Today, if your cancer no longer responds to conventional treatment, your oncologist may sequence the tumor cells and direct your treatment according to the profiles of genes that have acquired the mutations. Formerly, your treatment would have been based on the appearance of your cells under a microscope or on a pathologist’s theory about the tissue of origin. And nowadays, if your child gets cancer and goes to St. Jude Children’s Research Hospital in Memphis, Tennessee, his or her care will be excellent and heavily subsidized — they will almost surely take a sample of cancer cells and sequence their genome soon after you get there. If your child has an unusual clinical profile that physicians think might be a “syndrome,” although they can’t pin it down, you are apt to find your way to a center that starts by sequencing your child’s entire genome, tripling the odds that doctors will be able to find a diagnosis and shortening your “diagnostic odyssey” by months to years, and probably also saving tens of thousands of dollars.
All these clinical practices did not happen a few years ago. What do they have in common? They all depend on fast and cheap DNA sequencing. Uses of sequencing are proliferating wildly, because sequencing is cheap and therefore seemingly everywhere. Most medical goods and services are expensive when they are introduced — a feature they have in common with many other high-tech goods and services. But in medicine, they tend to get even more expensive over time, while digital goods and services get cheaper. DNA sequencing diverged from the medical pattern. Its reduced cost and increased use more closely resemble information technologies like computer memory, disk drives, and integrated circuits.
Drawing on the “undiagnosed disease” scenario, the typical patient entering treatment recently has been careening from doctor to doctor, specialist to specialist, center to center for two or three years, and has spent $5,000 to $10,000 on laboratory tests alone (and far more on visits with health professionals). Those tests may generate data about a handful of genes, proteins, and metabolites; for more or less the same money, whole-genome sequencing produces data on more than 20,000 genes, as well as on all the other sequences in the genome. In some cases, the genomic data are the most clinically important information. Often the genomic data are not the crucial missing elements; in those cases, the genomic analysis will not be terribly useful. But compared to other laboratory tests, the information-to-dollar ratio is much higher.
Uses of sequencing are proliferating wildly, because sequencing is cheap and therefore seemingly everywhere.
In sequencing, the original human reference genome for the Human Genome Project, which ran from 1990 to 2003, was largely based on technology that had been prototyped at Caltech in 1986 and then developed and refined at Applied Biosystems (ABI) between 1987 and 1998. That technology was based on fluorescent labeling of the ends of DNA molecules being copied using naturally occurring enzymes, and then running them on gels or through capillaries that separated the DNA molecules by size. It was breathtakingly elegant, building on the quiet genius of Fred Sanger and Alan Coulson at the University of Cambridge. And while it was miraculous that it actually worked, it did. Lee Hood, the Hunkapiller brothers, Lloyd Smith at Caltech, and the ABI team made it possible to develop an automated sequencing instrument. ABI sold its sequencing instruments to both the private company Celera and the publicly funded high-throughput sequencing centers that raced to do the initial human reference sequence, along with mouse, fruit fly, worm, yeast, mustard weed, and E. coli bacteria sequences. But it was still expensive.
That all began to change, even as the human reference sequence neared completion. Entirely new approaches to DNA sequencing appeared. They were much faster, much less expensive, and much more amenable to scaling for high-throughput sequencing. Was this magic? Hardly. Rather, it was the result of making DNA sequencing itself the target of research.
The National Human Genome Research Institute (NHGRI) incorporated this idea and developed a research program to produce technology to do science; yes, technology to do science. DNA sequencing was front and center. It was also somewhat countercultural at NIH.
Peer review of individual projects initiated by investigators is the solid foundation for the NIH and the National Science Foundation. But peer review is not always the best recipe for pushing engineering and technology forward. Sometimes the federal government is central to developing technologies with very broad applications that reach well beyond their origins. The salient example is the Information Processing Techniques Office of the Defense Advanced Research Projects Agency (DARPA). DARPA was the origin of many technologies: the internet, digital communications, and global positioning systems, for example. Its style of funding depended on experts who gave out federal funds to promising projects and held the program managers who in turn dispensed the funds accountable for results, while also tolerating failure. Most of those experts rotated in and out of academia, industry, and government. In the 1990s, some of us asked if NIH needed a DARPA to supplement its peer-reviewed, investigator-initiated research programs.
It turned out that the NIH already had something like DARPA, albeit in embryonic form and adapted to the norms of peer review. At its National Cancer Institute, the NIH had developed an expert community focused on genomic technologies and informatics. And over at NIH’s genome center, founding director Jeff Schloss had done something similar with advanced genomic technologies, with heavy emphasis on DNA sequencing. This program began small, but it became the single largest program focused on DNA sequencing technology.
It was not a question of academia versus government versus industry. All strands of the triple helix were essential.
The program funded work at many universities, and also at companies that developed what became known as “next generation” DNA sequencing. The NHGRI money was not always the source of the idea, but NIH’s DNA sequencing technology program helped fund most of the technologies that gave rise to the current DNA sequencing technologies.
The current market leader in DNA sequencing is Illumina, which started out by developing bead-array technology. It moved into sequencing by acquiring Solexa, a British start-up. NHGRI’s program supported many of the investigators through the perilous early stages.
It was not a question of academia versus government versus industry. All strands of the triple helix were essential. Governments played a crucial role, particularly in the U.S. and the U.K. They supplied capital to support research that seemed promising. Would the academic and early start-up research that gave rise to the next-generation sequencing technologies have happened without government funding? Perhaps, but not nearly as soon and possibly not at all. And the academic researchers and the angels, venture capitalists, and entrepreneurs who gave birth to the start-up firms relied on government employees who distributed public dollars for support. Those start-ups were later gobbled up by larger firms who now dominate DNA sequencing, and that is another lesson from DNA sequencing technology: It does demand capital.
The role of government is worth special attention. It was not just the money. It was also the unique power of government to pull together a community of rivals who were also colleagues — a community driven to deliver technologies for real-world use. Government employees could talk to everyone, and if the experts respected and trusted the government employees, they could foster a sense of community. Jeff Schloss organized meetings of the DNA sequencing technology mavens. The Advances in Genome Biology and Technology conference became the event at which companies and academic researchers displayed the biggest new thing in DNA sequencing. At the 2006 meeting, Schloss told Bio-IT World: “If we give people grants and they don’t get any further than a publication in Analytical Chemistry … we’ve basically failed. We want these technologies out there.”
If you wonder how a human genome can be sequenced for a few thousand dollars in a few days, it is not by accident. It is the result of deliberate efforts to develop new technologies for DNA sequencing, initiated by government but involving academic research, industrial engineering and development, and private capital. And one of the crucial drivers of the technology that made it possible was highly expert, visionary government employees, aka the Schloss Effect.