Illumina CEO Jay Flatley on How to Keep an Edge in the Fast-Paced World of Gene Sequencing
San Diego-based Illumina (NASDAQ: ILMN) is without a doubt one of the bigger success stories in biotech of the past decade. It makes biological research tools that significantly boost the efficiency of high-speed gene sequencing, instruments that spot subtle variations in long stretches of DNA, and products that analyze important ways in which genes can get turned on or off. The company, founded in 1998, has grown to almost 1,800 employees and a stock market valuation of more than $4.7 billion.
While getting to the mountaintop is hard in any business, it’s just as hard to stay on top. That’s particularly true in this era of mind-boggling innovation in gene sequencing, in which a number of companies are racing to bring the cost of an individual human genome sequence down to $5,000, or even $1,000 in the not-so-distant future. This is a technology race with profound implications for the healthcare system, and with the potential to usher in the long-envisioned era of personalized medicine.
That’s why I was happy to have a chance to sit down last week with Illumina CEO Jay Flatley for an interview at his company’s headquarters in San Diego. Flatley, 57, has been in charge at Illumina since its early days in October 1999. We covered a lot of ground in a little more than 20 minutes, so to make this easier to digest, I’ve broken the conversation into two parts. The first part is running today, with Part 2 coming tomorrow. Enjoy.
Xconomy: We hear so much about this being a really intense period of innovation in sequencing. It’s better, faster, cheaper. The $1,000 genome is apparently coming. Is this the most innovative period you’ve seen in sequencing?
Jay Flatley: Yes. Without a doubt. I was involved in the prior switchover from gel-based sequencing to capillary sequencing, in my prior company, and we launched a product called MegaBACE. It was the first capillary DNA sequencer. AB [Applied Biosystems] followed about a year later. Those technologies wound up sequencing the human genome. There was a lot of background research work going on in universities that was funded by the National Institutes of Health, but at that time, no products made it to market other than those two. So there was no market innovation.
I sold one company to Amersham, and Amersham didn’t invest in it. So AB had a run from 1998 until about three years ago, where they effectively had no competition in the high-end sequencing market. Innovation almost stopped. If you look at what transpired in those 10 years, it was close to zero in terms of real innovation. They made them smaller, or a little bigger. More capillaries, fewer capillaries. No dramatic changes. It was really with the advent of next-gen sequencing that made all the venture people realize there are other ways to do sequencing. Huge amounts of money poured into the space. And the elasticity of the market became apparent, because of the potential price points. People said, ‘Wow, we’re going to be sequencing millions of people here, and millions of crops, and animals. It’s going to be an enormous market. And that’s what has fueled the money coming into the space.
X: What time frame do you think this really started happening?
JF: Maybe three years ago, or maybe a little earlier than that. Solexa was obviously the first one, but there were a couple of other ones coming along in parallel that didn’t quite make it. PacBio got funded about six years ago. Now you see a rush of people going into sequencing here in the last couple years, trying to get a toehold. Because the general view is that ‘the market will be so large, that even if I get 10 percent, I can be a very successful company, or I can sell myself to someone else and get a big return.’
X: Who do you consider your major competitors now?
JF: The primary one today is Life Technologies (NASDAQ: LIFE). We think we’re way ahead. We’ve got about two-thirds of the market share, roughly, in next-gen sequencing. On the upstart side, [Mountain View, CA-based] Complete Genomics is probably trying to take part of the market that we think is ours. So there will be a lot of competition. I think there’s an emerging collection of companies at the low end now. We’ve been focused up until now on super high-throughput, cheapest, most number of bases possible.
This emerging segment is of low-capital access, faster turnaround, targeting diagnostic applications, with maybe a little bit longer [DNA] reads. It’s going to take over the other part of the existing market that our current systems don’t service. That’s where the Ion Torrents, and the Starlights [Life Technologies' new sequencer dubbed "Project Starlight"], and our Avantome product, which we haven’t launched yet. That’s going to play into this market segment.
X: Who among the crop of startups is doing exciting work, and how might that fit in strategically with what you guys are doing? I’ve written some about Oxford Nanopore, who I know you’ve invested in, but who else do you consider interesting?
JF: Obviously we think Oxford is the most interesting, because it’s the one we bet on [as part of a $28 million investment announced in February]. We think it’s the most elegant of all the solutions. It’s not passed its feasibility points yet, but we think it will be the cheapest and the fastest if it works the way we hope. Ion Torrent is an interesting approach; it’s another one that gets rid of lasers and dyes and relies on electrical detection. It has its set of challenges, but it’s a very interesting, directionally interesting, technology.
Starlight, if you want to take a single molecule in the existing regime of technology, is probably the most interesting. If you’re still using lasers and dyes, I think it’s going to be huge competition for PacBio. It puts PacBio in a really challenging spot if [Life Tech] is successful with Starlight.
X: Is it harder for you guys to maintain an innovation edge when you become a bigger company?
JF: We hope not. It’s characteristic, for sure, of companies as they grow, that they generally regress to the mean in lots of ways. As CEO, my No. 1 job is to make sure that doesn’t happen. We work really hard to make sure that we out-innovate all competitors, including the startups. It takes a lot of work, and it’s an organizational challenge to do it, but so far we have been successful. As long as I remain CEO that’s something I’m going to be really pushing on. We have to remain agile, we have to make decisions quickly, we can’t get bogged down in bureaucracy. It’s a challenge in a large company to prevent that.
X: This period of innovation, how long do you think this will last before sequencing becomes commoditized?
JF: At least five years, if not 10. After that point in time, sequencing will become so routine that you’ll be focused on low-cost provider technology. It will be very mature, and much of the value will be in the informatics and distribution of the data, how you manage the data, deliver it. Those market pieces will become increasingly important. In the mid-term, we clearly have identified that as sequencing cost goes down, sample prep and bioinformatics become unbelievably important in the cost equation. We’ll put a lot of energy in those areas, because there’s a potential bottleneck on the back end if we don’t solve the bioinformatics problem. Similarly, when you think about doing millions of samples, you have to have sample prep that’s close to trivial.
X: How do you continue to keep the innovative edge for the next five to 10 years? Is it a combination of internal R&D and strategic investments in interesting startups, or a combination?
JF: It’s a combination. We certainly are investing very heavily in sequencing. You saw it with our HiSeq launch, and what we’ve been able to do to keep our own road map going. We will make strategic investments like with Oxford, and we’ll make the occasional acquisition like Avantome. So I think it will be a pieced-together strategy of those three things. One of the things that’s important for Illumina is that we remain the kind of company that other small companies would like to be part of. If they had to be part of somebody else, we’d want to be the one they’d pick. That has to do with a culture, and our external perception of how innovative and creative we are, and the kind of attention we pay if we get into an acquisition discussion. In some of the acquisitions we’ve made, the feedback we’ve gotten is that we deal with that process much more effectively than the larger companies do. And that’s important for us to maintain.
X: One of the things Paul Schimmel brought up at the Xconomy event [March 31] was that the pace of data coming off these machines is far faster and greater than our ability to make sense of it with informatics. Why hasn’t somebody stepped up and become the Illumina of bioinformatics? Or will Illumina become the Illumina of bioinformatics?
JF: It’s a great question. There are a couple pieces to the answer. One is that from a venture capital perspective, if you look historically at bioinformatics companies, it’s road kill. There’s almost no examples of very successful bioinformatics companies. The main reason is that people don’t want to pay for software. It doesn’t have the perceived value that instruments or reagents have. So people are used to the $29 price point for software, or less. Now it’s the 99 cent app. The mentality that people have just doesn’t support a big business. That tends to push you more toward the model where it’s done by the large existing companies, like Microsoft, Oracle or Google, or it’s done in the companies developing the sequencing technology like us and Life [Technologies]. There will continue to be little startups, but the chance of developing a billion-dollar company? That’s close to zero, I think.
The other comment I’d make about the data streams, I’d say one thing. We’ve done a lot of work to reduce the volume of data from when we started. As the throughput has gone up by a factor of 100, data coming to a disk drive is less than it even was three years ago. That’s because of pre-processing, before it has to be written to a disk. So from a data volume management standpoint, the market is keeping up there. The challenge is in creating actionable information from the data. That’s really the big issue. Frankly, it’s a challenge everybody will face in the next couple years.