Friday, January 6, 2017

Genomic Prediction Patented???

Hat tip to Dr. John Cole, USDA, and Jesse Hoff for bringing this to my attention on Twitter.

I need to make an important point: I am not a patent lawyer. I do not have a deep understanding of patent law and nothing in this post should be considered legal advice or council.

On November 8, 2016 Canada issued a patent to Ben Hayes and Michael Goddard for genomic prediction. Genomic prediction is estimating an animal's breeding value (i.e. genetic merit) using DNA variants. See the Canadian Intellectual Property Office website for more detailed information about the patent.

Meuwissen, Hayes and Goddard first published their work on genomic prediction in 2001.
Hayes and Goddard filed for a patent of the method on December 21, 2007 with the United States Patent and Trademark Office.
About 23 days later, the Illumina BovineSNP50 Beadchip, a DNA test that can test 54,000 DNA variants quickly and cheaply for cattle, was released at the Plant and Animal Genome Conference. This 50K SNP chip was developed, in part, to deploy genomic prediction in North American dairy and beef cattle breeds.
This USPTO patent has not been granted. See https://assignment.uspto.gov/patent/index.html#/patent/search/resultAbstract?id=20120144508&type=publNum. However, the Canadian patent has.

A few points I want to make.
In 1997, Nejati-Javaremi, Smith and Gibson published a method they call total allelic relationship, which is basically genomic prediction by a different name. They even state in their article that genetic markers (not the actual DNA variants responsible for differences, i.e. causal variants) could be used to measure total allelic relationship. This method is basically the single-step genomic prediction that we use today, and in my non-legal opinion this probably represents significant prior art to the work of Hayes and Goddard.

Further, the Hayes and Goddard patent is limited to populations in which the effective population size (a measure of diversity, the size of a randomly mating population) is smaller than 1,000. I believe it could be argued that genomic predictions of multiple-breed data sets could have effective population sizes larger than 1,000.

In related news, Meat and Livestock Australia is fighting a Cargill USA and Branhaven LLC that basically allows them to patent all cattle DNA, see "MLA launches Federal Court bovine genome patent fight" by James Nason. In another article by Sue NealesRob Banks, University of New England, stated the Cargill/Branhaven patent is “intellectually and morally indefensible”. With this patent, the amount of prior art is even more staggering, basically all of modern cattle genetics. I don't know where to start or end with this patent...


Cattle breeders have fought other patents of animal breeding technologies. In 2014, legal battles started between the American Simmental Association versus Leachman Cattle of Colorado, LLC and Verified Beef, LLC. Ultimately, in August of 2016, the American Simmental Association won these cases and the Leachman patents were overturned.

Unfortunately, I worry that the American Simmental Association and their genetic evaluation arm, International Genetic Solutions, may have additional legal battles on their hands. Of the twelve breed associations that are included in the IGS multiple-breed evaluation, five are Canadian associations. International Genetic Solutions recently announced that they have released a multiple-breed single-step genomic prediction of stayability. They will continue to release multiple-breed genomic evaluations for additional traits in the coming months. In Canada, these genomic predictions could be viewed as infringing upon the Hayes and Goddard patent. 

We live in interesting times in the beef breeding world... 

Here is hoping for more patents to be overturned.

UPDATE: 6 January 2016
I received word on Advantage Cattle from JohnD that the US patent on genomic prediction had been rejected just yesterday on January 5th. http://www.advantagecattle.com/forum/topic.asp?whichpage=24.6666666666667&TOPIC_ID=10783#166702 

Most of the US rejection cites a Monsanto patent application titled "Marker assisted best linear unbiased prediction (ma-blup) software adaptions for practical application for large breeding populations in farm animal species." This patent was abandoned as an incomplete application in April of 2012. From the USPTO website, "Abandoned means that the application is no longer pending and, thus, cannot mature into registration."
The examiner referred to the Monsanto patent application as prior art, and rejected the Hayes and Goddard patent application in the US. Perhaps the same arguments can be used in Canada to overturn the granted patent?

Also, I should point out that I believe both dairy and beef industries, especially in the affected countries, should band together to fight these patents. I do not think that the American Simmental Association should be asked to carry the water again on this issue. But, I do fear that they might be a target of litigation. 

Post a Comment