3 ms·

I'm a biologist who works in the pharmaceutical industry. I like the idea of open source insulin as much as the next guy. But. I just took a look at their we

I'm a biologist who works in the pharmaceutical industry. I like the idea of open source insulin as much as the next guy. But. I just took a look at their website and it's a complete joke, frankly. It's basically a high-school science experiment. The sort of thing that a single undergrad could do in a few days. Compared to the thousands of man-years required to produce and test a new pharmaceutical product. I wish them luck, but c'mon. Biopharma is not IT where a smart kid can compete from his garage. It just isn't. The NYTimes article says "the hackers hope to be able to demonstrate the technological feasibility" rather than manufacture a drug. But everyone knows that it's technically feasible -- the challenge (and cost) is in the doing.

dnautics 11y ago
"Biopharma is not IT where a smart kid can compete from his garage." Well, I am currently producing one early-stage (xenograft) drug candidate literally in my garage, and have made two new molecules based on the structure, also in my garage, about to be submitted to structural characterization (and eventually, in vitro studies and maybe even xenograft)... But yeah, I do have a PhD in chemistry, and spend 10-20h/wk on the project. (also with the caveat that all my studies hit a "brick wall" of can't do in the garage when I have to start worrying about GMP).
- roadnottaken 11y ago
  Don't get me wrong: synthesizing compounds and expressing proteins is not that hard. Those things can be done in a garage. Extensive pharmacological testing in cultured cells and animals is much, much harder. I respect what you're doing, but you obviously know what GMP is and have some idea of what drug research entails, so to claim that a garage hackers could produce a drug candidate is disingenuous. At our company, a drug candidate is something that has shown efficacy in animal pharmacology models and has gone through 13-week tolerability studies in animals appears safe. And that's all before GMP.
  - dnautics 11y ago
    I don't think there are specific guidelines as to when you start calling something a candidate or not, although I personally wouldn't call a "hit" from a screen a "candidate". Everything I am working with is a natural product derivate, In fact, engineered to reduce tox of a compound that started off in late-stage preclinicals, so the odds of success are (qualitatively) higher to begin with. >Extensive pharmacological testing in cultured cells and animals is much, much harder. And that's why I'm contracting CROs for those parts. I guess technically that's "taking out of the garage", but I never said the whole thing is being done in the garage!
    - roadnottaken 11y ago
      >I don't think there are specific guidelines as to when you start calling something a candidate True, that's why I said "at our company". It differs everywhere, as you mention. My only point was that we screen thousands of compounds, and it's only after a compound has shown desirable activity, safety, ADME, PKPD, etc, that it becomes a real "candidate". Not everyone realizes this, and thinks you can synthesize a compound and the next step is to file an IND. Your point about CROs is a good one, and I thought of it after my post. We use lots of CROs, too. And there are obviously lots of successful virtual companies that don't do any wet-work at all. But CROs cost lots of money. Not really cheaper than doing it yourself, just more convenient sometimes. Virtual companies can work because they have millions in VC. Anyway good luck with your endeavor. I don't mean to sound cynical or negative, but not everyone in IT realizes that the two fields are qualitatively different in many ways and I think the differences are worth pointing out since they're often overlooked around here...
      - dnautics 11y ago
        CROs are really expensive, but it's surprising what is and what isn't. An NMR goes for ~$160 (but usually because you have unremovable features like GLP, which you may or may not care about depending on the step). So for some of these things I'm just "calling over a friend who is a professor at X university" and getting a bit of spectrometer time and a grad student. OTOH, I'm getting xenograft for just about $30k, which I think is about right. > not everyone in IT realizes that the two fields are qualitatively different in many ways and I think the differences are worth pointing out since they're often overlooked around here... Totally agree.
snewk 11y ago
> Biopharma is not IT where a smart kid can compete from his garage. It just isn't. It isn't today, but it may be at some point. We tend to underestimate the effects of monumental shifts like we've had in IT. This project represents the first steps in that direction. > The sort of thing that a single undergrad could do in a few days. Even if this were true, the OpenInsulin project doesn't have a large institution's lab facility. Nor does the project have a government grant as benefactor. They are operating only off crowdfunding. It's obviously not enough to get through the FDA approval process, but IS enough to construct a roadmap to that end and make headway on a functional insulin producing organism. The project already has E. coli making proinsulin. The experimental focus has now moved to methods of purification. They are also investigating the feasibility/cost of getting this through the FDA approval process for biosimilar molecules. I think the real magic of this project is its open nature, which I find to be lacking in the scientific community. This is research that anyone can make meaningful contributions to. Science (specifically biology) can learn a lot from open source software.
- gozur88 11y ago
  >It isn't today, but it may be at some point. We tend to underestimate the effects of monumental shifts like we've had in IT. This project represents the first steps in that direction. But what makes pharma expensive and time-consuming is the regulatory process. Producing drugs is easy - meth dealers do it all the time. But spending a decade getting your drug through the federal system is expensive, and there's no technological shortcut. I'll take more interest when they stop "investigating the feasibility/cost of getting this through the FDA" and publish a plan to actually do it.
- roadnottaken 11y ago
  > construct a roadmap I just don't understand what this means here. Cloning a gene and expressing it in e.coli is something that has been routine for decades. Literally. These simple experiments won't provide a roadmap to do anything that we haven't known how to do for a long time. It's like a crowdsourcing project to raise money to write a "hello world" program. Seriously. Open-ness is great. But "meaningful contributions" are generally made at the cutting edge of scientific research. The experiments described are nipping at the heels of science that was worked-out in the early 80s, so I fail to see how a meaningful contribution will be made here. Open source is great, but making headway in drug discovery or development is a Hard Problem and this doesn't strike me as a serious effort. Just my two cents.

dnautics 11y ago

"Biopharma is not IT where a smart kid can compete from his garage." Well, I am currently producing one early-stage (xenograft) drug candidate literally in my garage, and have made two new molecules based on the structure, also in my garage, about to be submitted to structural characterization (and eventually, in vitro studies and maybe even xenograft)... But yeah, I do have a PhD in chemistry, and spend 10-20h/wk on the project. (also with the caveat that all my studies hit a "brick wall" of can't do in the garage when I have to start worrying about GMP).

roadnottaken 11y ago
Don't get me wrong: synthesizing compounds and expressing proteins is not that hard. Those things can be done in a garage. Extensive pharmacological testing in cultured cells and animals is much, much harder. I respect what you're doing, but you obviously know what GMP is and have some idea of what drug research entails, so to claim that a garage hackers could produce a drug candidate is disingenuous. At our company, a drug candidate is something that has shown efficacy in animal pharmacology models and has gone through 13-week tolerability studies in animals appears safe. And that's all before GMP.
- dnautics 11y ago
  I don't think there are specific guidelines as to when you start calling something a candidate or not, although I personally wouldn't call a "hit" from a screen a "candidate". Everything I am working with is a natural product derivate, In fact, engineered to reduce tox of a compound that started off in late-stage preclinicals, so the odds of success are (qualitatively) higher to begin with. >Extensive pharmacological testing in cultured cells and animals is much, much harder. And that's why I'm contracting CROs for those parts. I guess technically that's "taking out of the garage", but I never said the whole thing is being done in the garage!
  - roadnottaken 11y ago
    >I don't think there are specific guidelines as to when you start calling something a candidate True, that's why I said "at our company". It differs everywhere, as you mention. My only point was that we screen thousands of compounds, and it's only after a compound has shown desirable activity, safety, ADME, PKPD, etc, that it becomes a real "candidate". Not everyone realizes this, and thinks you can synthesize a compound and the next step is to file an IND. Your point about CROs is a good one, and I thought of it after my post. We use lots of CROs, too. And there are obviously lots of successful virtual companies that don't do any wet-work at all. But CROs cost lots of money. Not really cheaper than doing it yourself, just more convenient sometimes. Virtual companies can work because they have millions in VC. Anyway good luck with your endeavor. I don't mean to sound cynical or negative, but not everyone in IT realizes that the two fields are qualitatively different in many ways and I think the differences are worth pointing out since they're often overlooked around here...
    - dnautics 11y ago
      CROs are really expensive, but it's surprising what is and what isn't. An NMR goes for ~$160 (but usually because you have unremovable features like GLP, which you may or may not care about depending on the step). So for some of these things I'm just "calling over a friend who is a professor at X university" and getting a bit of spectrometer time and a grad student. OTOH, I'm getting xenograft for just about $30k, which I think is about right. > not everyone in IT realizes that the two fields are qualitatively different in many ways and I think the differences are worth pointing out since they're often overlooked around here... Totally agree.

snewk 11y ago

> Biopharma is not IT where a smart kid can compete from his garage. It just isn't. It isn't today, but it may be at some point. We tend to underestimate the effects of monumental shifts like we've had in IT. This project represents the first steps in that direction. > The sort of thing that a single undergrad could do in a few days. Even if this were true, the OpenInsulin project doesn't have a large institution's lab facility. Nor does the project have a government grant as benefactor. They are operating only off crowdfunding. It's obviously not enough to get through the FDA approval process, but IS enough to construct a roadmap to that end and make headway on a functional insulin producing organism. The project already has E. coli making proinsulin. The experimental focus has now moved to methods of purification. They are also investigating the feasibility/cost of getting this through the FDA approval process for biosimilar molecules. I think the real magic of this project is its open nature, which I find to be lacking in the scientific community. This is research that anyone can make meaningful contributions to. Science (specifically biology) can learn a lot from open source software.

gozur88 11y ago
>It isn't today, but it may be at some point. We tend to underestimate the effects of monumental shifts like we've had in IT. This project represents the first steps in that direction. But what makes pharma expensive and time-consuming is the regulatory process. Producing drugs is easy - meth dealers do it all the time. But spending a decade getting your drug through the federal system is expensive, and there's no technological shortcut. I'll take more interest when they stop "investigating the feasibility/cost of getting this through the FDA" and publish a plan to actually do it.
roadnottaken 11y ago
> construct a roadmap I just don't understand what this means here. Cloning a gene and expressing it in e.coli is something that has been routine for decades. Literally. These simple experiments won't provide a roadmap to do anything that we haven't known how to do for a long time. It's like a crowdsourcing project to raise money to write a "hello world" program. Seriously. Open-ness is great. But "meaningful contributions" are generally made at the cutting edge of scientific research. The experiments described are nipping at the heels of science that was worked-out in the early 80s, so I fail to see how a meaningful contribution will be made here. Open source is great, but making headway in drug discovery or development is a Hard Problem and this doesn't strike me as a serious effort. Just my two cents.