AI Patent Trolls Now on the Job for Drug Companies

Love it or loathe it, the pharmaceutical industry is really good at protecting its intellectual property. Drug companies pour billions into discovering new drugs and bringing them to market, and they do whatever it takes to make sure they have exclusive positions to profit from their innovations for as long a possible. Patent applications are meticulously crafted to keep the competition at bay for as long as possible, which is why it often takes ages for cheaper generic versions of blockbuster medications to hit the market, to the chagrin of patients, insurers, and policymakers alike.

Drug companies now appear poised to benefit from the artificial intelligence revolution to solidify their patent positions even further. New computational methods are being employed to not only plan the synthesis of new drugs, but to also find alternative pathways to the same end product that might present a patent loophole. AI just might change the face of drug development in the near future, and not necessarily for the better.

Many Paths to Progress

In most industries, a patent is a simple concept: come up with a new idea, and if it proves to be novel, non-obvious, and useful, chances are good that a patent will be granted that prevents anyone but the owner from making, using, selling, or importing the covered invention for a certain period of time. The rub to the patent process is that the application must reveal everything about the invention publicly, which means that after the exclusivity period has expired, anyone can profit from the original inventor’s work.

Pharmaceuticals, though, are treated differently. Since it’s relatively easy to reverse engineer a chemical compound using analytical chemistry tools and methods, patents for drugs concentrate on the process used to arrive at the desired endpoint. Most drugs are relatively simple organic compounds whose creation is a long, complicated series of reactions. They’ll often start with a couple of simple compounds, reacted together under just the right conditions to yield an intermediate compound. That product is then perhaps purified before being mixed with a fourth compound, and the process continues. Functional groups are added or subtracted at each step until the final compound is created in sufficient quantity and quality.

Every step in the process is claimed in the process patent application so that the resulting patent is as broad as possible. But it doesn’t stop there. There may be more than one way to skin the synthetic cat, and every single feasible alternative synthesis needs to be covered by the application too. Chemists at pharmaceutical companies spend a lot of effort looking for and plugging these potential patent loopholes.

AI to the Rescue

Both the design of the best, most commercially viable synthesis and the search for loopholes are perfect applications for AI. Syntheses can be broken down into well-defined steps governed by rules that an expert system can rapidly churn through, searching for a path from a known starting point to the desired product. Researchers at the Polish Academy of Sciences and the Ulsan National Institute of Science and Technology in South Korea had previously shown that an application called Chematica can autonomously generate a synthetic plan for a group of seven well-known drugs from simple starting materials. (Chematica was recently purchased by Merck and seems to have been rebranded as Synthia.) Each plan was generated in about 20 minutes and verified by chemists, who followed the synthetic instructions in the lab and came up with the right endpoints.

As a follow-up, the same team turned that process around, using Chematica to search for “retrosynthetic” paths to three new endpoints. This time, the products were blockbuster drugs with billions in sales and presumably air-tight patent positions. The researchers gave Chematica some rules, making certain key synthetic steps off-limits to the algorithm and forcing it to find alternate ways to the same product. To their surprise, paths to each of endpoints were discovered that successfully evaded the patent-infringing steps.

The implications of this development are potentially far-reaching. In the first instance, it seems like Chematica and similar tools will quickly become an aid to the intuitive, creative process of designing an organic synthesis. Such applications could also put a fair number of chemists out of a job, when it can do in 20 minutes what a chemist might take weeks or months to accomplish.

On the other hand, AI applications like these stand to stifle competition. The more airtight the patent position for a drug, the longer the patent owner can maintain a monopoly on that drug. Using AI to test for loopholes around the synthetic process only solidifies the claims, making it less likely that generic versions of a drug will reach the market in a timely fashion.

Taken at face value, the use of AI to both explore new routes to drug synthesis and find potential patent loopholes is a fascinating use of technology. But like anything else, the devil is in the details, and when such systems are inevitably put into widespread use, it’s likely that the ones that end up paying the price of progress will be the consumers.

39 thoughts on “AI Patent Trolls Now on the Job for Drug Companies

  1. I acknowledge the potential for abuse here (as with any new tech) but I think you’re minimizing one important step in their process:
    “The researchers gave Chematica some rules, making certain key synthetic steps off-limits to the algorithm and forcing it to find alternate ways to the same product. To their surprise, paths to each of endpoints were discovered that successfully evaded the patent-infringing steps.”
    While this potentially puts some experimental chemists out of a job & concentrates drug discovery in the labs of the rich, it increases drug turn over & competition between firms.
    If you can make a similarly acting drug without having to license the process, in theory, this could increase the number of generics on the market.

    Sadly this has another unintended consequence, at least in the US. Lots of medical products (more commonly implants or devices instead of drugs) can be sped through the approval process by citing prior work that has been approved. Unfortunately, regardless of if it’s been withdrawn due to side effects or not.

    1. Remember CYA step earlier? Them finding those paths first allows for that.

      As for.

      ” Patent applications are meticulously crafted to keep the competition at bay for as long as possible, which is why it often takes ages for cheaper generic versions of blockbuster medications to hit the market, to the chagrin of patients, insurers, and policymakers alike.”

      Socialize drug discovery instead of complaining about the few backed by a legal system created by the same policymakers.

      1. Add… with former legal and/or leadership from companies that cause illness and disease as the same policy makers. Conflict of interest violations? A Right to Know?

        The industries that cause disease seem to work in tandem with the synthetic industry whether it was mine tailings or runoff or petrochemical waste in the olden days dumped down the stream… to mine tailings and/or runoff and/or petrochemical waste used as precursors to food and drugs (more likely as poisons). There are connections with products that are brought to the main stream as being beneficial with more-so now days being observed to profiteering motives.

        I almost think there is an act of war going on as there is some report a Chinese Military official released where they planned to weaponize pretty much everything they can… including via drugging to wage war against the U.S. and abroad. https://www.youtube.com/watch?v=Q6SIzeB3qy0

        Their rich were buying out the U.S. officials. Liberals have been investing in China longer than conservatives.

        This can be connected with the cartels and the large volume of flow now coming from malicious China versus domesticated. More reason liberals want flows coming into the U.S. illegally and elsewhere so they can traffic more around the ports.

        Not thinking James Burke would get a budget for a new series regarding. However, would be an interesting series.

        This is about as close as it’s getting at the moment:

    2. “If you can make a similarly acting drug without having to license the process, in theory, this could increase the number of generics on the market. ”

      No. Because each generic would have to go through the full testing process. That makes it cost prohibitive.

  2. This:
    “Each plan was generated in about 20 minutes and verified by chemists, who followed the synthetic instructions in the lab and came up with the right endpoints.”

    And this:
    “Such applications could also put a fair number of chemists out of a job, when it can do in 20 minutes what a chemist might take weeks or months to accomplish.”

    The latter are too busy verifying the former. But anyway this CYA seems like a supreme waste of effort that could be directed more productively elsewhere.

  3. Alternate/obfuscated pathways to “designer” recreational pharmaceuticals in 3…2…1…

    Having just dealt with the LOC copyright office (that doesn’t yet accept large uploads, nor 12 year old file formats), I’m not holding my breath for the DEA or ATF to pick up on this anytime soon.

    1. They only selectively enforce the laws… in my experience wrongfully delegating downstream to the lower court rackets to obstruct justice more with lower level gross frauds and cheats. The USC exists like with designer drugs… though they acted like they needed new legislation like the last two designer drug acts that weren’t required to enforce the USC. Literally, criminally insane and unlawful is in enforcement positions. DEA are the most dangerous historically with DHS now days co-operating with malicious Sheriff’s taking over I’m guessing with FBI. ATF seems administratively corrupt and unconstitutional… though on down they seem valid.

    2. There’s not a huge market for that despite the massive markup in the black market. Legislation moves slowly so simply swapping out a minor functional group creates an analogue that isn’t illegal. Club drugs are particularly prone to this even now.
      Following that, the faster ‘alternate’ pathway is generally switching from a Chinese pharmaceutical supplier to an Indian one that isn’t under customs scrutiny..

      1. I don’t believe that is what is explicitly stated in the USC. Enforcement, Prisons and Courts make excuses to bust people because they “smell” like marijuana and aren’t the pedo incest homo junkie kleptos masked armed robbing with more dangerous narcotics, controlled substances analogues and per 21 USC 802 section 7 & 9D:
        “any drug which contains any quantity of a substance which the Attorney General, after investigation, has found to have, and by regulation designated as having, a potential for abuse because of its depressant or stimulant effect on the central nervous system or its hallucinogenic effect.”

        Is not meeting 21 USC 802 section 15:
        “except that such term does not include the preparation, compounding, packaging, or labeling of a drug or other substance in conformity with applicable State or local law by a practitioner as an incident to his administration or dispensing of such drug or substance in the course of his professional practice. ”

        Is a 21 USC 802 section (32)(A) “controlled substance analogue”

        Prosecutors make excuses since some are like as bad as the Michigan Judge that got a mild Prison sentence for committing ‘Sexual Torture’. I mean… you have to be on narcotics and dangerous drugs to get into that. It’s like Jeffrey Dahmer officials making serious mental facilities injuries to enslave to do unlawful acts. Some prefer to drug poison their pandered and trafficked victims more masked to rob and maliciously destroy personality, property… you know… people. Who knows what the gross frauds and cheats false pretense claims will be… oh wait… there are only so many combinations and permutations they can iterate through and be logically… uh… huh… be per the spirit, intent and letter of the law???

        There is a reason the county where I almost was murdered in after working in pharma that I thought was more ethical… lost their seizure war profiteering scheme to the State of Michigan. The counties are maliciously unlawful intentionally and to me meet the definition of imminent threats.

        1. The short answer is, it depends.
          There is the chance (or high likelihood in some cases) of prosecution under the Federal Analogues act but that sort of depends on how much you piss off Barney Fife or your local DA. They might not know what you have, or really care to find out. Then there are other drugs like LSZ ( an LSD derivitive, but not quite / maybe analogue) that are as yet, unclassified. The line between ‘analogue’ & ‘molecule which has had significant large additions to it’ is very, very, legally grey.

          At any rate, if you’re determined to buy club drugs, spend the extra few bucks on Marquis Reagent or some other test kit to make sure what you’re buying is what you’re seeking.

          1. Right, test for sure. That’s why I used to grow… I hated buying and like food… who knows variables with not just pissed off narcotics divisions, thinks they’re the warden/guardian/conservator/parent/somethingthey’renot, courts and/or prosecutors.

            Don’t forget Freedom of Religion Rights also can be used as a defense.

  4. If pharmaceutical companies take this approach moving forward then when the patents expire there will be a wealth of all known ways to make this helpful compound. I don’t mind the greed driven long path to open source if it benefits my kids and grandkids. At some point in the future, they will be making more and more complex compounds and the wealth of knowledge will expand exponentially. The benefit of the AI will not cash out immediately for me, but it will cash out at some point, right? In fact, our infatuation with obtaining the best possible health will not only cash out locally but globally. Or am I missing the point?

    1. Seems like the perspective is from the raw material manufacturers and the pharmaceutical operations label claims for the entities. Raw material manufacturers might have patentable routes to make the precursors or raw materials and have proprietary rights and claims for to sell to the pharmaceutical companies who might have market exclusivity rights for the label claim if not the same owners. So isomers or label claims might be an issue on the pharma end and synthetic routes might be an issue in the raw material or precursor side of the business. This systems looks like more the synthetic operations for precursors and raw materials.

      Like I like to think of from a C.I.S. degree perspective. Say I have a computer information system (CIS) that is the hardware and software with an operating system that meets my requirements. That system can become corrupted with maliceware, spyware, diseaseware and other issues that can kill my functionality and capabilities to perform my required CIS role and responsibility. Next thing I know I have a completely different OS that stole my property for a ransom if not completely destroyed the system (used to be hardware destructive methods… or at least for drives when I was at Tech there was in the late 90’s still). Any system can have similar potential malicious issues.

      I was never trained in “financial instruments” being a survival requirement. I was trained in survival requirements being like “medicine” and not “poisonous”.

      Might be a graduate school like scenario with socioeconomic trends that most won’t comprehend and some can sell. Might be value added benefit with a tangible return on investment without market hype for longer term survivability with increase health and safety while not infringing on our Rights and Liberties.

      I’m thinking there is going to be review internally, maybe third party and with government agencies… and systems like this are better suited to make Lean Sigma government so government regulatory agencies aren’t corruptible.

      Seems interesting, where I hope that the AI can be used to keep the best minded focused on rational, logical ways and means to increase public health in safe ways considering a range of diverse cultures that prefer to not harm and are more domesticated with the longest term visions for others also. Sometimes… the best teams need the funding for a period to get their systems to the market at a price point and volume that keeps the infrastructure developed to a point that can allow the system to stay with society long term.

    2. Agreed. No matter how many ways to make the drug they patent, the patent has an expiration date on it, and a bunch of “Little Pharma” outfits tooling up to sell their own version the day it expires. They may be reducing the odds someone will circumvent the patent for its duration, but the trade-off is that after it expires, the world gets more open source data on different ways to make the medicine, potentially driving down the cost.

  5. Seems they can perform a QSAR type approach with the receptor sites of the different classes of biochemicals that affect human physiology… then compare to parent molecules and what would dock. Maybe even with electrophysiological approaches. Then the synthetic methods for drug candidates as well as raw material pharmecodiligence recon. I’m not an AI algorithm expert at all. I need to read into more.

  6. There is a bright side of such method. When many possible ways to produce some drug will be patented, someone who wants to synthesize it and does not bother with patents would choose simplest method available with his technology instead of trying to replicate the one described in the patent :)

    1. The likely issue however, is that companies will attempt to patent ALL methods found to arrive at a solution, and no other companies will be allowed to bring a solution to commercial production as it will be fined out of existence for patent violation.

      We will need to start making our own drugs from the patent applications, as it will basically spell out every possible method to an endpoint.

  7. Point of contention : a ‘Patent Troll’ is someone that owns a patent on something not to protect its monopoly on production of that thing, but to only sue other that would try to make it. Patent Trolls are easy to hate, because they do not operate in ‘good faith’ of the patent system an only serve to stifle innovation.

    If you are hating on companies protecting their IP through the Patent system, try to abolish the Patent system. As a wise man once said “don’t hate the player, hate the game.”

      1. An even wiser person realizes human nature will not change (just how many billions of years does one need) and designs accordingly. Pretending people are something they aren’t is a recipe for failure.

        1. Yes, that is why the patent system was invented to allow those who invest in the creation of new things to be able to have exclusive rights for a time to repay the R&D costs. (Prevent scummy people from undercutting the originator because they don’t have to pay the R&D)

          I think it could use some revisions to block abuses, but the current patent system does more good than harm.

  8. Drugs aren’t the main interest, nor is helping people, it’s about selling a product for maximum profit. New products are more profitable, since previous products weren’t perfect in everyway, one dose cures, not to mention side effects. AI might help find new drugs, or cheaper ways make them, but eventually the AI will be used to find the most profitable drugs, without the costly lawsuits. Good drugs aren’t profitable, limits the market. They need drugs that are only marginally effective, and need to be taken as often as possible, for a lifetime. A good drug, is one that requires several other drugs to be used in conjunction with it, so instead of a single medication, the patient gets to pay for several prescriptions. Grug companies only care about their products not killing too many of there customers, or obviously making them worse, than had they never used the product.

    1. Nice conspiracy theory, Mr. tin-foil hat! The problem isn’t that the companies intentionally strive to do this, the issue is that small organic molecules can just do so much. They might be easily able to block a histamine receptor but they can’t prevent the overly complicated and not understood path towards seasonal allergy to happen.
      No company would produce a drug against the symptoms if they thought one could be made to cure a disease. Why? Because if they just have the symptom blocker but the company next door has the cure, they wasted billions of dollars in R&D and certification. They can still make huge profits from curative drugs (ever heard of antibiotics?) because people keep getting sick!

      1. Seems like most adults have some sort of prescribe medication, blood pressure, cholesterol, quite a few things, which could be taken care of with daily exercise, and healthier diet. But, people would rather take a pill, than take responsibility. When I was a kid, they didn’t call acting up in class, a medical condition, but poor parenting. A little corporal punishment usually cured in one or two doses. When done before the whole class, tended to get the entire class cured as well.

        Pharmaceuticals is a business, not a charity. They want to maximize profits, expand their markets. Sales come first, then risk management. If the cared about the end user of their drugs, do you think it’s okay to medicate a 6 year old kid, who’s a little energetic, or has some other behavior problem? Wonder what happens after one of these kids, grows up, and decides he/she doesn’t need the meds anymore, Never learned to control themselves, since the pills took care of the problem.

        It’s not conspiracy, just plain business.

        1. There is criminal conspiracy… so technically a for profit business conspiracy. Potentially in warfare ways and means called chemical warfare. Some just forget or deny the war on drugs isn’t paramilitants and military SWAT trained more than ever ways and means. All with extra jurisdictional out of bounds with international take overs of assets. Who owns what now or majority stakes?

  9. Don’t drug patents only last about 20 years? That isn’t actually a very long time. If there was no money to be made, few drugs would ever get developed. Most useful drugs have been developed in capitalist countries.

    1. Exactly. It varies from country to country, but they all keep pretty close to the same to maintain reciprocity patent honoring deals. I want to say the US is 22 years.

      If they really patent the drug and all similar developments on day 1, they’re limiting the risk of short term loss on investment while actually seriously harming the current long term profit model, which is to patent a slightly more effective variant as expiration approaches.

      Drug patents are an easy straw man to blame for the rising cost of medicine, but the evidence doesn’t support that. I knew someone who took a medication that costs about 400 USD a month cash price. Co-pay is 30 USD. With some insight into their books, the pharmacy appears to be getting about 60 from the insurance company. So all told, the price is less than a quarter of the cash price. The problem is the insurance companies and the government programs expect discounts, so the base prices go up to allow room for massive negotiated discounts.

      And some of the favorites for pointing out expense are truly insanely expensive to make. Banning medical insurance would force prices to drop FAST, but you’d have people suffer, and more than a few die during the transition. And the transition itself would raise unemployment as varies employment inefficiencies shook out.

    2. Not only does the patent have a limited time horizon, but the patents have to be filed relatively early in the research process, since you have to file for a patent before publishing anything about what you’re patenting. You may end up with only 10 or 15 years left on that 20 year patent by the time the drug is approved.

  10. This is why the patent system is hopelessly broken. And why I am working on interstellar travel to provide a way for like minded people to go elsewhere when the time is right.

  11. It seems we’re headed toward a system where more and more and eventually almost everything will be owned by a few wealthy individuals or mega-corporations and the average person owns nothing and is also fairly useless. All process become more complex over time, making it harder and harder for small companies or individuals to keep up. I wonder what the societal implications of that are.

  12. “Pharmaceuticals, though, are treated differently. Since it’s relatively easy to reverse engineer a chemical compound using analytical chemistry tools and methods, patents for drugs concentrate on the proces…“

    Not really, it is also possible to patent the compound, composition and their use.
    Just because someone may reverse engineer it it does not mean that the patent rights will be lost.

  13. Fun fact: The NIH (National Institute of Health) funds the majority of pharmaceutical research here in the US.
    The companies develop these drugs, and then rake in enormous profits giving little to nothing back to the government or the consumer who must ultimately pay these costs!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.