Agree with your conclusion! Neuroimaging/neuroscience is another field in which there are lots of physicists and engineers but not much acceleration of progress.
I think another thing that we don’t take into account is that we have so many (necessary) ethical things in place to protect human trial participants. If we had fewer restrictions, we could probably fly through far more options, but at a greater human cost. Hence why we have so many things that work on mice: as much as I hate to say it, we don’t really care what the impact on them is outside of what it means to us.
It also takes a huge amount of time for a drug to be approved, in part because of large wait times on applications.
While I would love for it to be as simple as ‘do more maths’, anyone anywhere near medical research knows it really isn’t that simple 😅
And many "ethical" things that are totally unnecessary. Privacy is the obvious one, but also - did you know that there seems to be a regulation that you can't sign forms in pencil, seemingly also for ethical reasons? https://slatestarcodex.com/2017/08/29/my-irb-nightmare/
Thanks for the post, I especially appreciate the survey of progress.
“Eric Lander… the first scientist to occupy this position.”
Is this supposed to say biologist?
I find Peter Thiel’s suggestion that biologists are just failed physicists insulting (and a bit sexist). I say this as a physicist! I actually moved from experiment to theory because I could not deal with the slow feedback loop, and in my field, that’s 1 month-1 year not multiple years! Nothing but respect for you all who can do experiments far more complicated and time-sensitive than anything I did, especially because the number of variables you have to consider is so much larger.
There’s a lot of variables in DNA mutations. The advances we’ve made on targeted tomotherapy, immunotherapy, as well as vaccines that prevent infection with the viral precursor to cancer have drastically reduced our chances of getting cancer or dying from it, or the treatment. So we’re miles away from the early 70s neoplastic interventions. Genetic therapy will probably eliminate some familial causes of cancers, once we’ve teased out clear relationships. So there’s a lot to be excited about and grateful for. Shit, we can cure AIDS with stem cell therapy. I think we’ve just gotten less susceptible to excitement and impression about truly monumental advancements in science in the age of such rapid change.
I think it's fair to disagree with the entire premise that we are too slow... but I think things could always be better so it's worth asking what prevents them from being even better
This is an interesting piece. My own hot take (as a biologist) is that doing well in biology is less about raw cognitive power and more about knowing and remembering lots of stuff. This attitude that if you're good at math, you are just generally smart and good at everything holds decently well in physics, comp sci, and physical chemistry, but in organic chemistry and biology I've noticed it's less true.
As for the slow pace in genomics, we've run into two the problems. First, individual alleles are generally less predictive than we all thought in the 1990s. BRCA mutations matter of course, but there's no "gay gene" or "depression gene" or whatever. It's way more complex than that. Second, understanding genetic and epigenetic regulation is not the same as being able to manipulate the genome/epigenome, although with CRISPR progress is actually starting to happen. We've understood the biology of sickle cell anemia for decades: we knew the mutation that causes it, we knew expression of fetal haemoglobin ameliorates the disease, we knew the regulatory elements that affect fetal haemoglobin expression etc. etc. Now with CRISPR we finally have curative gene therapy, including both treatments that repair the mutation and treatments that reactive fetal haemoglobin. So on the genetic/epigenetic front, we are finally starting to do more than just publish papers about systems we are helpless to manipulate.
Really interesting piece. Couple of random thoughts:
1) As far as biology in general, the more we learn the more we see we don't know.
2) Sequencing the genome was a great achievement, and immediately then generated questions as to how does that work when folded up inside a cell. We're still trying to figure that out, both in terms of normal processes but also how it goes off the rails in pathological conditions. The newer multiplex methods you mention, especially ones that allow single cell analysis, including mathematically complex analysis, look promising, especially ones that can image the central dogma.
3) Peter Theil obviously has no idea what is required to experimentally collect and analyze biological data.
4) I'm a little skeptical about your idea that the IRA will lead to less drug discovery research getting into trials phase.
As far as drug prices themselves are concerned, more could be gained by at least more closely regulating, if not just rid of, pharmacy benefit management companies.
Thanks for your well informed essay. Indeed biology is complex. One issue you understandably avoid is the role of hype, which leads the public to believe that the cure for cancer is just around the corner. I’m not a cancer biologist, but (spoiler alert): it isn’t.
What do you think of the claim Sydney Brenner made in 2010 in "Sequences and consequences" (https://royalsocietypublishing.org/doi/10.1098/rstb.2009.0221) that the gene is too low of a level of abstraction, and that the correct level of abstraction is the cell? He argued that this would allow for a "middle-out" approach to biology ("from the vantage point of the cell we can look down on the molecules that constitute it and look up at the organism that contains it"). Moreover, this would be amenable to both data-driven and theory-driven methodologies since the problem of converting data to knowledge would presumably be a lot less daunting than it is with all the various types of -omics out there.
It's I think well reiterated in the book The Master Builder that I have yet to read (but read many blog posts by its author). I think it's plausible, but I also think we do not have as many experimental techniques for treating cells as the core thing.
“For example, it showed that humans have only about 30,000-35,000 genes, two times less than a fly” is the fly here drosophila? The numbers here are a bit iffy, sorry to pick nits but this is an article about math/bio divide…
Regarding feedback loops: is there anything that can be done to speed these up? I’ve seen claims before about drugs being designed and tested virtually, but this sounds like a sci-fi pipe dream. But if drugs could be designed and tested virtually I’d imagine that feedback loops could be shortened.
Well-done piece. It conveys a feeling that biological paradigms are insufficient to model the “wet complexity” irrespective of source inside or outside biology.
I enjoyed your recap of the history of genomics. It took me back to when the Drosophila genome was sequenced at the beginning of my doctoral project.
I’m a former cancer biologist who used to work in a cancer hospital and shadowed oncologists in their rounds with survivors of childhood cancer. Trying to solve a not only biological problem with biological solutions is showing its limitations. What if biology is a readout of underlying conditions, not the source? We can’t interrogate the sociological, emotional, spiritual, environmental, and psychological contributions to tumour formation and metastasis using current methods and thinking. I agree we need a multidisciplinary approach but one that extends beyond the hard sciences.
I’m a not even close to being a rank amateur in this discussion, having reached my humble mathematical peak in high school. But your comments caught my eye because my mother died of something resembling Non-Hodgkins Lymphoma at 43 - my dad said at the time that it was ‘reticulum cell sarcoma’ though my understanding is that this is now considered an anachronism. Coincidentally she had been born with one malformed, under-developed kidney, which was surgically removed in her mid-20s. She had previously manifested chronic inflammation - told my dad she had reliably been able to call on the school nurse to send her home for running a fever whenever she wished - and the removal cured that. I have since read that chronic inflammation is a major risk factor for cancer, but also that ‘cancer’ is such a broad umbrella term that it barely suffices to accurately identify a range of metastatic disease that is quite broad. I am searching for more insight into her illness and any guidance would be much appreciated. My father, a surgeon, held a hunch that exposure to radiation may have been a catalyst, specifically possible proximity to atomic weapons-related material. (During a year-long stay on a South Pacific island military base she had held a high security clearance as an executive secretary and her desk was next to a mysterious vault she was occasionally tasked with opening.) So, through all the angst accompanying her death, identifying the point of origin has, at times, been a bit of a family obsession. The ‘radiation theory’ seems a stretch, albeit not an utterly implausible one, given that she may well have been vulnerable anyway.
>Given that 90% of therapies that advance into human trials fail, information from these human trials is particularly important.
While this shows the importance of trials, it doesn't necessarily show that they represent the bottleneck (not that you're claiming it does), rather than e.g. the intelligence of biologists. Imagine if the rate of therapies that failed human trials were 99.9%. While that would indeed highlight the importance of trials, and the value to be gained in speeding them up, it would highlight even more the poor determination a priori of which therapies hold promise.
That said, if as you argue, there are hard limits on understanding wet messy therapies a priori, then the speed of trials may well be the current bottleneck.
>On the other hand, the slowing of feedback loops in medicine will have largely been the result of policy decisions.
You referenced a number of policy decisions that slow drug development, but do we have any rough estimates for their impact on the slowing of feedback loops? The IRA passed less than two years ago, so shouldn't have caused that much damage, yet. HIPAA was passed in the 90s, while IRBs have been in force for decades before that. When did the feedback loops start slowing, and how does that timing relate to specific policy decisions?
Particularly if one accepts the thesis that biology is messy and wet and doesn't lend itself to rapid progress through ingenuity alone, one might suspect that as low hanging fruit are picked, further progress will slow.
Apparently (https://en.wikipedia.org/wiki/Eroom's_law, https://www.nature.com/articles/nrd3681) R&D efficiency has been decreasing, since at least 1950. Ostensibly that suggests that the particular highlighted policy decisions leave much of the story unexplained. While it's posssible that the trend in general is mostly the result of an increasing regulatory burden, I'm curious to what extent low hanging fruit have simply been picked.
> You referenced a number of policy decisions that slow drug development, but do we have any rough estimates for their impact on the slowing of feedback loops? The IRA passed less than two years ago, so shouldn't have caused that much damage, yet. HIPAA was passed in the 90s, while IRBs have been in force for decades before that. When did the feedback loops start slowing, and how does that timing relate to specific policy decisions?
Thanks! I said the culprits are scattered across the medical world and numerous precisely because I do not think we have a good estimate of how much each of them matters and so on. IRA is more "trends for the future" and I referenced that econ paper. If revenue decreases and clinical trial prices increase, I think it's reasonable to infer there will be less trials
Speed of trials or rather the absolute sluggishness of the FDA, as well as cherry picking then”healthiest” dying patients, limiting the trial pool, are huge problems. Average time to approval of even a fast track drug is 10 years. On a health timescale that’s glacial. I write about it here if you’re interested. Attempts at policy change is so important. https://bessstillman.substack.com/p/the-drugs-killing-dying-patients?r=16l8ek
Of course! Even if the policies that Ruxandra listed in particular, and the collective regulatory burden as a whole, are responsible for only a minority of the slowing of drug development, that's still a violation of basic human rights and the cause of absolutely horrific human suffering at a large scale. I hope you're successful in bringing attention to this issue and wish the best to you and your husband.
Agree with your conclusion! Neuroimaging/neuroscience is another field in which there are lots of physicists and engineers but not much acceleration of progress.
oh that's interesting to know.
I think another thing that we don’t take into account is that we have so many (necessary) ethical things in place to protect human trial participants. If we had fewer restrictions, we could probably fly through far more options, but at a greater human cost. Hence why we have so many things that work on mice: as much as I hate to say it, we don’t really care what the impact on them is outside of what it means to us.
It also takes a huge amount of time for a drug to be approved, in part because of large wait times on applications.
While I would love for it to be as simple as ‘do more maths’, anyone anywhere near medical research knows it really isn’t that simple 😅
And many "ethical" things that are totally unnecessary. Privacy is the obvious one, but also - did you know that there seems to be a regulation that you can't sign forms in pencil, seemingly also for ethical reasons? https://slatestarcodex.com/2017/08/29/my-irb-nightmare/
Great article, Ruxandra. We really need more biologists writing op-eds, and you're clearly becoming a master of the art form.
thank you! great praise coming from you :)
Thanks for the post, I especially appreciate the survey of progress.
“Eric Lander… the first scientist to occupy this position.”
Is this supposed to say biologist?
I find Peter Thiel’s suggestion that biologists are just failed physicists insulting (and a bit sexist). I say this as a physicist! I actually moved from experiment to theory because I could not deal with the slow feedback loop, and in my field, that’s 1 month-1 year not multiple years! Nothing but respect for you all who can do experiments far more complicated and time-sensitive than anything I did, especially because the number of variables you have to consider is so much larger.
Thanks! Sorry I deleted. He was the first scientist to occupy after it had been made a cabinet position.
As in he was the first scientist to be a member of the cabinet
Yes, let's get the brilliant mathematical physicists who have made such a laughable mess of physics to save medicine.
There’s a lot of variables in DNA mutations. The advances we’ve made on targeted tomotherapy, immunotherapy, as well as vaccines that prevent infection with the viral precursor to cancer have drastically reduced our chances of getting cancer or dying from it, or the treatment. So we’re miles away from the early 70s neoplastic interventions. Genetic therapy will probably eliminate some familial causes of cancers, once we’ve teased out clear relationships. So there’s a lot to be excited about and grateful for. Shit, we can cure AIDS with stem cell therapy. I think we’ve just gotten less susceptible to excitement and impression about truly monumental advancements in science in the age of such rapid change.
I think it's fair to disagree with the entire premise that we are too slow... but I think things could always be better so it's worth asking what prevents them from being even better
This is an interesting piece. My own hot take (as a biologist) is that doing well in biology is less about raw cognitive power and more about knowing and remembering lots of stuff. This attitude that if you're good at math, you are just generally smart and good at everything holds decently well in physics, comp sci, and physical chemistry, but in organic chemistry and biology I've noticed it's less true.
As for the slow pace in genomics, we've run into two the problems. First, individual alleles are generally less predictive than we all thought in the 1990s. BRCA mutations matter of course, but there's no "gay gene" or "depression gene" or whatever. It's way more complex than that. Second, understanding genetic and epigenetic regulation is not the same as being able to manipulate the genome/epigenome, although with CRISPR progress is actually starting to happen. We've understood the biology of sickle cell anemia for decades: we knew the mutation that causes it, we knew expression of fetal haemoglobin ameliorates the disease, we knew the regulatory elements that affect fetal haemoglobin expression etc. etc. Now with CRISPR we finally have curative gene therapy, including both treatments that repair the mutation and treatments that reactive fetal haemoglobin. So on the genetic/epigenetic front, we are finally starting to do more than just publish papers about systems we are helpless to manipulate.
Broadly agree
Really interesting piece. Couple of random thoughts:
1) As far as biology in general, the more we learn the more we see we don't know.
2) Sequencing the genome was a great achievement, and immediately then generated questions as to how does that work when folded up inside a cell. We're still trying to figure that out, both in terms of normal processes but also how it goes off the rails in pathological conditions. The newer multiplex methods you mention, especially ones that allow single cell analysis, including mathematically complex analysis, look promising, especially ones that can image the central dogma.
3) Peter Theil obviously has no idea what is required to experimentally collect and analyze biological data.
4) I'm a little skeptical about your idea that the IRA will lead to less drug discovery research getting into trials phase.
As far as drug prices themselves are concerned, more could be gained by at least more closely regulating, if not just rid of, pharmacy benefit management companies.
4) this is speculative. I mean the idea is that companies would have less money so less money to invest in R&D, too
Thanks for your well informed essay. Indeed biology is complex. One issue you understandably avoid is the role of hype, which leads the public to believe that the cure for cancer is just around the corner. I’m not a cancer biologist, but (spoiler alert): it isn’t.
yeah true!
I disagree, many cancer "cures" have helped a lot of people but are then run out of the US. Get a flight to Mexico.
Methylation cycle function and quantum energy flow within the cytoplasm needs to be restored and cancer will reverse itself.
What do you think of the claim Sydney Brenner made in 2010 in "Sequences and consequences" (https://royalsocietypublishing.org/doi/10.1098/rstb.2009.0221) that the gene is too low of a level of abstraction, and that the correct level of abstraction is the cell? He argued that this would allow for a "middle-out" approach to biology ("from the vantage point of the cell we can look down on the molecules that constitute it and look up at the organism that contains it"). Moreover, this would be amenable to both data-driven and theory-driven methodologies since the problem of converting data to knowledge would presumably be a lot less daunting than it is with all the various types of -omics out there.
yeah I know this argument.
It's I think well reiterated in the book The Master Builder that I have yet to read (but read many blog posts by its author). I think it's plausible, but I also think we do not have as many experimental techniques for treating cells as the core thing.
like the reason we treated "sequences" as the main thing seems to me that we had access to many of them hah
I hadn’t heard of that book before, will check it out!
Arias is imo the best critic of the current omics obsession (I say this as an omics PhD :) )
This is a very convincing argument. Your writing does a great job at striking a balance between general appeal and deep dives into your expertise
Thank you!
“For example, it showed that humans have only about 30,000-35,000 genes, two times less than a fly” is the fly here drosophila? The numbers here are a bit iffy, sorry to pick nits but this is an article about math/bio divide…
this is very embarassing, I wrote this from memory from my undergrad days (revised now)
It happens! Great post, beautifully written
Regarding feedback loops: is there anything that can be done to speed these up? I’ve seen claims before about drugs being designed and tested virtually, but this sounds like a sci-fi pipe dream. But if drugs could be designed and tested virtually I’d imagine that feedback loops could be shortened.
There's a case to be made that safetyism slows down clinical trials
Well-done piece. It conveys a feeling that biological paradigms are insufficient to model the “wet complexity” irrespective of source inside or outside biology.
I enjoyed your recap of the history of genomics. It took me back to when the Drosophila genome was sequenced at the beginning of my doctoral project.
I’m a former cancer biologist who used to work in a cancer hospital and shadowed oncologists in their rounds with survivors of childhood cancer. Trying to solve a not only biological problem with biological solutions is showing its limitations. What if biology is a readout of underlying conditions, not the source? We can’t interrogate the sociological, emotional, spiritual, environmental, and psychological contributions to tumour formation and metastasis using current methods and thinking. I agree we need a multidisciplinary approach but one that extends beyond the hard sciences.
I’m a not even close to being a rank amateur in this discussion, having reached my humble mathematical peak in high school. But your comments caught my eye because my mother died of something resembling Non-Hodgkins Lymphoma at 43 - my dad said at the time that it was ‘reticulum cell sarcoma’ though my understanding is that this is now considered an anachronism. Coincidentally she had been born with one malformed, under-developed kidney, which was surgically removed in her mid-20s. She had previously manifested chronic inflammation - told my dad she had reliably been able to call on the school nurse to send her home for running a fever whenever she wished - and the removal cured that. I have since read that chronic inflammation is a major risk factor for cancer, but also that ‘cancer’ is such a broad umbrella term that it barely suffices to accurately identify a range of metastatic disease that is quite broad. I am searching for more insight into her illness and any guidance would be much appreciated. My father, a surgeon, held a hunch that exposure to radiation may have been a catalyst, specifically possible proximity to atomic weapons-related material. (During a year-long stay on a South Pacific island military base she had held a high security clearance as an executive secretary and her desk was next to a mysterious vault she was occasionally tasked with opening.) So, through all the angst accompanying her death, identifying the point of origin has, at times, been a bit of a family obsession. The ‘radiation theory’ seems a stretch, albeit not an utterly implausible one, given that she may well have been vulnerable anyway.
Thanks for this interesting and compelling piece!
>Given that 90% of therapies that advance into human trials fail, information from these human trials is particularly important.
While this shows the importance of trials, it doesn't necessarily show that they represent the bottleneck (not that you're claiming it does), rather than e.g. the intelligence of biologists. Imagine if the rate of therapies that failed human trials were 99.9%. While that would indeed highlight the importance of trials, and the value to be gained in speeding them up, it would highlight even more the poor determination a priori of which therapies hold promise.
That said, if as you argue, there are hard limits on understanding wet messy therapies a priori, then the speed of trials may well be the current bottleneck.
>On the other hand, the slowing of feedback loops in medicine will have largely been the result of policy decisions.
You referenced a number of policy decisions that slow drug development, but do we have any rough estimates for their impact on the slowing of feedback loops? The IRA passed less than two years ago, so shouldn't have caused that much damage, yet. HIPAA was passed in the 90s, while IRBs have been in force for decades before that. When did the feedback loops start slowing, and how does that timing relate to specific policy decisions?
Particularly if one accepts the thesis that biology is messy and wet and doesn't lend itself to rapid progress through ingenuity alone, one might suspect that as low hanging fruit are picked, further progress will slow.
Apparently (https://en.wikipedia.org/wiki/Eroom's_law, https://www.nature.com/articles/nrd3681) R&D efficiency has been decreasing, since at least 1950. Ostensibly that suggests that the particular highlighted policy decisions leave much of the story unexplained. While it's posssible that the trend in general is mostly the result of an increasing regulatory burden, I'm curious to what extent low hanging fruit have simply been picked.
Incidentally, it's "HIPAA," not "HIPPA."
Speaking of IRB bureaucracy (outside of a biology setting), Scott Alexander wrote an entertaining piece on his attempt to navigate it: https://slatestarcodex.com/2017/08/29/my-irb-nightmare/.
> You referenced a number of policy decisions that slow drug development, but do we have any rough estimates for their impact on the slowing of feedback loops? The IRA passed less than two years ago, so shouldn't have caused that much damage, yet. HIPAA was passed in the 90s, while IRBs have been in force for decades before that. When did the feedback loops start slowing, and how does that timing relate to specific policy decisions?
Thanks! I said the culprits are scattered across the medical world and numerous precisely because I do not think we have a good estimate of how much each of them matters and so on. IRA is more "trends for the future" and I referenced that econ paper. If revenue decreases and clinical trial prices increase, I think it's reasonable to infer there will be less trials
Speed of trials or rather the absolute sluggishness of the FDA, as well as cherry picking then”healthiest” dying patients, limiting the trial pool, are huge problems. Average time to approval of even a fast track drug is 10 years. On a health timescale that’s glacial. I write about it here if you’re interested. Attempts at policy change is so important. https://bessstillman.substack.com/p/the-drugs-killing-dying-patients?r=16l8ek
Of course! Even if the policies that Ruxandra listed in particular, and the collective regulatory burden as a whole, are responsible for only a minority of the slowing of drug development, that's still a violation of basic human rights and the cause of absolutely horrific human suffering at a large scale. I hope you're successful in bringing attention to this issue and wish the best to you and your husband.