Mar 21, 2012

The Holes in Whole-Genome Sequencing

  • This week, the Supreme Court handed down a unanimous decision overturning Prometheus Laboratories' dosage calibration patent. Where Prometheus saw a solid contribution to the technological arts, the Court saw an attempt to patent a formula.

    The amicus briefs were fascinating: Mayo Clinic, the ACLU, Cato Institute, AARP, LabCorp, and Verizon on one side; and Prometheus, BIO, Novartis, Myriad, the National Venture Capital Association, several intellectual property associations, and the Association of University Technology Managers on the other. The United States, Roche, Abbott, Microsoft, and several other IP associations wrote briefs in favor of neither party.

    The briefs made all sorts of arguments—from noticing useful aspects of the way lightbulbs were made were not a patentable idea, to how using a formula to determine how much acid to use in treating rubber was. They centered around an idea of whether or not patentable novelty was present. There were some interesting quirks—for the engineering cases cited, the threshold for novelty seemed higher. Where the folds in lightbulbs were decried as finding a new quality in someone else's old product, finding new uses for old molecules is the best when it comes to drugs.

    Older drugs have already run the regulatory gauntlet, and finding new indications can only extend their usefulness. But where the lightbulb argument fell down, Mayo's argument carries on: If Prometheus can patent some formula to use the numbers from a blood test, how do you compete? Mayo claimed some of the formula's parameters were too high or did not apply to a diverse set of patients. However, using a slightly different formula would obviously infringe on the claims.

    What about a more complex formula that still uses that blood test as one of its elements—does the use of the blood test for determining the dosage of this drug apply? Does this patent effectively limit the value of a test that Prometheus does not own?

    In a way, it's akin to discovering that some known blood parasite caused some sort of disease in particular patients. Your knowledge helped alter treatment regimes—when doctors see those symptoms, they know to send blood out for a look under the microscope—but if you had invented the microscope they used to look at it, you'd be in a much better patent situation.

    But this scenario hasn't acknowledged one big gorilla—medicine. Were this an industrial thing that you wanted to form into a company, you'd patent the equipment, keep the finer details to yourself, and go into business with the advantage of being the expert. In finance, computing, or anything, this information would (and should) be a trade secret, like the source code to a program—because that's what it is. This is medicine, however—you can't use the information secretly, but at the same time doctors cannot unlearn what they've learned—so informational patents like these are sure to cause grief on both ends.

    That's the hazard of developing a way to analyze data produced by a test the patent doesn't cover. I suspect that this ruling will extend to similar patents including Myriad. The ability to say that certain alleles correspond to certain added risks sounds uncomfortably similar, especially with the wide variety of genome sequencing techniques available.

    Remember Roche? The are trying to buy out Illumina, one of the big sequencing tech companies. Don't forget they still own 454, another big sequencer. I also didn't see Life Technologies, which has the Ion Torrent DNA sequencers, filing an amicus in favor of Prometheus.

    The $1,000 genomes, what use are those if you can't use the information without negotiating a patent thicket? I can't imagine a case where a patient would have their doctor run a whole genome sequence on them and then request anything but the latest and greatest in analysis, using all of the information currently available.

    If we are to enter the next stage of personalized medicine, we cannot do so by parceling out information—genomes do not work that way, pleiotropy doesn't stop at the lines drawn by the USPTO. And by allowing patents that stand on generic tests—given X data from whereever, do Y—we will have to wait a very long time to receive our test results.

    To be fair, the companies should be compensated for their work, but the field they work in is unlike any other. The comparisons to light bulbs, computer code, protocols for treating rubber, they do not fit perfectly for the same reason that drugs do not fit into the same category: the FDA. Complying with electrical codes and obtaining FCC approval for the latest gadget simply does not compare with FDA's requirements. The research wasn't cheap to begin with, either. I can't imagine how CLIA approval would work for a whole-genome leave-no-nucleotide-unturned approach.

    The consumer electronics/computer businesses addressed the problem of patent thickets with patent pools, consortia, and open standards. Even so, we saw gadgets appear and disappear from the market over patent disputes. Perhaps the FDA's exclusivity provisions could work to reward confirmed tests, or perhaps a consortium of testers themselves can pay a bounty, or the patients and their organizations could.

    This ruling was no surprise—there were signs that a patent process cleanup was necessary all the way back when the first human genome was sequenced, and Celera attempted to get patents on medically relevant human genes.

    So yes, we'll eventually have affordable whole-genome sequencing that will revolutionize personalized medicine. But, between the patent negotiations, regulation, and the complexity of the scientific problems at hand, here's a bit of free medical advice: don't hold your breath.

Add a comment

  • You must be signed in to perform this action.
    Click here to Login or Register for free.
    You will be taken back to your selected item after Login/Registration.

Related content