Hi David, great article, I worked in diagnostics for 20 years, I know it was not the focus of your piece, but I think another reason 23andMe failed was because of their "one and done" business model - once you used their test and got your results, you never have to do it again. Very few businesses survive if their customers never return. So I think they were looking for ways to monetize the data, then ran into the genetic problem you describe. I still commend the founders for the aspiration to make genetic testing available to the average person though.
Dave, thanks so much for this comment. I think you're right about that aspect of the business model. Or, at least, to the extent they were relying on drug development to be the real business model, it didn't leave them much to fall back on, although perhaps they can find some (smaller) foothold with ancestry? More importantly, I appreciate your point about the aspiration. I felt that way too. Mike Joyner, who mentioned above, disagreed with me on that — he felt that people would be more prone to adverse learning based on results they couldn't do anything about, feeling fatalistic, or, alternatively invincible. When I was at Sports Illustrated, I interviewed a bunch of pro athletes in sports like football, boxing, and mma, and asked if they'd get a test like ApoE to see if they might more predisposed to lasting damage from brain trauma. They generally said that it might not change their decisions, but they'd be interested to know. And some suggested that as they got close to retirement, perhaps it would be one factor to put into the decision-making stew. That seems pretty reasonable to me. Additionally, 23andMe and others did a great job at making certain tests affordable. I was able to get a test to see if I was a Tay-Sachs carrier very easily and cheaply. For others I know who have had to do that through the traditional route, it takes forever and is really expensive. That's a spot where 23andMe should already be disruptive, in my opinion. Thanks again for this comment, and I'm glad to be aware of your Substack! Looks fascinating
The psychology around testing is interesting, some people want to know their genetic profile/risks, and others don't. At one of the companies I worked at, we made HIV tests, and I was surprised to learn that many people avoid getting HIV tests because they just don't want to know, even if knowing might save their life. Same for people who avoid getting their colonoscopy, not just because the procedure is gross (although that is a factor), but many people just don't care to know.
A genetic test moves slightly into the realm of asking people the strange philosophical question of whether or not they would like to know the exact day they're going to die. Some would, but many wouldn't. There is something both psychologically and philosophically freeing about NOT knowing certain things. This might be an intriguing area of study - why do people avoid certain kinds of information that could help them? Is it a net pro or con to know you are going to have a degenerative disease later in life? Sometimes there is value in knowing, so that steps might be taken to avoid future harm, but sometimes there is an interesting kind of freedom in intentional ignorance. I do think it's better to have the option to know available, as opposed to withholding it, and let people make their own decision.
Thanks for checking out my Substack, I'm going all in on embracing generalism for both life and "portfolio career". Thanks for writing Range, it's opened up many more people to the value of generalism in the same way that Susan Cain's Quiet opened up the discussion on the value of introverts.
Dave, I think you and I may have similar brains! By the end of your first paragraph I was thinking, "kind of like asking people if they want to know when they'll die," and then there it was. I agree it's better to have the option. I recall that Robert Green did a few studies, again regarding ApoE, looking at different aspects of disclosure. And I think a conclusion was that the option of testing is a positive for a lot of people, and when it's a negative it's often a transient one. Of course, this was among people who opted in to testing, but I don't think anyone is talking about anything other than that anyway. (Here's one example: https://www.nejm.org/doi/full/10.1056/NEJMoa0809578). ...Lastly, in terms of our similar brains: I take the Susan Cain comparison as a huge compliment. As an introvert, I really appreciated that book (and have since gotten to know here a little and she's truly a lovely person), and also think my writing style is probably about as similar to hers as anyone I can think of. The title and subtitle, as you may have gleaned, are sort of patterned after her example.
Hi David, haha nice to meet a "brain twin", maybe there is something to this generalist mindset after all. Oh and I'm also an introvert :)
I heard you on a podcast a while back and enjoyed how you were answering the questions, you were explaining things the same way I like to learn them, with a sense of curiosity, at depth, with thoughtful analysis, plus some amount of fascination in discovering unexpected things. In some sense it was frustrating though, because I felt like you had more information and insights in your pocket than you were given time to convey, but such is the nature of podcast interviews.
I never noticed that your book title/subtitle was similar to "Quiet", but now that you mention it, I see the similarity. I follow Susan on here and she does seem like a genuinely kind person. I enjoyed her book as well.
The ApoE study results are interesting but not necessarily surprising, that people who were negative experience some relief, while those who were positive some anxiety, but maybe not as much lasting difference as one might've expected. I do think this has something to do with the fact that Alzheimers is a) far off into the future, b) something that we generally accept might happen as we age, and c) at the moment we're lacking strong evidence on what you can do to prevent it. A more devastating, worrisome disease diagnosis e.g. HIV or cancer would've obviously produced different results in this type of study.
Loved it! Very well put. By the way, I am looking at a copy of your The Sports Gene :)
I got a PhD in Genetics and Molecular Biology more than 10 years ago and I am still mesmerized by how a molecule as simple as DNA can hold so much information, heredity and evolution power. BuT I´ve never wanted to get my cheeks swabbed. It´s a long, looong way from genotype to phenotype for most cases and that´s probably because we’ve underestimated how different (if not all!) loci interplay to produce the "seeable". Not to mention the mighty environment. It´s a shame that 23andMe killed the drug discovery initiative. I don´t see how we can get there if not by pushing the science.
Prianda! Oh geeze, I try to pretend that nobody who actually knows what they're talking about reads my ramblings here;) In any case, I agree with you on all fronts. I minored in astronomy in college, and was just in awe of the largeness and complexity of the cosmos. And yet, to your point, I might be in even more awe of what a simple molecule can do. ...And I agree about 23andMe. I wonder whether, if they had started a little smaller in a certain sense, or maybe with some of the "experiment of nature" cases, they'd be in better shape. But the company ramped up so rapidly and I think that makes it hard to learn lessons along the way and pivot. ...Do you work in medicine or science now?
Keep your ~ramblings~ coming, David! They surely inspire my science writing effort 😊 About 23andMe, I recall attending a lecture in my graduate school days and being overwhelmed by their ramping publication numbers. My professors were very skeptical... Having studied a little quantitative genetics back then, so was I about how far we could go on translating a bunch of genotypes to day-by-day life. But the resident Polyanna in me hoped for the best. Anyway, not all is lost, right? Gathered knowledge is always a win and I´m still hoping it´ll serve us sometime.
Today I´m working in a biotech company specialized in custom-made antibodies and heterologous proteins. Science is fun in all its facets, but I must admit a miss playing around more with DNA :)
Great article. I've been following 23&Me for years and was so excited when they went public. The mom of the founder of 23&Me wrote a great parenting book. Anyways, I ended up investing in the company with high hopes but lost out big time. I think your explanation makes sense. It seems like when founders come from tech or finance and try to enter medicine/pharma, it never goes anywhere. Which is unfortunately because, as someone who works in healthcare, it feels so ripe for disruption. As Prianda said, I hope people keep trying to improve healthcare from a genetics standpoint, pharmaceutical standpoint, and a systems and payments standpoint.
I've never read but also heard that book is great! ...Ah, I'm sorry to hear that about the investment. And I do think there's still loads of genetic work to be done that will influence medicine, but I think 23andMe should've pivoted a long time ago to a different type of research, basically. There's a neat company, I can't remember the name, but I know Lux Capital invested in them, that's taking the experiments-of-nature approach. And, speaking of investments, I made a little investment in a CRISPR company that's looking at a genetically simple condition that a ton of people have. So I think there's at least a chance that could make a big difference. My intuition (which probably isn't worth much of anything in this area), is that the systems and payments area is where there's some lower-hanging fruit that could impact an enormous number of lives. It does seem like some tech founders have made some inroads. For example, I don't know how to assess the quality of the medical care at One Medical, but when I lived in New York and used them, the administrative stuff was fantastic.
So well written! I am sorry about your friend. It makes me (and possibly all other parents) anxious about our children who are involved in demanding sports......
Hey Kelly, thanks for reading! In terms of being anxious about kids in sports, I certainly understand that. That said, for the vast majority of kids, the benefits outweigh the risks. And you can lower the risk. One of the simplest thing anyone can do is investigate their own family history a little. When I first went to conferences for families who had experienced a sudden death, I'd meet people who would say, for example: "Well, uncle Bob died in a one-car accident, and cousin Jimmy was a varsity swimmer but he drowned in a pool." And then they would realize they should get checked for conditions like HCM, and could prevent other deaths in the family. If someone in the family has died unexpectedly for unknown reasons, or for known cardiac reasons, that's a reason to get checked out. The trouble is a lot of people only learn this after a tragedy. So knowing a bit of your family history can really help determine if there's cause for concern in many cases.
Thank you! What you write seems to be in agreement with what several doctors in sports do today, they keep asking again and again about family history and this sounded a bit strange at first. It is good to know that you and science in general corroborates this attitude.
Thanks for this great post David. We were naive to think that there would be single genes coding for problems like severe mental illness and alcoholism. Even gene mutations that have some predictive value, like ApoE4 in Alzheimer's disease, aren't bullet proof.
Thanks so much for reading, Patrick. I prefer to think of us as having been optimistic;) But yes, you're right. ...This is only sort of related, but I'm reminded of years ago when I saw some press release where Mark Zuckerberg was making a huge donation for medical research, and saying that in a century we'd have no disease, or something like that. And among medical breakthroughs he mentioned is one that had already basically been blown up by new data. It gave me the feeling that because he knows how to make a website work, he got the idea that the human body is also just like a complicated website we need to figure out. But we didn't build it, and we don't understand a lot of it, and it doesn't even all make sense, so I think sometimes even the well-intentioned reductionist impulses are lacking in a certain humility.
Hey David, I've been mulling this one over. It just so happens that next Monday, we have a student with progeria transferring into our school. Today, the head of the Progeria Research Foundation came to speak at our school to name and norm the disease so that we can make the transition for the new student as smooth as possible. Just like HCM, it sounds like progeria is caused by a single mutation. (Your comparison to one typo in an entire book store is excellent. It reminds me of every space analogy that gives perspective to just how capital-B Big these big numbers are.) I was also astounded to learn there are only ~400 people in the world with progeria. 1 in 20,000,000 people are born with it, and it just so happens there will be one at my tiny charter school.
The uninformed hypothesis this is leading me to is that it seems like most genetic mutations (which my bio teacher once framed to me as experiments by evolution) seem like they do more harm than good. I just tried to look up the most recent human evolutions and they were things like lactose tolerance and skin pigmentation, but those happened 10's of 1000's of years ago. But then I also remember Thomas Lund or Priscilla Lopes-Schliep, so maybe I just have a negativity bias here. I don't know much about the topic, but what do you think?
Wow, Matt, that's amazing. I knew progeria was rare, but I didn't realize it was quite that rare. (HCM is 1 in 500, to put it in perspective, although not all of those people will be at risk of sudden death.) I hope it's a growing experience for the new student and the school! ...as far as mutations, I think your hypothesis is correct, with a caveat. I think it's understood that most mutations are probably neutral. (Or, at least they don't do anything obvious...I expect that, just as we found that "junk DNA" is not actually junk, we may find in the future that some apparently neutral mutations actually do something that is just difficult to discern, or maybe with very small effect.) But between harmful and beneficial, I think it's true that mutations are more likely to be harmful than helpful. The way this was explained to me back I was reporting on genetics was that a lot of biological systems are fairly or very highly optimized, and as with any complex system it's easier to break it by taking a hammer to one spot in the system than to improve it by taking a hammer to one spot. So we're talking about the rare cases here, which are single mutations of large effect. As far as recent evolution, there are those cases, like lactose tolerance and skin pigmentation, and other things...the sickle cell mutation turns out to confer some malaria resistance...the mutation that confers some HIV resistance predates HIV, but is thought to be not very old — perhaps younger than lactose tolerance even —and perhaps arose due to some other virus. But again these are single mutations with huge effect. I think the idea that evolution has stopped is mistaken, though. The fact that there are way more people now means there are way more mutations happening, and mutations with even modest benefit can spread very quickly, and mutations can also spread via "genetic drift," which just means chance spreads a mutation through a population. With a lot more mutations around, that happens. But most of these mutations are not hammers, and, of course, even the fast changes are slow on the scale of a human life. There's a controversial book, called The 10,000 Year Explosion, that argues that evolution has been speeding up. In any case, I think your hypothesis about the greater prevalence of deleterious mutations compared to beneficial ones is accepted knowledge, for now, anyway;)
Hey Jeff, great point. I've been thinking about that, in light of what I think has been the most astounding AI success so far — the protein folding problem. Even so, I think there may be some fundamental limits here, or at least points at which accounting for a slightly larger amount of variance that doesn't have much predictive value is extremely difficult. But I don't think it'd make sense to curtail the future possibilities at this point. Who knows what will be possible! Perhaps we'll have good enough basic models of basic building blocks that AI can do simulations and predict outcomes the way it does now for text, but for gene variations. I think AI has already made an impact in determining so-called "polygenic risk scores," which may be useful in some cases...and already being used by people who want to select embryos that they think will produce a human who is a bit smarter, or taller.
Very good summary of challenges of genetic medicine. While there have been success stories (see BRCA mutation), genetics are very complicated, as you said. It will be interesting to see where 23andme goes from here.
Completely agree. I should say, I think hitting all sorts of dead ends is a feature, not a bug in scientific research. That said, in this case, some of the pivots could've come a lot earlier. And I worry that the disconnect between hype and reality can damage either research efforts or public perception going forward. I hope that's not the case.
I would be interested to hear your thoughts on pharmacogenomics. Whilst not specifically designing personalised drugs, there are studies that show response to drugs changes depending on the genome and being able to better predict drug response pre-prescription could lead to more timely and effective treatment with more optimal medication rather than the trial and error that currently exists.
Tom, I think conceptually this is a great area, and makes a ton of sense. I also think it has the distinct advantage of allowing a type of experimentation that can provide clear feedback in way that seems unlikely to me in other areas of genetic research. At the same time, I'm cautious. Not that I'm up on the cutting edge here, but I remember particularly the enthusiasm about determining dosing for the blood thinner warfarin based on genetics. (If I recall, the FDA even updated a label to give some info on genetics and dosing.) It looked great at first, but over time I don't think it has lived up to what looked like the initial promise. After the first rush, there were studies with conflicting findings; sometimes gene-based dosing helped, sometimes it seemed to do nothing, or maybe it only worked in certain subpopulations that aren't that easy to identify ahead of time. And I think ultimately the feeling was that other patient factors contributed more to overall patient variability. In the end, I think while there's something real there it mostly went by the wayside in practical terms. I don't think it means that'll always be the case, but I guess I'd be cautious about the practical benefit.
Oh, very interesting about warfarin! I had first heard about warfarin and genetic link about 15 years ago, and I had not kept up. Thanks for this update! Agree that it warfarin and genetic dosing was promoted as great example of personalized medicine.
Hi David, great article, I worked in diagnostics for 20 years, I know it was not the focus of your piece, but I think another reason 23andMe failed was because of their "one and done" business model - once you used their test and got your results, you never have to do it again. Very few businesses survive if their customers never return. So I think they were looking for ways to monetize the data, then ran into the genetic problem you describe. I still commend the founders for the aspiration to make genetic testing available to the average person though.
Dave, thanks so much for this comment. I think you're right about that aspect of the business model. Or, at least, to the extent they were relying on drug development to be the real business model, it didn't leave them much to fall back on, although perhaps they can find some (smaller) foothold with ancestry? More importantly, I appreciate your point about the aspiration. I felt that way too. Mike Joyner, who mentioned above, disagreed with me on that — he felt that people would be more prone to adverse learning based on results they couldn't do anything about, feeling fatalistic, or, alternatively invincible. When I was at Sports Illustrated, I interviewed a bunch of pro athletes in sports like football, boxing, and mma, and asked if they'd get a test like ApoE to see if they might more predisposed to lasting damage from brain trauma. They generally said that it might not change their decisions, but they'd be interested to know. And some suggested that as they got close to retirement, perhaps it would be one factor to put into the decision-making stew. That seems pretty reasonable to me. Additionally, 23andMe and others did a great job at making certain tests affordable. I was able to get a test to see if I was a Tay-Sachs carrier very easily and cheaply. For others I know who have had to do that through the traditional route, it takes forever and is really expensive. That's a spot where 23andMe should already be disruptive, in my opinion. Thanks again for this comment, and I'm glad to be aware of your Substack! Looks fascinating
Hi David,
The psychology around testing is interesting, some people want to know their genetic profile/risks, and others don't. At one of the companies I worked at, we made HIV tests, and I was surprised to learn that many people avoid getting HIV tests because they just don't want to know, even if knowing might save their life. Same for people who avoid getting their colonoscopy, not just because the procedure is gross (although that is a factor), but many people just don't care to know.
A genetic test moves slightly into the realm of asking people the strange philosophical question of whether or not they would like to know the exact day they're going to die. Some would, but many wouldn't. There is something both psychologically and philosophically freeing about NOT knowing certain things. This might be an intriguing area of study - why do people avoid certain kinds of information that could help them? Is it a net pro or con to know you are going to have a degenerative disease later in life? Sometimes there is value in knowing, so that steps might be taken to avoid future harm, but sometimes there is an interesting kind of freedom in intentional ignorance. I do think it's better to have the option to know available, as opposed to withholding it, and let people make their own decision.
Thanks for checking out my Substack, I'm going all in on embracing generalism for both life and "portfolio career". Thanks for writing Range, it's opened up many more people to the value of generalism in the same way that Susan Cain's Quiet opened up the discussion on the value of introverts.
Dave, I think you and I may have similar brains! By the end of your first paragraph I was thinking, "kind of like asking people if they want to know when they'll die," and then there it was. I agree it's better to have the option. I recall that Robert Green did a few studies, again regarding ApoE, looking at different aspects of disclosure. And I think a conclusion was that the option of testing is a positive for a lot of people, and when it's a negative it's often a transient one. Of course, this was among people who opted in to testing, but I don't think anyone is talking about anything other than that anyway. (Here's one example: https://www.nejm.org/doi/full/10.1056/NEJMoa0809578). ...Lastly, in terms of our similar brains: I take the Susan Cain comparison as a huge compliment. As an introvert, I really appreciated that book (and have since gotten to know here a little and she's truly a lovely person), and also think my writing style is probably about as similar to hers as anyone I can think of. The title and subtitle, as you may have gleaned, are sort of patterned after her example.
Hi David, haha nice to meet a "brain twin", maybe there is something to this generalist mindset after all. Oh and I'm also an introvert :)
I heard you on a podcast a while back and enjoyed how you were answering the questions, you were explaining things the same way I like to learn them, with a sense of curiosity, at depth, with thoughtful analysis, plus some amount of fascination in discovering unexpected things. In some sense it was frustrating though, because I felt like you had more information and insights in your pocket than you were given time to convey, but such is the nature of podcast interviews.
I never noticed that your book title/subtitle was similar to "Quiet", but now that you mention it, I see the similarity. I follow Susan on here and she does seem like a genuinely kind person. I enjoyed her book as well.
The ApoE study results are interesting but not necessarily surprising, that people who were negative experience some relief, while those who were positive some anxiety, but maybe not as much lasting difference as one might've expected. I do think this has something to do with the fact that Alzheimers is a) far off into the future, b) something that we generally accept might happen as we age, and c) at the moment we're lacking strong evidence on what you can do to prevent it. A more devastating, worrisome disease diagnosis e.g. HIV or cancer would've obviously produced different results in this type of study.
Loved it! Very well put. By the way, I am looking at a copy of your The Sports Gene :)
I got a PhD in Genetics and Molecular Biology more than 10 years ago and I am still mesmerized by how a molecule as simple as DNA can hold so much information, heredity and evolution power. BuT I´ve never wanted to get my cheeks swabbed. It´s a long, looong way from genotype to phenotype for most cases and that´s probably because we’ve underestimated how different (if not all!) loci interplay to produce the "seeable". Not to mention the mighty environment. It´s a shame that 23andMe killed the drug discovery initiative. I don´t see how we can get there if not by pushing the science.
Prianda! Oh geeze, I try to pretend that nobody who actually knows what they're talking about reads my ramblings here;) In any case, I agree with you on all fronts. I minored in astronomy in college, and was just in awe of the largeness and complexity of the cosmos. And yet, to your point, I might be in even more awe of what a simple molecule can do. ...And I agree about 23andMe. I wonder whether, if they had started a little smaller in a certain sense, or maybe with some of the "experiment of nature" cases, they'd be in better shape. But the company ramped up so rapidly and I think that makes it hard to learn lessons along the way and pivot. ...Do you work in medicine or science now?
Keep your ~ramblings~ coming, David! They surely inspire my science writing effort 😊 About 23andMe, I recall attending a lecture in my graduate school days and being overwhelmed by their ramping publication numbers. My professors were very skeptical... Having studied a little quantitative genetics back then, so was I about how far we could go on translating a bunch of genotypes to day-by-day life. But the resident Polyanna in me hoped for the best. Anyway, not all is lost, right? Gathered knowledge is always a win and I´m still hoping it´ll serve us sometime.
Today I´m working in a biotech company specialized in custom-made antibodies and heterologous proteins. Science is fun in all its facets, but I must admit a miss playing around more with DNA :)
Great article. I've been following 23&Me for years and was so excited when they went public. The mom of the founder of 23&Me wrote a great parenting book. Anyways, I ended up investing in the company with high hopes but lost out big time. I think your explanation makes sense. It seems like when founders come from tech or finance and try to enter medicine/pharma, it never goes anywhere. Which is unfortunately because, as someone who works in healthcare, it feels so ripe for disruption. As Prianda said, I hope people keep trying to improve healthcare from a genetics standpoint, pharmaceutical standpoint, and a systems and payments standpoint.
I've never read but also heard that book is great! ...Ah, I'm sorry to hear that about the investment. And I do think there's still loads of genetic work to be done that will influence medicine, but I think 23andMe should've pivoted a long time ago to a different type of research, basically. There's a neat company, I can't remember the name, but I know Lux Capital invested in them, that's taking the experiments-of-nature approach. And, speaking of investments, I made a little investment in a CRISPR company that's looking at a genetically simple condition that a ton of people have. So I think there's at least a chance that could make a big difference. My intuition (which probably isn't worth much of anything in this area), is that the systems and payments area is where there's some lower-hanging fruit that could impact an enormous number of lives. It does seem like some tech founders have made some inroads. For example, I don't know how to assess the quality of the medical care at One Medical, but when I lived in New York and used them, the administrative stuff was fantastic.
So well written! I am sorry about your friend. It makes me (and possibly all other parents) anxious about our children who are involved in demanding sports......
Hey Kelly, thanks for reading! In terms of being anxious about kids in sports, I certainly understand that. That said, for the vast majority of kids, the benefits outweigh the risks. And you can lower the risk. One of the simplest thing anyone can do is investigate their own family history a little. When I first went to conferences for families who had experienced a sudden death, I'd meet people who would say, for example: "Well, uncle Bob died in a one-car accident, and cousin Jimmy was a varsity swimmer but he drowned in a pool." And then they would realize they should get checked for conditions like HCM, and could prevent other deaths in the family. If someone in the family has died unexpectedly for unknown reasons, or for known cardiac reasons, that's a reason to get checked out. The trouble is a lot of people only learn this after a tragedy. So knowing a bit of your family history can really help determine if there's cause for concern in many cases.
Thank you! What you write seems to be in agreement with what several doctors in sports do today, they keep asking again and again about family history and this sounded a bit strange at first. It is good to know that you and science in general corroborates this attitude.
Thanks for this great post David. We were naive to think that there would be single genes coding for problems like severe mental illness and alcoholism. Even gene mutations that have some predictive value, like ApoE4 in Alzheimer's disease, aren't bullet proof.
Thanks so much for reading, Patrick. I prefer to think of us as having been optimistic;) But yes, you're right. ...This is only sort of related, but I'm reminded of years ago when I saw some press release where Mark Zuckerberg was making a huge donation for medical research, and saying that in a century we'd have no disease, or something like that. And among medical breakthroughs he mentioned is one that had already basically been blown up by new data. It gave me the feeling that because he knows how to make a website work, he got the idea that the human body is also just like a complicated website we need to figure out. But we didn't build it, and we don't understand a lot of it, and it doesn't even all make sense, so I think sometimes even the well-intentioned reductionist impulses are lacking in a certain humility.
Hey David, I've been mulling this one over. It just so happens that next Monday, we have a student with progeria transferring into our school. Today, the head of the Progeria Research Foundation came to speak at our school to name and norm the disease so that we can make the transition for the new student as smooth as possible. Just like HCM, it sounds like progeria is caused by a single mutation. (Your comparison to one typo in an entire book store is excellent. It reminds me of every space analogy that gives perspective to just how capital-B Big these big numbers are.) I was also astounded to learn there are only ~400 people in the world with progeria. 1 in 20,000,000 people are born with it, and it just so happens there will be one at my tiny charter school.
The uninformed hypothesis this is leading me to is that it seems like most genetic mutations (which my bio teacher once framed to me as experiments by evolution) seem like they do more harm than good. I just tried to look up the most recent human evolutions and they were things like lactose tolerance and skin pigmentation, but those happened 10's of 1000's of years ago. But then I also remember Thomas Lund or Priscilla Lopes-Schliep, so maybe I just have a negativity bias here. I don't know much about the topic, but what do you think?
Wow, Matt, that's amazing. I knew progeria was rare, but I didn't realize it was quite that rare. (HCM is 1 in 500, to put it in perspective, although not all of those people will be at risk of sudden death.) I hope it's a growing experience for the new student and the school! ...as far as mutations, I think your hypothesis is correct, with a caveat. I think it's understood that most mutations are probably neutral. (Or, at least they don't do anything obvious...I expect that, just as we found that "junk DNA" is not actually junk, we may find in the future that some apparently neutral mutations actually do something that is just difficult to discern, or maybe with very small effect.) But between harmful and beneficial, I think it's true that mutations are more likely to be harmful than helpful. The way this was explained to me back I was reporting on genetics was that a lot of biological systems are fairly or very highly optimized, and as with any complex system it's easier to break it by taking a hammer to one spot in the system than to improve it by taking a hammer to one spot. So we're talking about the rare cases here, which are single mutations of large effect. As far as recent evolution, there are those cases, like lactose tolerance and skin pigmentation, and other things...the sickle cell mutation turns out to confer some malaria resistance...the mutation that confers some HIV resistance predates HIV, but is thought to be not very old — perhaps younger than lactose tolerance even —and perhaps arose due to some other virus. But again these are single mutations with huge effect. I think the idea that evolution has stopped is mistaken, though. The fact that there are way more people now means there are way more mutations happening, and mutations with even modest benefit can spread very quickly, and mutations can also spread via "genetic drift," which just means chance spreads a mutation through a population. With a lot more mutations around, that happens. But most of these mutations are not hammers, and, of course, even the fast changes are slow on the scale of a human life. There's a controversial book, called The 10,000 Year Explosion, that argues that evolution has been speeding up. In any case, I think your hypothesis about the greater prevalence of deleterious mutations compared to beneficial ones is accepted knowledge, for now, anyway;)
This seems like a prime puzzle/target for the next generation of AI.
Hey Jeff, great point. I've been thinking about that, in light of what I think has been the most astounding AI success so far — the protein folding problem. Even so, I think there may be some fundamental limits here, or at least points at which accounting for a slightly larger amount of variance that doesn't have much predictive value is extremely difficult. But I don't think it'd make sense to curtail the future possibilities at this point. Who knows what will be possible! Perhaps we'll have good enough basic models of basic building blocks that AI can do simulations and predict outcomes the way it does now for text, but for gene variations. I think AI has already made an impact in determining so-called "polygenic risk scores," which may be useful in some cases...and already being used by people who want to select embryos that they think will produce a human who is a bit smarter, or taller.
Very good summary of challenges of genetic medicine. While there have been success stories (see BRCA mutation), genetics are very complicated, as you said. It will be interesting to see where 23andme goes from here.
Completely agree. I should say, I think hitting all sorts of dead ends is a feature, not a bug in scientific research. That said, in this case, some of the pivots could've come a lot earlier. And I worry that the disconnect between hype and reality can damage either research efforts or public perception going forward. I hope that's not the case.
I would be interested to hear your thoughts on pharmacogenomics. Whilst not specifically designing personalised drugs, there are studies that show response to drugs changes depending on the genome and being able to better predict drug response pre-prescription could lead to more timely and effective treatment with more optimal medication rather than the trial and error that currently exists.
https://www.nature.com/articles/s41576-022-00572-8
Tom, I think conceptually this is a great area, and makes a ton of sense. I also think it has the distinct advantage of allowing a type of experimentation that can provide clear feedback in way that seems unlikely to me in other areas of genetic research. At the same time, I'm cautious. Not that I'm up on the cutting edge here, but I remember particularly the enthusiasm about determining dosing for the blood thinner warfarin based on genetics. (If I recall, the FDA even updated a label to give some info on genetics and dosing.) It looked great at first, but over time I don't think it has lived up to what looked like the initial promise. After the first rush, there were studies with conflicting findings; sometimes gene-based dosing helped, sometimes it seemed to do nothing, or maybe it only worked in certain subpopulations that aren't that easy to identify ahead of time. And I think ultimately the feeling was that other patient factors contributed more to overall patient variability. In the end, I think while there's something real there it mostly went by the wayside in practical terms. I don't think it means that'll always be the case, but I guess I'd be cautious about the practical benefit.
Oh, very interesting about warfarin! I had first heard about warfarin and genetic link about 15 years ago, and I had not kept up. Thanks for this update! Agree that it warfarin and genetic dosing was promoted as great example of personalized medicine.