!20151022  Will drug discovery tools (e.g., IBM's KnIT) make 103 attacks easier?

But first, Washington Post has an interesting article on a case before the Supreme Court involving wholesale and retail sales of electricity, as governed by state and local laws. The case is FERC v. EPSA. The case has nothing to do with patent law, but as it involves a lot of issues dealing with technology and economics (as does patent law), it is worth following to get further insight into the Supreme Court's very strange thinking about technology in general. Article at: http://www.washingtonpost.com/news/energy-environment/wp/2015/10/20/the-electricity-innovation-so-controversial-that-its-now-before-the-supreme-court/.

The New York Times has an article about a strange tax-sheltering game that was played by Starbucks in Europe, involving corporate shells in the Amsterdam and London. Article at: http://www.nytimes.com/2015/10/22/business/international/european-inquiry-focuses-on-a-mysterious-starbucks-business.html. Here is part of the manuevers involving IP:

Emerald City owned Alki, which was set up in London to house Starbucks’ intellectual property. The intellectual property included logos and the recipe for roasting coffee beans, which Starbucks subsidiaries pay Alki a royalty to license. Because of its structure, Alki was not subject to corporate tax in the Netherlands or Britain.

In 2008, Dutch tax authorities issued a new tax ruling for Starbucks. After that, the Amsterdam coffee roasting business began to send large royalty payments to Alki for use of the bean roasting recipe, which the commission said on Wednesday appeared vastly inflated for what it was.

The recipe was basically the temperature for roasting beans, and appeared to be more like instructions than intellectual property. Yet counting it as such allowed Starbucks’ roasting unit to reallocate most of its profit to Alki in the form of royalties, the commission said, nearly wiping out the Dutch tax bill. No other Starbucks companies or roasters paid royalties for the same information, the commission said.

---    Will drug discovery tools (e.g., IBM's KnIT) make 103 attacks easier?

      I have talked occasionally on how automatic drug discovery tools will make it easier to attack drug patents as being the "predictable" (KSR) obvious result of activities of one skilled in the art. Indeed, one discovery question that should be asked now in all drug patent infringement lawsuits is "To what extent did you use drug discovery tools in your research - which tools, which databases, etc.". That information, in the hands of an analyst such as myself, is quite useful for preparing 103 invalidity arguments consistent with the research-nonsense of KSR (the Supreme Court's use of "predictable" doesn't correspond to the realities of R&D).

      To put this in perspective, here is an example of one automated drug discovery, based on an automated drug system, KnIT, being developed by IBM. As you read the following excerpts, keep in mind the "predictable" language of KSR.

Automated hypothesis generation based on mining scientific literature
Scott Spangler et. al. (IBM Research San Jose, Univ. Texas, Baylor College)
ACM KDD Conference 2014, 1877-1986
[Abstract]: We present an initial case study on KnIT, a prototype system that mines the information contained in the scientific literature, represents it in a queriable network, and then further reasons upon these data to generate novel and experimentally testable hypotheses.
[Abstract]: KnIT combines entity detection with neighbor-text feature analysis and with graph-based diffusion of information to identify potential new properties of entities that are strongly implied by existing relationships.
[Section 4:: To approach this problem [dealing with p53 kinases], we first note that there are over 240,000 papers that mention one or more of 500+ known human kinases in their Medline abstract. ... In our initial exploration, we searched all kinases but removed abstracts that make any reference either to p53 or to a second kinase. We thus excluded data that would trivialize the predictions. This left us with 259 kinases in all. Of these, 23 were known to be p53 kinases.
[Section 6.4: The algorithms described above were used to rank 252 kinases (those with at least 20 publications) by likelihood of being p53 kinases. As expected, most kinases known to phosphorylate p53 were near the top of the ranking list. Five kinases on the list not known to phosphorylate p53 were then tested by biochemical and molecular biology experiments for their ability to interact with and phosphorylate p53. ... The two sets of experiments argue strongly that computationally predicted top p53 kinase candidates PKN1 and NEK1 are indeed true p53 kinases.

      Question: if IBM tried to patent PKN1 and NEK1 as new, useful p53 kinases, (which they are), and ignoring the vomit nonsense of current 101 caselaw (All-lice, Bilgski, Hold The Mayo) - would this patent have to be ruled KSR-obvious because the use of the tool against known databases and then standard chemical testing of computer chosen candidates - is that something that is the "predictable" result of what one skilled in the art is supposed to do? I will argue that for these two specfic chemicals, they are not patentable as p53 kinases discoveries because of KSR and all of the above automation.

      Does this tool (and an increasing number of others) become an easier way for rejections based on hindsight analysis? Someone patents a new drug, and I gather all of the possible journal article abstracts (and texts) of relevances - easy to do once I know the patented drug - load them into IBM's KnIT, get some list with the patented drug near the top, and do the simple confirmation experiment. At this point, I would then argue that at the time of the invention, it would have been obvious for one skilled in the appropriate area of drug chemistry to have done this, and thus no patent.

      So I see two lines intersecting in the future. The first line is the increasing power of automated drug discovery tools. They will become more and more powerful, as programmers integrate more databases together, and drug modelling tools become more powerful. This curve will continue to increase. The second is the decreasing coherence of 103 (and 101) caselaw, where judges continually ignorant of the scientific process continue to write decisions as poorly reasoned as KSR (which is almost as vomit-y as Alice). This obviousness problem will become increasingly worse for the chemical/pharma industry, as companies such as IBM push the development of such tools to find profit somewhere.

      But IBM won't be happy just undermining pharmaceutical patenting. From their "Conclusion and Future Work" section:

We also feel that the general approach of mining literature to identify hidden relationships between entities is not confined to cancer or even to biology, but is a general tool that can be applied in almost any science. The potential for dramatic acceleration of discovery in all sciences ....
In short, IBM foresees a future where every future invention is nothing more than the predictable results of one skilled in any art using such tools - that is, EVERYTHING WILL BE OBVIOUS AND THUS NOTHING WILL BE PATENTABLE. One more note: "This work was supported by DARPA, the NSF and the NIH." Why do these government agencies want to help fund what could undermine the entire U.S. patent system and its role in growing the economy? Do they understand the long term risks to the U.S. economy if the above two curves intersect?

      Now, consider the following kinase drug discovery patent, and ask yourself, "Are these people idiots for patenting software that could make it impossible for anyone using the software to patent anything discovered in part by the software?":

Drug discovery methods for Aurora kinase inhibitors
U.S. Patent Application 2011/269732; 2014/141099
Vertex Pharma (filed April 2008) Abstract:
The present invention relates to drug discovery methods, particularly methods for assaying compounds for activity as Aurora kinase inhibitors. This invention also relates to a pharmacophore describing compounds that are able to promote a conformational change in the protein AuroraB and whose binding constant for the two-step process is given as Ki*. Finally, this invention also relates to compounds having the features of the pharmacophore.
Why waste money on the method patent? If your drug discovery tools actually work, DONT TELL ANYONE ESPECIALLY PATENT BUSTERS LIKE ME, keep the methods secret, use the software methods to find new Aurora kinases, and apply for patents on these new kinases. But geesh, if I see you guys trying to apply in later years for such patents, I am going to argue: "You want to claim a product your PHOSITAs discovered with your PHOSITA drug discovery tools?"

      Speaking of Hold the Mayo, et al, consider the following claim issued in July of 2015:

U.S. Patent 9,074,016
G protein-related kinase mutants in essential hypertension

Claim 3. A method of identifying putative anti-hypertensive agents, comprising: comparing electrolyte output of a first transgenic mouse of claim 1 administered said agent with the electrolyte output of a second transgenic mouse of claim 1 not administered said agent, whereby an increase in the electrolyte output of said first transgenic mouse compared to said second transgenic mouse identifies a candidate anti-hypertensive agent.
How did this get by Alice/Mayo? It is nothing more than:
- test a new drug in a test mouse
- compare blood of test mouse with control mouse not given drug
Since all tests of something new require comparison to a control, this claim reduces to: TEST A NEW DRUG IN A TEST MOUSE.

IPISC - your patent insurance experts:

-- Need protection from frivolous patent infringement lawsuits?
-- Need protection from AIA attacks? From troll attacks? From IPR-troll attacks?
-- Need help asserting/leveraging your patent(s) without being forced to sell out to trolls and "monetizers"?
-- Need to borrow funding using your IP portfolio as collateral?

Call IPISC to learn about how these cost-effective, risk-minimizing, insurance-based protections are available for your quality patents in U.S., Canada and Europe.
IPISC - http://www.patentinsurance.com , 800-537-7863.
This PATNEWS sponsored by Intellectual Property Insurance Services Corp.

Greg Aharonian
Internet Patent News Service
Your "Judicial Counter-Errorism Expert"
In U.S. – 415-981-0441
For a comprehensive collection of patent quality statistics, see:
For a brochure on my patent search services, see: