What? Patents are not secret. Patents are the opposite of secret. The very word "patent" means "open". Are drug patents some kind of trick to do the opposite of what a patent is supposed to do?
> However, when the researchers designated some atoms and bond as untouchable, the program proposed plans that avoided the patented ones.
Can someone explain just exactly how do drug patents work? Can this legally work? I understand that normally patents cover processes or working inventions. I didn't think that the precise chemical bonds were required to be produced in the final output of a patent.
Also, since apparently the researchers have read the patents in question and designed their software to produce drugs that avoid those patents, have they willfully infringed? Since they were aware of the patent and the process by which the drug worked, so using this knowledge they produced similar drugs that do the same thing?
I assume that the researchers know drug patent law well enough and are confident that their method avoids patents, so there's obviously a lot I don't understand.
As you can imagine there are thousands of ways a moderately complex molecule could be deconstructed and the bank of known reactions is more than any one person can really grasp. That and the myriad patent literature and how cryptic and dense it can be make the problem particularly suited for algorithmic treatment. This is only being done now because cataloging all the research and patent literature (which goes back to the late 1800s) and digitally formatting in a way that allows it to be computationally analyzed and processed requires manual translation of each report by a human being. Not many people have the expertise required for this and those that do usually prefer less menial work. So it has taken this long to amass a digital library of sufficient size to give one confidence that the answers it provides are comprehensive and that further searching would be pointless.
It seems what they are attempting to sell is a IP protection service: let us show you other molecules and synesis routes you should also patent defensively, without actually attempting them in practice. Pretty soon you will have algorithmically defined patents.
It's IP anti-science at it's worst.
A big part of it is that drug patents cover the process of production that gives the final molecule with viable yields. Pharmaceutical companies try to write patents broadly enough, or just file enough different patents, that there's no way to get to the final product without infringing one or more of their patents. The idea here is to come up with a sufficiently novel synthesis pathway that the original patents don't cover it, and thereby don't apply.
Derek Lowe has a good discussion: http://blogs.sciencemag.org/pipeline/archives/2019/01/18/aut...
The patent can also be thrown out if it's to broad, no? The broader it is , the better the chances that there's prior art. Sure it costs money, but then there's money to be made so someone will sue
> since apparently the researchers have read the patents in question and designed their software to produce drugs that avoid those patents, have they willfully infringed?
the patents cover process, not the resultant molecule. the program finds a different process to make the same thing.
I was really surprised by that. So when Elizabeth I sold a patent for the monopoly of salt to some duke in order to raise gold for the crown the word "patent" meant that it was a public rather than a secret decree.
I'd like to see any evidence at all for this.
The closest thing we have to a controlled experiment is the software industry, where patents are a) difficult to obtain, at least in a way analogous to patenting a molecule and b) not culturally common among large portions of the R&D community. There is a huge amount of R&D into better databases, better operating systems, better distributed systems, better UIs, better programming languages, better compilers, etc. etc. that happens with no patent protection, and often with no license fees or other direct monetary compensation at all. Microsoft being willing to keep their source code secret and patent their techniques didn't help NT defeat Linux one bit, and certainly has caused no loss of people willing to work on Linux.
As for the article, changing the molecule can have unpredictable effects on how the compound acts in the body. Even changing one bond/atom can result in dramatically different properties. The article doesn't address that.
Doctor: I judge you to have [medical condition]. Here is a prescription for [expensive on-patent drug].
Patient: I would prefer to take this less expensive drug, which a computer program thinks might be equivalent.
Doctor: Hmm. Doesn't look like there are any clinical studies for this analogue drug. Since I don't want to risk a malpractice lawsuit, I'm not going to prescribe it. Without a prescription, your insurance won't pay for the analogue drug either, so it'll have to be cheaper than your copay in order for it to make sense for you.
DOCTOR: You have [medical condition]. We can manage it with [medication which is on the appropriate tiered formulary for the patient’s pharmacy benefits manager].
PATIENT: I would prefer to take this medication they advertised to me during the last football game.
DOCTOR: This medication is only a slight incremental improvement over the generic one approved by your insurance. If you want the newer one, it may have higher copays and will have more paperwork hassle as we try to get the insurance to cover it. In the end, it may not be covered. Do you still want it?
PATIENT: Yes. [Leaves angry Yelp review over how the physician wouldn’t prescribe them X medication].
END SCENE
As for the final point, if there are no clinical studies (at all) for a given drug, then it is not FDA approved and would be illegal to prescribe unless under the auspices of a registered clinical trial with a filed IND with the FDA.
If it is FDA approved but not for that specific indication, then it depends on the clinical practice patterns at large whether this is a reasonable course of action vs malpractice. Case in point: Avastin is way cheaper than Lucentis for macular degeneration and works via similar mechanism, but was never tested or approved for that indication. Many Retinologists will inject Avastin anyway because some trials have shown benefit for cost conscious patients or patients from whom insurance will not cover Lucentis. This is not malpractice.
Using another drug from another class entirely without any clinical evidence to back up its use outside of a trial could be considered malpractice, but only provable if there is a bad outcome.
Without a prescription, they aren’t getting the drug legally. With a prescription without insurance/PBM approval, then yes, it could be very expensive for unconventional treatments.
If this software can come up with sufficiently novel pathways other processes can be used to come up with the same drug.
This is the idea anyway.
Sounds like this is for finding different synthetic routes to the same molecule. So it might help avoid method claims, but not claims for the product itself.
