3 ms·

Is there any good/better way to derisk this process? It's always seemed quite inefficient to me

giarc 5y ago
Read up on Andrew Lo. He used his background in finance to develop better biotech financing models. https://www.bloomberg.com/news/articles/2021-02-05/out-of-grief-mit-s-andrew-lo-invented-a-better-way-to-finance-biomedical-innovation https://www.bloomberg.com/news/articles/2021-02-05/out-of-gr...
cshimmin 5y ago
It seems that at some level, that is exactly the purpose of (publicly-funded) basic research on the academic side. For example, the idea of messenger RNA therapy has been around for decades in the academic literature, and of course this idea was only possible with preceding advancements in general biology.
osigurdson 5y ago
https://www.schrodinger.com/ https://www.schrodinger.com/
- theonething 5y ago
  totally off topic, but can't resist. on their careers page, under Software Engineers there is this listing: > Developer, Force Fields haha, yes, the scifi nerd dream job. :)
  - Ajef 5y ago
    I assume it's referring to this: https://en.wikipedia.org/wiki/Force_field_(chemistry) https://en.wikipedia.org/wiki/Force_field_(chemistry) And developing software similar to: https://openforcefield.org/software/ https://openforcefield.org/software/
csours 5y ago
Protein folding helps some. Genomics help some. Computer models help with pharmacokinetics. But all these things help you get to the expensive part more easily: testing safety and efficacy. https://www.science.org/content/blog-post/big-problems https://www.science.org/content/blog-post/big-problems https://www.science.org/content/blog-post/ai-and-drug-discovery-attacking-right-problems https://www.science.org/content/blog-post/ai-and-drug-discov... https://www.science.org/content/blog-post/another-ai-generated-drug https://www.science.org/content/blog-post/another-ai-generat...
- ethbr0 5y ago
  Or to put it in other terms, "better" ways mean billions of dollars of avoided loss / additional profit to pharma companies doing drug development. Suffice to say, there are a lot of people always looking at this, and a lot of money being spent on it. It's just hard.
cehrlich 5y ago
Currently, no. Bringing a new drug to market takes several years and a billion dollars. And I haven't seen any compelling evidence that anyone is working on a "solution". A large part of the cost and complexity is the clinical trials, of which several stages of increasing complexity are required to get FDA approval. While it might seem inefficient, it's also the reason why vaccines work so well and have such small side effects - about half of Phase 3 candidates either fail the clinical trial or are rejected by the government for some other reason. Each round of clinical trials requires significantly more money than the one before it, but it's also arguably easier to get investors on board in the later rounds as risk is lower. Smaller pharma companies often aim to get bought outright before their product is brought to market because it makes more sense for everyone involved. Maybe someone will come up with an algorithm that lets us simulate clinical trials with sufficient precision, but that seems like a far-off sci fi fantasy to me.

Read up on Andrew Lo. He used his background in finance to develop better biotech financing models. https://www.bloomberg.com/news/articles/2021-02-05/out-of-grief-mit-s-andrew-lo-invented-a-better-way-to-finance-biomedical-innovation https://www.bloomberg.com/news/articles/2021-02-05/out-of-gr...

cshimmin 5y ago

It seems that at some level, that is exactly the purpose of (publicly-funded) basic research on the academic side. For example, the idea of messenger RNA therapy has been around for decades in the academic literature, and of course this idea was only possible with preceding advancements in general biology.

osigurdson 5y ago

https://www.schrodinger.com/ https://www.schrodinger.com/

theonething 5y ago
totally off topic, but can't resist. on their careers page, under Software Engineers there is this listing: > Developer, Force Fields haha, yes, the scifi nerd dream job. :)
- Ajef 5y ago
  I assume it's referring to this: https://en.wikipedia.org/wiki/Force_field_(chemistry) https://en.wikipedia.org/wiki/Force_field_(chemistry) And developing software similar to: https://openforcefield.org/software/ https://openforcefield.org/software/

csours 5y ago

Protein folding helps some. Genomics help some. Computer models help with pharmacokinetics. But all these things help you get to the expensive part more easily: testing safety and efficacy. https://www.science.org/content/blog-post/big-problems https://www.science.org/content/blog-post/big-problems https://www.science.org/content/blog-post/ai-and-drug-discovery-attacking-right-problems https://www.science.org/content/blog-post/ai-and-drug-discov... https://www.science.org/content/blog-post/another-ai-generated-drug https://www.science.org/content/blog-post/another-ai-generat...

ethbr0 5y ago
Or to put it in other terms, "better" ways mean billions of dollars of avoided loss / additional profit to pharma companies doing drug development. Suffice to say, there are a lot of people always looking at this, and a lot of money being spent on it. It's just hard.

cehrlich 5y ago

Currently, no. Bringing a new drug to market takes several years and a billion dollars. And I haven't seen any compelling evidence that anyone is working on a "solution". A large part of the cost and complexity is the clinical trials, of which several stages of increasing complexity are required to get FDA approval. While it might seem inefficient, it's also the reason why vaccines work so well and have such small side effects - about half of Phase 3 candidates either fail the clinical trial or are rejected by the government for some other reason. Each round of clinical trials requires significantly more money than the one before it, but it's also arguably easier to get investors on board in the later rounds as risk is lower. Smaller pharma companies often aim to get bought outright before their product is brought to market because it makes more sense for everyone involved. Maybe someone will come up with an algorithm that lets us simulate clinical trials with sufficient precision, but that seems like a far-off sci fi fantasy to me.