Big day for gene patents
I’m back from defending my thesis (apparently successfully), just in time for big news in the biomedical patent world. The summary judgment ruling against Myriad Genetics and its BRCA gene patents, announced yesterday, brings up many questions about the future of patents covering genes and, potentially, other biological phenomena. Obvious, however, is that the US patent office (USPTO) was shamefully uncritical of the claims from the original patent applications, a problem that extends to many, many similar patents.
Valid patent claims are meant to be narrow and novel but, by claiming invention of a short (15 nucleotide) DNA sequence, the BRCA1 patent clearly violates these criteria. This is the conclusion that Duke researchers elaborate in a recent investigation, where they elegantly demonstrate what should be obvious to anyone with a minimal understanding of statistics (which was clearly not the case for the patent examiner). The BRCA1 patent claims (in language similar to many gene patents): “An isolated DNA having at least 15 nucleotides of the DNA of claim 1″ (claim 1 being the protein sequence BRCA1). Thomas B. Kepler, Colin Crossman, and Robert Cook-Deegan at Duke calculated that BRCA1 contains 5,575 individual 15-nucleotide sequences (15-mers). Since the same protein sequences can be encoded in different DNA sequences, there are actually 1.6 × 106 15-mers that could encode the patented sequence of BRCA1. As there are only 1.07 × 109 possible 15-mers (DNA only has four letters), the patent could actually cover 1 in 600 of all possible 15-nucleotide DNA sequences (1.6 × 10e6/1.07 × 10e9). The authors calculate that an average human gene would have 15 of the 15-mers covered by the patent! It should be noted (and this is a compliment, not a criticism) that this theoretical analysis requires nothing more than a calculator. Searching for DNA or protein sequences can be done, for free, by anyone with a web browser; it was a bit more difficult, but not impossible, when the Myriad patents were filed in the late ’90s. The merit of patenting such sequences may be debated as a point of policy; their novelty and uniqueness may not.
The court didn’t look into such specifics, but rather at the broader applicability of patent law to genetic information Yesterday’s ruling focuses on whether the BRCA patents fall within the realm of “products of nature”, which have been held to be non-patentable. The court found the genes to, indeed, be such a natural product.
Though there will certainly be much to debate in the ruling (which, no doubt, will be appealed), one of my favorite parts is a not-so-subtle rebuke of the USPTO, as well as Myriad’s argument that the government’s wisdom in granting patents should be respected:
The Federal Circuit has previously held that it owes no deference to USPTO legal determinations. See, e.g., Arnold Pship v. Dudas, 362 F.3d 1338, 1340 (Fed. Cir. 2004) (“This court reviews statutory interpretation, the central issue in this case, without deference.”). While Congress has created a presumption of validity for issued patents, approximately 40% of patents challenged in the courts have been found invalid, demonstrating that this presumption is far from absolute.
