. But at $600 per run ($300 for prep and $300 for sequencing), the cost per Gb just isn’t compelling. Things won’t start getting exciting until they come out with a completely automated 144M chip. I don’t think they’ll pull that off in 2020.
Startups
GenapSys was able to launch in 2019 after a long delay. Will one of the other darkhorse startups surprise everyone and launch in 2020? There’s Omniome, Element, Ontera (née Two Pore Guys), Quantapore, Roswell, etc (see here for others). Of these I think maybe Omniome has the best chance of launching a sequencing platform this year, mostly based on how long they’ve been working on the technology, the money they’ve raised, and the management team they’ve put together.
Looking forward to seeing how laughably wrong I get these predictions (and maybe even one or two spot on).
Bionano Genomics has an option to in-licence PacBio long-reads technology therefore I see it as a likely takeover target from Illumina
I suppose this acquisition is possible, but I don’t think Illumina could use it as a way to work around the FTC and CMA decision. Illumina is much more likely to enter into some sort of distribution deal directly with PacBio. Therefore, I don’t really see this potential acquisition as very likely.
In my opinion such a buyout is highly likely.
In order to achieve “Platinum Genome sequencing ” you need Illumina’s short-reads, PacBio, long-reads, ONY Nanopore sequencing, 10x Genomics sequencing approach, Bionano’s optical mapping and even Dovetail Genomics’s Chicago.
Bottom line: Bionano’s optical mapping solution is highly complementary to Illumina’s short-reads and indispensable to she’d light on Structural Variations that stretch out to mega base pairs in length.
What do think about Genscript Biotech (China)?
The focus of this post was on sequencing platform providers, so I haven’t focused on service providers like Genescript.
I absolutely agree that distributed sequencing is where ONT is untouchable. But I think there is at least 50% chance they never get around to tackling that problem. ONT has raised sums approaching $1B and the VCs are probably starting to itch for an exit of some sort. If they IPO the implied valuation will be $4B+ which requires some serious revenue. Selling a few thousand $1k instruments and $500 flowcells is not going to cut it, which is why they are wasting money, resources and opportunities on the Promethion – in order to compete with Illumina where Illumina is strongest. This is reminiscent of the trajectory PacBio took and subsequent disillusionment of the public markets.
Yes, their projected revenue models must look really, really different from Illumina. Even the “large” PromethION has very low capital costs. I’m not sure what their plan is to achieve billions in revenue, but my assumption is that it’s based on a very widely deployed distributed sequencing platform. Looks like you think the better bet is to just focus on the PromethION. Could be. It will be interesting to see where they go once they start really having investor/revenue pressure.
I actually agree with Keith R that in the long term the Promethion is a costly exercise in taking on the dominant player on their own turf. Current clinical applications, where the money is, do not really benefit from long reads except in niche cases. And the latest direction that ONT is taking is straight for the heart of the lion’s den – short reads. Capex in that market is probably a less important consideration than RuO. Don’t see how ONT wins there in the next several years. Some day when SVs become important the story changes but ONT needs to live long enough to see that day. The shrewd long term play would have been to raise far less revenue, grow far more slowly, and develop the Minion and Flongle in the markets where Illumina can’t touch them. But that takes patience that few investors have. I understand what they are doing and why they are doing it, but it does not feel like the best long term outcome. The CTO’s intense disdain for everything Illumina also seems to get in the way of rational decision making. We shall see.