Founder Stories: Jay Flatley, Former CEO of Illumina

Good morning, and thank you very much. Last night I mentioned that when I had received word that I had been acknowledged as the honoree this year, I went back and looked at the website to see who the prior winners were and was immediately humbled by even being considered in the company of those who have gone before me for this award. So I want to thank you once again for that and thank the PMWC for the opportunity to present this morning. Over the past 10 years, the pace of technology and program development has been absolutely astonishing, while at the same time most of us feel like the impact that we've been able to make on clinical decision-making has been limited, if not actually disappointing. The biology seems to seldom cooperate as we might hope, and we wonder when the layer upon layer of complexity in the underlying biology will ultimately succumb to our ever more powerful tools. There's reason in my mind to believe that we may be at a tipping point where we have the ability to look at genomes, we can look at complete human genomes economically, quickly, and at great scale. And coupled with progress that we've seen recently on the regulatory front and a growing list of examples where understanding the genome has made significant differences in treatment decisions, extended lives, or in some cases even cured patients, I have great hope. So in the few minutes that I have with you this morning, what I'd like to do is do a quick retrospective on progress, put that in the context of the technology developments of the time, and share with you a bit about where I think we are today and some predictions for the future. So 10 years ago in 2004, Illumina introduced a product that was at that time revolutionary. It allowed the ability to examine 1,536 SNPs simultaneously across 96 samples and obsoleted the prior technology which had the ability to genotype 3 SNPs at a time. That product enabled a SNP consortium and catalyzed what became known as the Genome-Wide Association Era, the GWAS era, the first really concerted quest to try to understand the connection between genomics and disease. The technology moved extremely quickly in this timeframe from us and from many others, and we quickly got to the ability to genotype about 1 million samples at a time. And in March 2005, as part of the Human Genome Project, the NIH hailed the first complete sequence of the X chromosome. The International HapMap Project began about this time as a global effort in an attempt to understand the common patterns of genetic variation across populations. Using haplotype maps. And at the end of 2005, the NCI and the NHGRI launched the Cancer Genome Atlas Project, a 3-year, $100 million effort aimed at understanding the molecular basis of cancer. In 2006, genomics made it into proposed policy when then-Senator Barack Obama introduced the Genomics and Personalized Medicine Act of 2006, proposing appropriate utilization of molecular genetic tests helping to secure the promise of personalized medicine for all Americans. That bill never passed, by the way. The NGS era began in earnest in 2006 with a product from 454, the first commercial NGS company, only to be bought in 2007 by Roche and to be closed down 6 years later. We bought Solexa. We introduced the 1G analyzer. And after 9 months of really hard work, we actually got it to sequence a billion bases in a single run. And that run took about 3 days. The FDA updated the label for Coumadin to recommend considering a person's genetic makeup before recommending dosage. Consumer companies launched their first array-based products, and James Watson got sequenced on the 454 platform at 7x coverage. The first full-coverage human genome was of a Han Chinese and was heralded in Nature magazine with a cover entitled Your Life Is in Your Hands: Instructions for the Personal Genomic Age. Later that year, the NIH announced the launch of the Human Microbiome Project to better understand the assortment of bacteria, fungi, and microbes and how that plays in human health and disease. In 2008, then-President George Bush signed the GINA Bill, the Genetic Information Nondiscrimination Act. The International Cancer Consortium was launched. And in 2009, the first consumer sequencing program allowed doctors to order a complete coverage human genome sequence from a CLIA laboratory for under $50,000. The 1000 Genomes Project was initiated to create an extensive catalog of human variation by sequencing 1,000 individuals. In 2009, less than 5 years ago, the first human tumor sample was sequenced on a GA-IIx, the first machine that could sequence a full human genome in a single run. In 2010, sequencing leapt forward to achieve an output of 200G, driving the cost of sequencing a human genome down under $25,000. The applications of the technology broadened quite rapidly to include RNA expression, methylation, and many other applications. Sequencing full coverage genomes began in earnest with tens of thousands generated over the next 2 to 3 years. The first child saved by DNA sequencing, Nicholas Volker, then 6 years old, hit the media out of the Medical College of Wisconsin. I believe Howard's here and will probably talk about that. That same year, whole genome sequencing was used to identify a fusion oncogene in a patient with a difficult diagnostic case of APL. The results changed treatment and was published in JAMA in April 2011. Hospitals such as Children's Mercy in St. Louis began to diagnose and treat rare disease in a neonatal hospital setting. A personalized medicine approach allowed a young girl to walk again, Shelby Valiant, a young lady in Phoenix who had her whole genome sequenced by TGen. The American Gut Project kicked off to understand how diet, microbes, and the environment affect human health. From technology to science, from policy to programs, there's been stunning progress in the last 10 years. And the pace of change, we think, is accelerating. NGS is moving into the clinic at a breathtaking rate across reproductive health, rare disease, and cancer. 5 years ago, this chart would have been blank. In total, we now estimate that there's nearly 1,000 separate NGS tests being done around the world. Perhaps the, uh, the fastest adoption of any single diagnostic test in history has been NIPT, and over just the last 2 years, We're now at a point where over half the women in the high-risk group in the United States opt for NIPT testing rather than using traditional amnio. In retrospect, I guess none of us should find that surprising. Just over a week ago, we announced a new system that allows full coverage genomes to be sequenced at $1,000, including all costs other than overhead. We think this is really something that has broken the sound barrier in human genomics. And it's not only important psychologically, but it's critically important because of the scale of the projects that it will enable. So not surprisingly, NGS is at the center of my universe. At the moderate output range of this scale, the MiSeqDx product is now cleared by the FDA, so the first NGS product cleared through the agency with an associated assay, and it's being widely adopted for clinical applications. We believe the demand for factory-scale sequencing of human genomes is about to explode, as evidenced by a number of recent programs that have been announced, including one at the University of Cambridge to sequence 10,000 genomes and a very exciting program in the UK to sequence 100,000 genomes. So we just launched the HiSeq X Ten at the high-performance end of this spectrum, designed for population power. The HiSeq X is capable of sequencing 1.8T in under 3 days, delivering more than 5 complete full coverage human genomes in a single day. In 2007, it took 3 days to generate 1G of data, and today that can be done in 2.4 minutes. It's sometimes difficult to grasp the magnitude of numbers as large as a terabase, so to represent this visually, I compared the output level of the prior generation capillary technology to that of a HiSeq X. To achieve that same level of output would take on the order of 500,000 capillary sequencers, or enough to stretch from our headquarters in San Diego to Disneyland and back. The level of output and cost, we think, brings transformative capabilities to the world of discovery. It will drive the most important factor, in our view, that limits the widespread adoption of personalized medicine, and that's enhancing the overall clinical utility of the genome. This schematic illustrates our vision of how this ecosystem needs to work. First and foremost, we need to build very, very large data repositories to accept genomic and genotypic data from these very large population projects around the globe. And we think critically that these datasets need to be shared through some sort of federation mechanism that allows local control of the datasets but provide the ability to create an integrated layer of metadata that's shared and fully open. Researchers would have access to these datasets at the raw data level, while clinicians would typically access the data at a metadata level. Individual patients get added to this database as diagnosis and treatment decisions get made, creating an adaptive learning environment where those before them improve each subsequent diagnosis and treatment decision. So we think convincing data generators to share data is a hard problem. Most continue to feel that proprietary control will provide some economic or social benefit that would be diluted if they decided to allow more general access. As a company, we fully support a number of efforts that are now underway to try to encourage sharing of these very large data sets that are getting created around the globe. In oncology, the long-awaited ability to sequence every cancer patient is now possible. Many of the large-scale programs that I talked about have cancer as a focus, and we're convinced that getting 1 million well-phenotyped oncology patients into a database in the next 5 years is not only doable, but is also necessary if we hope to accelerate progress. Such datasets will dramatically improve our ability to predict cancer, to affect the landscape of companion diagnostics, and potentially open up miraculous opportunities in using cell-free DNA as a screening test for oncology. The goal is to arrive at a future where diseases are classified in a new molecular taxonomy with companion therapeutics and a set of tools that allow disease to be monitored via a molecular stethoscope. To realize this vision, there's a whole lot more we have to do. From a technology perspective, the basic tool sets in hand, but improvements are necessary across quite a wide spectrum, including things like challenging samples, FFPE samples and low input, standardization around data formats and data ontologies, the ability to aggregate genotypic and clinical data, and very tough issues around data security, HIPAA standards, standards and compliance, and the ability to create improved annotation both clinically and biologically. So fortunately, we think these technological problems are fully tractable and I think are actually quite solvable over the next few years. The daunting challenge of dealing with payers, with reimbursement, with regulation and ethical questions, we think, remain. And most of you are much more qualified than I to take on those difficult issues. And I think over the next couple of days there will be many speakers from this podium that will talk to you about some of the frustrations on that front as well as the progress. Our hope is that what we can deliver in both— in terms of economic and medical value will overwhelm the hysteresis in the healthcare system and that force— that will force cooperative work all toward the common goal and that we as a company hope we can make that happen. Soon. So at the first consumer conference in Boston in 2010, I made a set of predictions outlining how I thought the field would evolve. And a couple of these have already happened. Sequencing clearly now is directly saving lives, or at least in many cases extending lives. Nations are beginning to sequence their populations. And complete human genome sequencing is now available for $1,000. But we still have an awful lot to do if we're going to meet my goal of driving cancer to a chronic disease in the next decade or get to the point where we have enough capability and solve the social and ethical problems of doing newborn screening of all babies at birth. In the meantime, those of us at Illumina will continue to focus on our mission of improving human health by unlocking the power of the genome. Thank you very much.