Illumina, Inc. (NASDAQ:ILMN) Goldman Sachs Global Healthcare Conference June 12, 2019 3:00 PM ET
Francis deSouza - CEO
Conference Call Participants
Patrick Donnelly - Goldman Sachs
I think we can look to get started. Thanks for being here. I'm Patrick Donnelly, the tools and diagnostics analyst here at Goldman Sachs. Happy to have Francis deSouza, CEO of Illumina, with us today. It's open Q&A so feel free to just raise your hand, and we'll try to get you. But Francis, I think you wanted you read a safe harbor statement before we get started.
Yes, thank you, Patrick, and good morning to all of you and thank you for attending our session this morning. Before we start I'd like to remind you that my comments today could include forward looking statements. You should refer to our SEC filings for a discussion of the risks and uncertainties that could cause results to differ materially from our current expectations. It's our intent that all forward looking statements regarding our financial results and commercial activity made during today's discussion will be protected under the Private Securities Litigation Reform Act of 1995.
All right. Yes, may be just to start Francis, maybe you can just kind of talk through how the year has gone so far for you guys. Obviously, the 1Q results were solid, you had some NovaSeq shipments kind of shift in and out of other quarters. So maybe start there and then we can kind of dive into the Q&A, top level.
Sure. So we came into this year obviously with good momentum coming out of last year. Well we had a strong growth last year and as we talked about this year and the last couple of years, one of the themes that is playing out is the upgrade cycle and high throughput part of our portfolio. The upgrade cycle from HiSeq X and HiSeq to NovaSeq. And we talked about the fact that we've designed that upgrade cycle which frankly is the biggest that's happened in the sequencing market ever to be a multi-year upgrade cycle and that the first year in '17, we expected so the very large high throughput labs we expect and the commercial labs to lead the upgrade cycle.
And then, we expected to see the overall upgrade cycle play out over a three to five year win. So we're in the midst of that this year. And coming into this year we did a number of things to activate a part of the HiSeq customer base to upgrade. So, we launched the S Prime at the low end of the first half portfolio which also has a longer read length. We lowered the price of the S1 and the S2 and the intent is this year to catalyze the upgrade of the smaller end of the high throughput labs the core lives. And so that's sort of one theme that's playing out.
Another theme that's playing out is when we talked about that this year we are going to start to see more revenue coming from a more diverse set of population sequencing estimates that the story of population sequencing for the last few years an exciting story was really driven by GeL in the U.K., but this year we expect it to see whether it's projects in the U.S. but maybe half a dozen projects around the world that would contribute to population sequencing revenue for us. So that's been playing out this year.
We're also talking about the fact that we're seeing continued momentum build in the clinical markets whether it's a build in the reimbursement frameworks for NIPT especially in Europe but also a good progress in terms of reimbursement and regulatory approvals in oncology. And so that's another theme playing out this year.
Maybe we start on population sequencing and seems to be the biggest focus for investors. Maybe just talk through the projects that you're seeing out there and you mentioned maybe half a dozen certainly some big ones in the U.S. that are in focus group. There's talk through the visibility you guys have again the second half you're talking about I think a $55 million step up that Sam broke out on the call. There's talk through those projects, the pipeline, and then the visibility and those of them playing out the back up?
Yes. So, population sequencing efforts just to sort of ground setting are efforts that are driven in a lot of cases by governments. But the intent is to sequence big chunks of the population and there are really two flavors of them. The first population sequencing effort in genomics England, started in the 2013, 2014 timeframe and was primarily a research project and it wound down sort of end of last year beginning of this year, doing a 100,000 genomes and the intent was to do good research and discovery around the clinical utility, the economic value of doing next generation sequencing testing in the health care system.
And so one flavor of population sequencing projects is to create the next GeL in other countries. So for example in the U.S. you have the All of Us research project that's modeled on GeL it's a research project and the intent is to learn about the economic value of using next generation sequencing and what we can get out of it.
We're seeing though and what's very exciting is we're seeing through the next generation of population sequencing efforts and the difference is projects like the project in France for example aren't research projects. They are driven by the health care system and that's also happening in the U.K. now.
So the next step of GeL, GeL 2 some people are calling it's really not a research project at all it's driven by the NHS, the National Health System where they are building into the standard of care in the U.K. the use of Next Generation Sequencing where people who have genetic diseases and who have cancer.
And so that's really exciting because you're no longer seeing it as a research project with a beginning and an end or a certain number of genomes. It's not just getting baked into the standard of care in the U.K. but the entire U.K. population. And you're seeing that flavor of population sequencing projects now in a lot of cases around the world. France is another example of that.
And so today when we talk about the fact that we have a team in our company that's specifically focused on population sequencing efforts. They are deeply engaged with the countries around the world that are looking to do these populations sequencing efforts.
In some cases they're helping design those efforts. And they're sharing the learnings we had in going through the journey with GeL. And so helping those together, and so that gives us a unique seat of table and a unique point of view around what projects are happening and when. We are working on over 50 projects around the world, not all of them are public but if you write them you'll see maybe 30 ish are public. And so we're working on them being rolled out and a lot of them as I said are really about national systems embracing sequencing as a standard of care for their populations going forward from now on.
And maybe just talk through what you learned from GeL, so I think the cadence was maybe a little chunkier than sort of I expected with large number of it coming through last year obviously I think around 75,000 to 100,000 and expectations of other POPSEQ initiatives relative to that.
Yes, we had - GeL is a terrific project, it's a really exciting project obviously U.K. is pioneering the world there. And we're really happy in terms of thinking - as an industry we're very happy about the learnings that came out of GeL, and the result is also promising. It's obviously being rolled out into the health care system.
And we learned a lot about the utility using next generation sequencing in genetic disease and in oncology. And I'll compare sort of where we are right. So, if you have a child born with a genetic disease and it's estimated that north of 5% of children today are born with a genetic disease. Some of them will present immediately after birth with symptoms like epileptic seizures, and it's not uncommon for a child to be placed in the NICU with the seizures, the doctors trying to figure out what's wrong and if they can't figure out what's wrong then the child is sent home with a diagnosis as a failure to thrive.
And then on average in the U.S., the child will go through a diagnostic odyssey that could be seven to nine years. On average the child will be misdiagnosed two to three times. In a 9% to 10% of those families will go bankrupt. That's the standard of care here. And what we found working with GeL in the U.K., that there's a better way to do this. And what they're looking to do in the NHS is they're saying, look if on pick it any day, five day, seven you still haven't found a diagnosis.
And then what you do is you do a whole genome sequence of the child and the parents, and 50% plus cases they're able to do a diagnosis right there. And so you're talking about now day seven to day nine versus year seven to year nine. And so the economics are so compelling that the pain avoidance is so compelling that they're building it into their health care system.
So we learned that to roll that out in the health care system and it's in the U.K., but it is the lesson that can translate everywhere. We've learnt a lot about the protocols that you have to put into place, the physician education you have to put into place. For example, one of the things we learnt in the U.K., was for the use NGS testing to select therapies for cancer patients you really want the tumor to be fresh frozen samples, not FFPE because you get too much degradation and we had to rework the accession protocol for tumors.
And so that's the kind of lesson you learn. Just having gone through that process and that's the kind of lesson we can share with population sequencing efforts around the world either through our experience but also connecting them with the U.K., and saying look here's how you think about it. Here's where the highest value is initially. But to do it here's what you need to do in the health care system.
Yes useful. I’ll just repeat the question. It's just around visibility on the population sequencing initiatives revenue timing?
Yes. In the cases that I talked about the 50 plus, we are engaged with the teams directly. We are working with those teams, we're helping plan the rollout, we're helping contribute thinking around the infrastructure for the rollout. And so we are actively in dialogue now, we know even from our experience with GeL these are a lot of cases government initiatives, they're complex, there are multiple stakeholders.
And so there is a timing that gets laid out that we work with and in some cases it goes later than expected. But generally we're at the table right with those conversations. In all of those cases what we're hearing is it's not a case of - if they will go forward, it's a case of when they will go forward.
Then maybe on last one on the population sequencing side obviously the volumes up front are important and sort of a big piece of revenue. But maybe talk a little bit about the back end - something like GeL where the data obviously analyzing that is a big piece of Illumina's role.
That's right. And GeL I think is very exciting from a research perspective and a project. But the knock-on effects are huge. The one that I touched on first was making it a part of the standard of care in the health care system is hugely strategically important. It dwarfs the financial value of the GeL project by a lot because now it's NHS and in perpetuity for genetic diseases for oncology testing.
Another big aspect of the value of GeL as you touched on is the data. The data that comes out from sequencing the 100,000 genomes that they did and matching it with good phenotypic data. And so you create this very rich database of genomic and phenotypic data that has other uses. They're making it available to the research community, they are charging players like pharmaceutical companies to get access to that data because that data presents a very rich data source or drug development. Look for novel targets and so you should expect that that data will continue to generate value going forward.
From Illumina's perspective, there are a number of roles we are playing right. So obviously we generate the data from our sequencers and the idea is to make it as easy as possible to process that data. And so we have a number of things we're doing on that front. And it becomes more and more important as we look at the other systems, the other national systems that are coming online. Because we talk about GeL and the U.K. and we talk about all of us here in the U.S. but there are initiatives that have been announced from companies like Bangladesh or India and even some of the developed countries like Finland or Denmark that may say look we don't want to go and figure out all the different parts. Show us the reference architecture and make it easy.
And so what we do is the sequencers today generate baseball and with the Edico acquisition we did last year what we are able to do is we're able to provide you a pipeline that takes you from base calls to variance. And that's a pretty big step forward.
And so, as a customer you no longer have to go and figure out what's the best designer tool or what's the best variable color tool. How do they all work together. We provide that standardized pipeline for you. And we've built it into hardware. And so we use FPGA. So you get hardware acceleration of that pipeline.
And I'll give you an example of what one customer is able to show is they're saying look, going through the standard leading pipeline today it took them about sort of 19 hours to get that processing done. By using the technology we bought for medical, they were able to do it and depending on the run between 20 and 46 minutes.
And so, you get that substantial acceleration, plus you get a standard pipeline that you don't have to go and choose the tools yourself. So, we're helping our customers and say here's how we'll take you from base calls all the way to variance and then we provide tools and go from variance to interpretation. We have a case lock tool for example.
In addition to that, we are cultivating an ecosystem of partnerships where if you want to apply more sophisticated analysis to the data we allow you to.
Q - Unidentified Analyst
…or so that we don't think of it as a product cycle company and that you concentrate on you have price tiered and architecture has changed that NovaSeq and that it would be more like software services, speed and things like that. Is that kind of the general thinking?
I think that's an evolution that is happening. And it's one that we are actively sort of moving along. Right. So, you think about it historically. You're right. One of the lines you look at us on and you'd see a product launch, you would see sort of significant pop in revenue in the product launch and then you would see the next pop at the next product launch and over time, you shouldn’t see quite the same the pop and the drop as you saw before and the number of things are driving it.
One as you look at our business today, our sequencing system revenue is just a much smaller part of our total revenue than it used to be and so what it used to be in the 30s a few years ago percentage of revenue, it’s now in some of the mid teens percentage of revenue.
And that mid teens revenue is now split across a broader portfolio, so in the old days we are a very small number of sequencers, today we have a much broader portfolio from the HiSeq, MiniSeq, MiSeq, NextSeq and the NovaSeq and so no single upgrade cycles is responsible for that mid teens sequencing revenues. And so the numbers alone will tell you that we’re moving away from the world if you know this is the big product launch and the big ramp and then sort of defining it. In addition as a company we are we are making sort of moves to move in that direction.
So for example if you look at the NovaSeq upgrades cycles, we have a very deliberate about driving that to be a multiyear upgrade cycle rather than big bowler in year one and then diminishing in year two going forward and how we have done that. We have done that by having a family of flow cells that target different parts of the HiSeq base and making those flow cells available in a measured way over time and so that we’re activating the upgrade cycle of different parts of the customer base at different times and so we launched with a flow cell that was ready geared at the higher end of the larger end of that customer base and then a couple of years later we launched the S Prime that was targeting the smaller end of that customer base and activating the upgrade cycle of that base.
And so that also helps have this be a more measured upgrade cycle rather than big and what you saw then is you saw more NovaSeq been sold in year two than year one of the launch and that’s similar but more in year two and so that is part of strategic intent is sort of drive as to be more smooth over time rather than popping up.
In addition you are seeing the clinical customer base be it bigger part of our customer base as we go forward and that customer base, that is not necessarily the first to jump on a new product that comes out and so that dynamic also will help us get to a world where we have a smoother ramp than you might see from a product lifecycle company.
Let me just expand on the instrument side, you guys had I think it was $6 million shortfall in 1Q primarily based on timing, maybe just talk through the conference you guys had was timing and if those play out in the next two quarters capturing those before obviously the big up tick in 4Q?
Yes, from our perspective the timing of specific sales is driven by our customers and so we want to make sure that we are very responsive to when our customers wants the system and so they tell us and in some cases there are issues where the lab may not be ready on time or personnel maybe not ready on time and we are very flexible with customers about if they tell us it’s next week rather than this week and this week again, that’s six systems were rather in the system our business has $4 million, $5 million that doesn’t change the shape of our business and so we are very responsive to our customers and say here is tell us wanted and we will work to make sure that happens. We look at those numbers but those kinds show signals big change in that.
Okay. And then I was thinking about 4Q, you guys have talked about 4Q placements being two times what they were in 1Q, maybe just talk through the drivers there and again visibility into that margin up tick which is probably little more than that?
Yes, so I will start by saying that seasonality, our hardware business and the way we see seasonality play out is we have end of the year budgets that play out at different times for example the Japan end of year budget is Q1, small part of our business but we definitely see interest in Japan pick up in Q1. The bigger seasonality for us is in Q3 and Q4 and Q3 is going to be end of the U.S. government budget cycle and so we see builders spending driven by that and in Q4 it’s typically the end of the corporate budget cycle and you see budgets happened there and so that dynamic is definitely part of our business and has been every year. And so there is an element playing out here. This year in addition to that there is a further back end loading given the projects that we talked about, the larger projects that are ramping up over the back half of the year.
And so we start with the regular seasonality and then you overlay some of your specific factors and that's what results in the numbers we talked about. Last Q4 we had over 100 NovaSeq and so it’s not unusual for us to see those kinds of numbers that play out.
Okay. And then maybe just on the staying on NovaSeq, we've got the long awaited consumable pull through number last quarter, caught some people by surprise but a million dollars maybe stuck through I guess the thought process on giving that number and where you expect it to go over that?
Yes, the NovaSeq pull through number, I recognize it's an important number in terms of how people model our business. So I'll talk about in terms of building their models. How many instruments and the average pull through per instrument. From our perspective though the pull through number is really an outcome, it's not a number we try and optimize for what we wanted to do with NovaSeq strategically is we want to make sure we are democratizing access to high throughput sequencing what we wanted to do was put the power of our highest end machines into more labs than we ever had before.
And if you look at where we were before, the most powerful movie machine we had is the HiSeq X and there were less than 50 customers around the world that bought the HiSeq X and if you bought the HiSeq X, you had a significant price advantage in doing whole genome sequencing than customers who bought HiSeq 2500, 3000, 4000 and so the dynamic that’s created in the market was you had the 800 HiSeq customers, the under 50 X customers and if these customers wanted to do whole genome economically the most economically possible, they would outsource to the fifth.
We wanted to know the secrets but that power in the hands of as many customers as possible. And so today we're at a stage where the price differential that a customer can get between running NovaSeq themselves or going through the very big labs is the narrowest it's ever been. And so that's very exciting for our customer. And so we want to get NovaSeq’s out there and want to catalyze the upgrade from the HiSeq base. And so as long as we're democratizing access as long as we're decentralizing access to high throughput sequencing, that's sort of the strategic win for us.
That will generate a pull through number and now we've had a couple of years of data and a million dollar numbers higher than some people thought and so we wanted to share that number. So people knew sort of where we were. That number will move around a little bit going forward. We just launched the S prime, the S1, S2 flow cell that we priced and we only have 25% of our HiSeq customers have bought NovaSeq so far.
So it's still three quarters to buy. And so that number will move around. We said that we expect that number to drift up over the course of this year and where it makes sense, we will provide that number as relevant but we won't be updating that number quarter.
Okay. And then maybe thinking about democratizing the flow cell changes. Let me just talk through how that is catalyzing some of the lower throughput HiSeq users again the S1, S Prime, the price stream, what went into that decision, how the reception has been?
Yes, so the way we thought about the NovaSeq upgrade cycle as we've talked about it being three to five year upgrade process. And then we'd launch the NovaSeq in 2017, the way we thought about it is the first customers that will upgrade are going to be the highest throughput labs. They're going to be and we expected to be the commercial lab. And the reason for that is a lot of academic customers have to go through the grant writing cycle. And so when we launched a product in Q1, we knew that a lot of the academic customers and Core Labs will have to wait until the next grant cycle before they can get the funds to buy NovaSeq.
We also knew that a lot of big projects were already underway from a research perspective. And so those labs that are doing those projects, those funding agencies that were funding those projects weren't going to want to switch midstream from an X to a NovaSeq. And so the first year was really primarily around making sure we were activating the very large high throughput customers and then we knew that starting the second year we'd start to see some of the grants kick in and more of the academic customers coming.
And then this year the third year the intent is to catalyze the low end of that. It's still the high throughput customer, it’s the low end of that customers in the Core Labs and to do that we had to do a number of things. One is we had to provide a full scale that was sized to the sample volume that those Core Labs would see and that's what you saw with the S Prime. We also had to lower the price of the S1 of the S2 to make it accessible to those customers. The other thing that we did with the S Prime was we gave the longer read capabilities to 250. And what we've heard from our customers was that was still one advantage that customers had in running the 2500 that they weren't able to match on the NovaSeq.
And so the portfolio that came out this year closes that gap. And so if you were using that and there's still some applications where that's relevant you now had access to that capability on NovaSeq and so those are the customers that were catalyzing this year. And that's sort of the reaction we're seeing their reaction to S Prime in Q1 as we said was very positive.
The thing that surprised us overlaid in all of these years as we saw it in 2017, we saw it in 2018, we saw in Q1 was the percentage of customers that were new to sequencing or straight from bench tops, the new to high throughput that came in with note, that came in larger than we expected and north of 25% in each of those in each of those periods. We weren't expecting to be quite that.
And maybe just shifting over to the PacBio side. Maybe just update us on where you guys stand there and maybe some quick background of deal for us?
Yes, so we announced in the fall of last year that we were acquiring PacBio, PacBio provide sequencers in the long read market. So that's about 5% of the overall sequencing market that we actually don't serve today. And it’s a segment of the market where customers will want to sequence for example a new species where there isn't a reference genome or they'll want to or applications like organ transplant, they want to read parts of the genome for example where it's hard to read using short read technology.
We play in different parts of the market. You wouldn't use a long read sequencer for example in markets like NIPT where you're reading 160 base fragments reading it 10,000 it doesn't make sense for those markets, it's a higher cost per sequencing to do the long read sequencing.
But in those markets where you want that data that's what you'll do. So we announced last year that we are looking to acquire PacBio. That acquisition will require approval both in the U.S. and in the U.K. and we're in process with both the FTC here in the U.S. and the CMA. We said we expect that acquisition to close mid-year and that we're expecting to hear back from both the CMA and the U.K. and the FTC in the U.S.
In the U.S., the decision will be made by the five commissioners of the FTC and in the CMA, we'll be hearing about they’re in the first phase of review and the choices are they can approve it, reject it or they can move into a second phase of review. And so we don't really have new news around that, we're still in that process.
Okay. Maybe just on the competitive landscape you guys have long held a pretty dominant market share. Maybe just talk through any changes there particularly maybe in China, we've seen some IP lawsuits buyback and forth with you guys. Maybe just an update?
Yes, look with the market as big as the sequencing market with a market that is rapidly growing as the sequencing market. We expect to see and we'll continue to see competition, we'll see existing players in our short rate market for example. You have companies like Thermo that acquired technology and they have a sequencer companies like Caogen in that market. You have PGI in China that's in that market. And then you see every year there are sort of the new companies that are funded by the venture community. So we look at those as well. Across the board and it's true for all our franchises or assays including our sequencers. We have a very strong IP position.
Both around our SPS, our sequencing technology as well as some of our assay technology and we are vigilant in protecting our position. And so we are currently, we filed IP infringement lawsuits against PGI in markets where we've seen them because we believe they infringe on our SPS technology and we'll continue to do that. We fully expect to compete. We expect people to innovate on their own technologies. But if we feel that somebody is infringing, we will protect ours.
And maybe since you touched on kind of ventured backed companies one market that's certainly got a lot of attention liquid biopsy side. Maybe just talk about that market. Your guys role in supplying a lot of those customers. And where do you think we are in the progression of coming there?
Yes, liquid biopsy is very exciting. And the idea there is that you can actually, you can sequence a tumor or tumor fragments by just taking a blood test from the patient rather than getting a biopsy, a tissue biopsy. So as you can imagine that's really exciting in a number of areas. And you have companies like Guardant that are really sort of pushing the envelope and then sort of leading the industry around using our sequencing technology and their assay technology to do blood tests to do a number of things to profile the tumor, to monitor the effectiveness of the therapy and they are very moving quickly, they’ve got MCD, they have got reimbursement in place for using their tests for conditions like non-small cell lung cancer where it’s very hard to get high quality tissue samples of lung cancer right and you certainly don’t want to be biopsying a patient multiple times whereas you can do multiple blood tests to look at the progression of a lung cancer.
But they’ve also got reimbursement for solid tumors as well and you have other players like Grail which we incubated internally to Illumina and then spun out that are looking to use liquid biopsy and blood test to do early stage cancer screening using a blood test, we thought obviously life changing because the survival rates if you catch somebody with a stage 1 cancer across a variety of cancer types is so much higher than if we catch three and four which you do a lot in cancer cycle rectal and pancreatic where there aren’t symptoms, Grail had terrific results which they showed at Pascal Labs and then ASCO again this year and so we’re excited about the progress they’re making in the use of liquid biopsy for cancer screening.
Our role in that is well in Grail’s case we incubated the company and then spun it out but across the we are talking dozens and dozens of companies now in liquid biopsy, we provide the sequencing and so in that space, the application plays very, very well to the strength of our sequencers, you need sequencers that are that can do high throughput sequencing, that can do very deep sequencing in some case, you need very accurate sequencers because of the end of that sequencing run, you are influencing the care of a patient and so those are the customers typically that are aggressive adopters of our high throughput sequencing system.
Okay. Maybe sneak one last one in here, in some the consumer genomics market that was big growth area last year as seem to had a bit of a low here, give us an update on your thinking of that market and then go forward?
Yes, to play it out we saw really strong growth in the direct to consumer market in 2017 and 2018 we had now north of 20 million people have had that is done through those DTC companies and the vast majority of that happened in 2017 and 2018 and we’re now seeing a pause in that market. And that’s happening as the market is transitioning from being a genealogy led market to being a health market and over time the health market is a much bigger market than the genealogy market but we’re seeing that transition happen now.
We are also seeing another transition play out which is the story of direct to consumer genomics has really been or see U.S. story and that’s been the vast, vast majority of the revenue of company and really a story driven by two companies Ancestry and 23andMe and you’re seeing both those drag, if those dynamics start to change. We are now tracking over 100 companies in DTC around the world, it’s an exciting stuff happening in China, Japan, Korea and so you’re seeing diversification of that customer base both in terms of number of customers and in geographies. But it will take time sort of that to build.
I think we’re out of time Francis, thank you so much.
Thank you so much. We appreciate it.