Hacker News Reader

Comment by JumpCrisscross | Hacker News Reader

When drug companies issue patents, do they not also publicise production methods?

It depends on the scope of the patent. But to some extent, kind of. See, e.g., the patent for an adalimumab (Humira) formulation https://patents.google.com/patent/US9085619B2/en. The patent discloses materials and methods for creating a formulation of the drug.

The real issue is probably that there are so many moving parts, especially when you get to the more complex drugs (e.g. antibodies). You would have to go through a million patents to find all the information required to create the actual molecule and also get it in a state to put into your body.

And even with all the information, it is unlikely anyone is sufficiently resourced (whether it's money, tools, or know-how) to actually make something viable.

as vonmoltke suggests, even if you've got some chemical synthesis procedure laid out in a peer reviewed journal, you are, in the best case, expert man hours away from successfully duplicating it. and you'll only know you've successfully duplicated it if you can run the result through really expensive analytical equipment to confirm that 1) you made the stuff you wanted to make and 2) it doesn't have impurities that are problematic.

(as a minor defense of the pharma companies, you're paying them not just to do a bunch of organic synthesis for you, but also to do ongoing QA that the synthesis process is working correctly. and not all processes are equally easy/amenable to industrial-scale duplication.)

Obviously this is true. There’s much that goes into industrial synthesis that someone with a setup like his at home just can’t currently reproduce, and safely.

What I find most dubious about the article is the lack of attention to synthetic biology. Because QA in that sense can become say, “this modified organism can produce the drug in question under most conditions and all future populations will carry the new genes with X% assurance” - ONCE -and you’re done.

This is quite possibly the real Pharma killer, IMO.

But only kind of. Even if you could get a strain of yeast to produce a specific antibody 100% of the time, how reliably could you get the yeast to grow? If you can get it to grow, how reliably can you extract a enough of the antibody to form just one dose? If you can extract, can you get it into a formulation that is pharmaceutically viable? Its not just about synthesis.

There's so much tooling and expertise that goes into front end, back end, deployment, db, QA, etc ,etc, etc just to get one app going. Why would would drug deployment be any different?

Well, yes you’re correct, there are associated complexities I’ve glossed over. My point however is simply that all of these complexities only really need to be solved once by an expert or team of experts such that non-experts can then utilize such a biotechnology (“yeast strain”) to produce some medicinal compound or other useful engineered metabolite. A yeast strain that can’t be reliably cultivated by a non-expert with relative ease and safety (in a non-laboratory) wouldn’t make the cut, in theory.

Also the same semi-synthetic organism could be applied in traditional pharmaceutical process settings as well, if for no reason other than cost reduction (and probably workforce reduction as well). So even “big pharma” has some incentive to contribute to this sort of technology.

It isn't even the basics of the synthesis chain that are difficult. Most (small molecule) stuff a lab chemist could whip up a batch of given a bit of time and reasonable access to precursors.

But this is exactly like saying "Reddit is just a few SQL queries, a reasonable programmer could whip up their own version in a weekend." Just like there is only one reddit/StackOverflow/Facebook even though plenty of people have the same basic idea, even knowing the chemical formula doesn't help that much. The trick is in the execution: generating reliable quantities of reliable strength and without any contaminants.

There is a whole field, process engineers, devoted to this quite difficult problem.

Generally, no. Even if they did, based on my experience with process-sensitive non-pharmaceutical products it takes a lot of experimentation and refinement to execute the process correctly.

Reply on news.ycombinator.com