>By contrast, intrinsic mortality stems from processes originating within the body, including genetic mutations, age-related diseases, and the decline of physiological function with age
So we put genetic diseases in the bucket of intrinsic mortality and then found that intrinsic mortality has a heritable component?
Yeah this paper came across to me basically as "if you ignore environmental causes of death, the heritability of death goes up"... which seems kind of circular.
Not necessarily. It could be the case that randomness plays a huge part in non-environmental caused deaths, and if that were the case we would see very little heritability.
I really like everything Uri Alon (last author) publishes, but these types of studies have a history of inflating genetic contributions to phenotypes. Decoupling genetics from environment is not easy as they are both highly correlated.
In fact, the article discussion states: "Limitations of this study include reliance on assumptions of the twin design, such as the equal environment assumption". My take on this is that the main result of the article is probably true, but the 50% figure is likely to be inflated.
I went in early one morning to get an echocardiogram and was practicing a technique that opens people up emotionally and had run into several damp squibs in a row -- but I was feeling irrepressible.
I hit the jackpot with the ultrasound technician who spoke passionately about what she believed about lifestyle risk for cardiovascular conditions and she believed quite strongly that heart disease runs in families more because lifestyle runs in families than because of genetics. She's not at the top of the medical totem pole but I can say she inspired me to take responsibility for my health than the specialist who I talked to about the results.
There's no prior reason to expect the cited conditions to have any specific relation to genetics. Any of them could easily be caused or accelerated by environmental conditions.
Yeah, it’s important to note that heritability is a statistic about today’s population, not a deep natural parameter that tells you about causality. Heritability of smoking went up when smoking became less socially approved, for example.
In case anyone was curious like me: the standard deviation of lifespan is ~12-15 years in developed countries.
So environmental effects, sleep, diet, lifestyle, etc (I.e. modifiable factors) maybe account for half of that, so like 6-7.5 years of variance. Which… sounds about right to me.
Lifespan is not even half the story though, health span is much more important. Your life is completely different if you can ski or split your own wood at 80+ vs being barely able to use stairs at 50. Both might die at 90 but one "lived" 30 years more
I'm not really afraid of getting old, but I'm afraid of becoming decrepit.
My grandma has been decrepit for over 5 years now. She can't walk and has no bladder or bowel control, so she just sits on the couch and shits herself all day. She's not living, she's merely surviving. She was living with my mom for a while, but my mom decided she couldn't handle it anymore and put her in an assisted living facility.
If I get to the point where I couldn't cook my own meals and wipe my own ass, just put a bullet in me. I do not fear dying, but I do fear spending years of my life not being able to actually do anything.
My dad died at the end of last year, and was not too different from your grandma. For him the main problem was chronic pain from his failing body. Even fairly powerful opioids from a pain management doctor only helped a bit. Basically all he could do was sleep, eat meals, and sit in his chair in pain.
I feel similar to you, but I wonder if it's one of those those things where age changes your perspective. Dad was in assisted living and had several stints in rehab/nursing home facilities, and in both there were quite a few people with what I'd call poor quality of life who were still holding on to life.
Something we youngsters (I'm 69) may not realize is that people in assisted living still have friends and frequently even sex lives while they are there. They read, play games, and watch movies, just like us. They might not be able to do all the things they could when they were younger, but their lives are not necessarily over.
Of course, some truly do “live” there, and good for them.
And others just sit there waiting to die, unable to even feed themselves.
I saw plenty of examples of both when my grandmothers were in assisted living homes. Unfortunately my grandmothers both tended towards the latter case.
It is probably more than half the story. Health span is strongly correlated to life span, although not completely. The median "health span gap" is about 10 years, and has widened by roughly one year over the past 20 years. However, this is probably just due to an aging population and not necessarily from any factors you can control fully.
I wouldn't be surprised if "health span" (although defining it is difficult) exactly mirrors the inheritability pattern of mortality.
It depends on the definition, if you're even just 20kg overweight you're living a wildly different life than you'd have if you were fit, you're closing so many doors by default and making a bunch of things much harder than they should be, But you're still considered "healthy" here
My Dad (age 81) tore his rotator cuff splitting wood recently. It's slow to heal and he's in a lot of pain which (along with his Alzheimer's) is really getting him down.
Maybe even if you're still fit and strong in your 80s you should let someone else split your wood for you
Yeah, been working in IT since forever (sitting work all day), but started lifting recently and it already made remarkable improvements in my wellbeing. Should've started sooner of course, but I'm still well in time.
This plus stretching / yoga has been amazing as I'm entering my 40s. For a while I was just lifting and I had strong muscles but they were short and tight. Not everyone has that problem, but just noting strong muscles are half the picture, being strong and flexible makes life feel effortless and years of being a desk jockey.
One of the most consistent health research findings Ive heard in recent years is the benefits of weight training for older adults. Hopefully the message is being received.
Life span is easier to measure. You get the offial birth dates table, you get the official death dates table, you just substract the numbers and call it a day.
Too many people think your life is a binary 'living or dead' when thats not the case at all. I didn't even understand it fully till I was hit by a car.
It is almost never reasonable to assume normality and make calculations like this. This is particularly the case when you are dealing with lifespan, which isn't normally-distributed even in the slightest. The actual ranges are likely smaller than you are stating here, and variance is just not a very practical or interpretable metric to use when dealing with such a skewed distribution.
We should be stating something like a probability density interval (i.e. what is the actual range / interval that 95% of age-related deaths occur within), and then re-framing how much genetic variation can explain within that range, or something like it. As it is presented in the headline / takeaway, the heritability estimate is almost impossible to translate into anything properly interpretable.
This is a nice example/re-stating of what the heritability % "means" here.
I'm curious, with something like smoking/drinking, how you can be confident that you've untangled genetic predispositions to addiction or overconsumption from those "modifiable factors". I guess that's just captured within the 50% heritability? And if you could confidently untangle them, you might find heritability is higher than 50%?
Heritability is a pretty funky concept because it's contextual to a certain point in time, environment, and population, effectively.
An example I like is that if you measured the heritability of depression in 2015, and then you measured the heritability of depression in 2021, you would likely see changes due to environmental effects (namely, there's the pandemic/lockdowns and this could conceivably cause more people to experience depressive symptoms). Let's assume we make those measurements and the rate of depression did increase, and we could tie it causally to the pandemic or related events.
In that scenario, the heritability of depression would have decreased. I don't think anyone would argue there were massive genetic changes in that 6 year time period on a population scale, but the environment changed in a way that affected the population as a whole, so the proportion of the effect on the trait which is genetically explained decreased.
For something like lifespan in the above example, you can imagine that in a period of wartime, famine, or widespread disease the heritability would also decrease in many scenarios (if random chance is ending a lot of lives early, how long the tail of lifespan is influenced genetically is much less important).
Given that note, it's generally tricky to talk about whether heritability increases or decreases, but with more accurate estimates of how genetic predispositions form you could see the heritability of certain traits increase with the environment held stable, as there's certainly ones that may be underestimated or genetic factors that aren't currently accounted for in many traits.
*edit: I realized I never mentioned the other thing I wanted to mention writing this! since you mentioned what the percent heritability means here, I think the best way to think of it is just "the proportion of phenotypic variation for this trait in a measured population which is explained by genetic variation." So it's dependent on the amount of variation in several aspects (environmental, genetic, phenotypic).
Some epigenetic effects are semi-heritable too, eg maternal exposure can be transmitted. That's in addition to environmental effects like you mentioned. Two otherwise identical cohorts can inherit the same genetic predisposition for depression where one manifests and another does not entirely due to their circumstances.
Environmental effects are not necessarily modifiable. It includes randomness, background radiation, unknown risk factors, anything which is not genetic.
> the standard deviation of lifespan is ~12-15 years in developed countries.
That seems rather higher than I would have expected, at least if one corrects for preventable accidents and other such things (that I would expect to shift the results away from a normal distribution).
> at least if one corrects for preventable accidents and other such things
You can't really correct for these. Yes there are genuine accidents that will kill you under any circumstances, but for a lot of things both your odds of having an accident and the odds of surviving it are strongly linked to age. As a simple example, despite driving significantly less, the elderly get into more car accidents and suffer worse injuries in those accidents than people earlier in life. Only the age range of 15-24 has higher car accident fatality rates.
There is no such thing as death by old age. At most there are deaths in the elderly that don't get attributed to a specific cause (typically because of so many different things going on at once and no desire to cut up grandma after the fact to see which straw broke her back) which we tend to refer to as "died of old age" but it's not a recognized medical cause of death. People die of diseases, injuries, and various other things, many of which are strongly influenced by age but also heavily influenced by other factors.
You can set a cutoff point and say these things don't count as age related deaths whereas these others do. As long as you're consistent with these choices, you can learn something useful. But a wide enough net that is widely agreed to cover what we think of as aging is going to include a lot of other maladies, whereas a narrower selection criteria is probably going to yield wildly different results from one analysis to the next.
Lifespan is a quite skewed distribution, so the SD looks large because it is in fact a poor summary of the bulk of the distribution. The actual part we care about for age-related mortality is narrower than such an SD would imply if we had a normal distribution (simple image example: https://biology.stackexchange.com/a/87851).
Lifespan isn't as important as healthy lifespan. Lifestyle can mean the difference between being able to complete an Ironman triathlon at age 80 vs being bedbound.
The 50% number is a bit mysterious, but if I understand the text of the article correctly, it essentially means that if we do not account for the noise added by accidents and such, we get a Pearson correlation of life expectancies of monozygotic twins of ~0.23. If we correct for accidents, the correlation rises to 0.5, hence 50% (with some further analysis they go up to 0.55, hence "above 50%" in the abstract). Now, in practical terms, this means that, given a MZ twin who died recently of natural causes, we could obtain an estimate for ourselves, but only if we make additional assumptions. A correlation coefficient alone is not very informative.
>Now, in practical terms, this means that, given a MZ twin who died recently of natural causes, we could obtain an estimate for ourselves,
Uh... am I misreading your comment, or are you suggesting that when your identical twin dies of non-accidental death, you can be pretty sure you're about to croak in the next wee days or weeks yourself? Very difficult to engineer that alarm bell (you either have a twin, or not), and too damned late to matter.
Keep in mind this research is based on correcting twin study heritability estimates for confounding effects. However, new research shows that heritability estimates derived from twin studies are themselves dramatically inflated: https://open.substack.com/pub/theinfinitesimal/p/the-missing...
I work in the field and I think he is quite good at providing a fair layman’s overview. He’s also famous (or notorious) for having few barriers in terms of who he reads/responds to.
One of these guys he’s citing has identified himself as a Nazi. Maybe Scott should start exercising a little discernment here. Regardless, he’s making a pretty weak argument against Sasha’s interpretation of the study.
There's a lot of genes that impact lifespan, both good and bad. For example my father has hereditary hemochromatosis due to 2 copies of the HFE C282Y mutation. He was diagnosed in his 50's, so I'd expect the damage it did to his body to impact lifespan.
In my case I don't have it (I'm just a genetic carrier). If I did have the genotype and took the necessary dietary measures to avoid the phenotype, then it likely wouldn't impact lifespan.
On one hand you can argue a heritable disease like HHC has an impact on lifespan, but with genetic testing and treatment you can argue it doesn't impact lifespan (or it's impact is significantly mitigated).
Geneticist would say it's intended design of twin studies. Your example doesn't show AIDS is a genetic defect, but that genetics may predispose/protect against it.
There's serious issues with heritability research in general, it's observability stuff -- not experimentation, so imv, its at best proto-science, and in many cases plainly pseudoscience. "Heritability" itself has little to do with whether something is inherited, and speaks only to correlation with genes. Since we have a vast amount of genes which are shared for all sorts of reasons (ie., mating is based on shared culture, wealth, geography, etc.) -- the metric is mostly useless.
Accents are highly heritable, since they always correlated with location which is always correlated with genes.
Even if you do these twin studies, you have to assume a model of how genes and the environment interact, and all such models are obviously false.
Thus even if you grant that heritability measures on high quality twin studies are 'sign correct', in the sense that they show P(genetic effect) > P(no genetic effect) -- any magnitude of this effect, or any theory of is, is more or less pseudoscience (unless there are experimental studies showing gene-trait mechanism).
For example, it is "obvious" that P(genetic effect) > P(none) for intelligence, since genes control the structure of the brain and body. But there is no evidence (I'm aware of...) that beyond provision of a functioning brain, our genetics play any role in intelligence stratification. ie., all correlation with task performance and IQ can be explained by correlations in the metal retardation / mental deficiency range.
This doesn't mean intelligence is very malleable beyond a certain age. My own views is that genes are basically providing functioning hardware to the womb, and after that point its early development (both pre-birth and probs up to at most 3yo) which locks in a lot of the observed intelligence stratification. This is a very different story than popularisers of IQ research communicate though, but be aware, none are very good scientists and most of this research is methodologically unfit
Genetics may predispose for nicotine addiction, obesity, alcoholism, etc. This is intended design of genetic studies which look at multiple genes, like twin studies, GWAS studies, etc.
I think the studies which find a single gene variant which would have large impact on lifespan would be interesting. Not sure if variants like that exist though.
This finding rectified my mental model of longevity after a long, perplexing period where longevity was estimated to be much less heritable than expected when comparing to other studied traits.
I know enough about heritability to know that the science people use words differently than I expected, but not enough to explain that so here's someone's article about it:
Heritability IMO falls into the same bin as "standardized/relative effect sizes" (e.g. correlation coefficients, Cohen's d, odds-ratios, "explained variance", relative risk, etc), in that a division / re-scaling is introduced to supposedly increase interpretability, but, in reality, this has precisely the opposite effect.
Heritability is a bit worse though because the variance is partitioned into three giant piles of mush, at least two of which piles are very poorly measured / controlled at all.
How is heritabiltity of life span useful if by the time the lifespan becomes known (eg at 80yrs old) the inheritance is not possible anymore (eg menopause)?
Heritability acts on lineages, not individuals (in general, not always) - a good rule of thumb is that traits that benefit 3 or more generations of a family have a good shot at being propagated. In this case, the advantage (of both menopause and longevity) is increased well-being of the tribe, ampliyfing the positive effects of culture and stability. Wisdom of the elders is implicit to the genetics. This is a tradeoff with the cost in resources; at some point the cost to keep someone around might exceed the benefit, but from an evolutionary standpoint, the accounting is over a lifetime; in a relatively stable environment, genes that improve longevity and healthspan will be reinforced by the positive feedback loops of culture and nurture and civilization and technology. Menopause is also prevalent in orcas and a handful of other mammals - and older females help rearing and protecting babies, and so forth, with a protoculture providing that feedback loop.
It's not useful. Indeed that's likely why we die of old age - there is no selective pressure to remove harmful mutations that don't reduce your ability to pass on your genes so such harmful mutations just accumulate over many generations. You might have a mutation that will cause your heart to rupture at age 150, but you'll never know it because you'll die of something else first.
It is possible though to selectively breed animals like flies for long lifespan. You wait to see how long one generation lives and cull the descendants of those that died early. It's inefficient but lifespan extensions of 50-60% have been demonstrated. One could imagine through gene editing that a species might be able to reap the benefits without the culling.
Note that evolution isn't about individuals, it's about genes (which we should further note is more than just DNA, but that's a different discussion). If it weren't valuable for humans to live at all without being fertile, then the average age of menopause and the average age of death would likely be much closer together. As it turns out, the human genes that were best able to pass themselves along were those that kept old people around despite being infertile, presumably to the benefit of helping to raise grandchildren, among other things.
Healthy grandparents that are around to support their children and take care of grandchildren increase the fitness of the entire lineage by helping their children have more children and those grandchildren to be healthier/safer.
If you are interacting with a carrier of your genes at all while they still might reproduce, you are having an impact on their fitness and thus evolutionary pressure exists.
We're social animals. Anything that benefits the extended tribal unit is advantageous. Adults beyond child-bearing age contribute significantly to child raising, education, leadership etc of the entire tribe.
In (quantitative) genetics literature, heritability is usually defined (simplifying a bit) as the proportion of variance of a trait (lifespan, height, etc), in a population, that can be explained by genetics. The rest, by environmental factors, or error.
If height were a 100% heritability means that all differences in height between individuals would be explainable by genetics.
Those old people used to be young and helped pay for their parents and grandparents to live into old age. Part of being young and productive is helping take care of those less able than you, including the elderly.
Unless you're volunteering to work for 40 years then be executed on retirement, I think you should delete that comment and that thought from your mind.
Given the opportunity, one wonders what you'd like to do with 'old people consuming resources without adding anything and supposedly 'holding back societies'. Earlier in the 20th century a significant cohort of intellectuals had decided ideas on this and the earlier generations.
Related somewhat to this: 'The Intellectuals and the Masses' by John Carey makes for truly shocking reading.
It's interesting on my mother's side of the family, most everyone lived well into their 80's and 90's. The execution being for my Mom and her sisters who smoked heavily. Her brothers both died in their 60's but were in the Vietnam war and were definitely exposed to Agent Orange and both had brain cancer. My dad lived until nearly 80 after smoking since he was 12 years old and 2-3 packs per day.
This is incorrect. Twin studies typically compare MZ twin similarity against (same sex, usually) DZ twin similarity. Assuming that there is nothing special about MZs for the trait (e.g. in this case if MZ twins lived longer by virtue of being MZ twins), you can estimate heritability free of shared environments.
> We estimated uncorrected heritability (uncorrected for extrinsic mortality) (materials and methods) in three independent ways: (i) MZ twins reared apart (n = 150), (ii) DZ twins reared apart (n = 371), and (iii) MZ versus DZ twins reared together (196 MZ, 325 DZ)
This is from _one_ of the datasets they examined, but there were also two others. n=150 twins reared apart in their small category, or n=520 twins reared apart total is the lower bound of data they had, and even that is not too shabby imo
I don't have an opinion to offer here other than the intrinsic limitations of studies that depend on twins raised apart (that there aren't many of them). It's an unusual instance of a stat where the obvious concern with the premise is underappreciated rather than overappreciated. I've got nothing on MZ/DZ controls.
ah I see, you're commenting on the general difficulty, not necessarily saying this study's results are bad due to the limitation. My apologies, I don't think we disagree.
Yes but those aren’t random samples. Children not raised with their birth parents had different circumstances. As did children who got split up, and families adopting children is also a selection bias.
Faster evolution does not necessarily translate to better outcomes. Exhibit A: the respiratory capabilities of the brachycephalic pug. Exhibit B: the rabbit fear response – they can get so terrified that they break their own spines trying to escape. Exhibit C: every creature with a hybrid r/K reproductive strategy involving child- or sibling-cannibalism.
>By contrast, intrinsic mortality stems from processes originating within the body, including genetic mutations, age-related diseases, and the decline of physiological function with age
So we put genetic diseases in the bucket of intrinsic mortality and then found that intrinsic mortality has a heritable component?
Yeah this paper came across to me basically as "if you ignore environmental causes of death, the heritability of death goes up"... which seems kind of circular.
Not necessarily. It could be the case that randomness plays a huge part in non-environmental caused deaths, and if that were the case we would see very little heritability.
But randomness comes from the environment, no?
I really like everything Uri Alon (last author) publishes, but these types of studies have a history of inflating genetic contributions to phenotypes. Decoupling genetics from environment is not easy as they are both highly correlated.
In fact, the article discussion states: "Limitations of this study include reliance on assumptions of the twin design, such as the equal environment assumption". My take on this is that the main result of the article is probably true, but the 50% figure is likely to be inflated.
I went in early one morning to get an echocardiogram and was practicing a technique that opens people up emotionally and had run into several damp squibs in a row -- but I was feeling irrepressible.
I hit the jackpot with the ultrasound technician who spoke passionately about what she believed about lifestyle risk for cardiovascular conditions and she believed quite strongly that heart disease runs in families more because lifestyle runs in families than because of genetics. She's not at the top of the medical totem pole but I can say she inspired me to take responsibility for my health than the specialist who I talked to about the results.
There's no prior reason to expect the cited conditions to have any specific relation to genetics. Any of them could easily be caused or accelerated by environmental conditions.
And, in fact, it looks like they half-of-are.
I thought the implication was lifestyle isn't as important as we previously believed.
On average! Start drinking a lot and find out.
Yeah, it’s important to note that heritability is a statistic about today’s population, not a deep natural parameter that tells you about causality. Heritability of smoking went up when smoking became less socially approved, for example.
In case anyone was curious like me: the standard deviation of lifespan is ~12-15 years in developed countries.
So environmental effects, sleep, diet, lifestyle, etc (I.e. modifiable factors) maybe account for half of that, so like 6-7.5 years of variance. Which… sounds about right to me.
Lifespan is not even half the story though, health span is much more important. Your life is completely different if you can ski or split your own wood at 80+ vs being barely able to use stairs at 50. Both might die at 90 but one "lived" 30 years more
Yup.
I'm not really afraid of getting old, but I'm afraid of becoming decrepit.
My grandma has been decrepit for over 5 years now. She can't walk and has no bladder or bowel control, so she just sits on the couch and shits herself all day. She's not living, she's merely surviving. She was living with my mom for a while, but my mom decided she couldn't handle it anymore and put her in an assisted living facility.
If I get to the point where I couldn't cook my own meals and wipe my own ass, just put a bullet in me. I do not fear dying, but I do fear spending years of my life not being able to actually do anything.
My dad died at the end of last year, and was not too different from your grandma. For him the main problem was chronic pain from his failing body. Even fairly powerful opioids from a pain management doctor only helped a bit. Basically all he could do was sleep, eat meals, and sit in his chair in pain.
I feel similar to you, but I wonder if it's one of those those things where age changes your perspective. Dad was in assisted living and had several stints in rehab/nursing home facilities, and in both there were quite a few people with what I'd call poor quality of life who were still holding on to life.
Something we youngsters (I'm 69) may not realize is that people in assisted living still have friends and frequently even sex lives while they are there. They read, play games, and watch movies, just like us. They might not be able to do all the things they could when they were younger, but their lives are not necessarily over.
I am looking forward to playing 3 decades of great computer games once I am too old to go out into the woods or do martial arts.
I love gaming, but I am still too young to do it properly.
Of course, some truly do “live” there, and good for them.
And others just sit there waiting to die, unable to even feed themselves.
I saw plenty of examples of both when my grandmothers were in assisted living homes. Unfortunately my grandmothers both tended towards the latter case.
It is probably more than half the story. Health span is strongly correlated to life span, although not completely. The median "health span gap" is about 10 years, and has widened by roughly one year over the past 20 years. However, this is probably just due to an aging population and not necessarily from any factors you can control fully.
I wouldn't be surprised if "health span" (although defining it is difficult) exactly mirrors the inheritability pattern of mortality.
> The median "health span gap" is about 10 years
It depends on the definition, if you're even just 20kg overweight you're living a wildly different life than you'd have if you were fit, you're closing so many doors by default and making a bunch of things much harder than they should be, But you're still considered "healthy" here
My Dad (age 81) tore his rotator cuff splitting wood recently. It's slow to heal and he's in a lot of pain which (along with his Alzheimer's) is really getting him down.
Maybe even if you're still fit and strong in your 80s you should let someone else split your wood for you
Yeah, been working in IT since forever (sitting work all day), but started lifting recently and it already made remarkable improvements in my wellbeing. Should've started sooner of course, but I'm still well in time.
This plus stretching / yoga has been amazing as I'm entering my 40s. For a while I was just lifting and I had strong muscles but they were short and tight. Not everyone has that problem, but just noting strong muscles are half the picture, being strong and flexible makes life feel effortless and years of being a desk jockey.
Lot of people think it's a niche exercise activity and it shouldn't be - for all ages including those in their 80s and 90s according to reports.
One of the most consistent health research findings Ive heard in recent years is the benefits of weight training for older adults. Hopefully the message is being received.
100% now that I get older I observe the even older people I know.
Some live a very painful and limited life. Others are 85+ and still go out to run, play soccer etc. Amazing to see.
Life span is easier to measure. You get the offial birth dates table, you get the official death dates table, you just substract the numbers and call it a day.
It's a remarkable tragedy how many people don't understand your point.
https://en.wikipedia.org/wiki/Disability-adjusted_life_year
Too many people think your life is a binary 'living or dead' when thats not the case at all. I didn't even understand it fully till I was hit by a car.
I'm sorry that happened.
As many of the health nutters say, the goal is "live well, drop dead."
It is almost never reasonable to assume normality and make calculations like this. This is particularly the case when you are dealing with lifespan, which isn't normally-distributed even in the slightest. The actual ranges are likely smaller than you are stating here, and variance is just not a very practical or interpretable metric to use when dealing with such a skewed distribution.
We should be stating something like a probability density interval (i.e. what is the actual range / interval that 95% of age-related deaths occur within), and then re-framing how much genetic variation can explain within that range, or something like it. As it is presented in the headline / takeaway, the heritability estimate is almost impossible to translate into anything properly interpretable.
https://biology.stackexchange.com/questions/87850/why-isnt-l...
One note: the standard deviation of the remaining effects would be sqrt(1/2) as large, not 1/2 as large. So more like 8.5-10.5 years.
This is a nice example/re-stating of what the heritability % "means" here.
I'm curious, with something like smoking/drinking, how you can be confident that you've untangled genetic predispositions to addiction or overconsumption from those "modifiable factors". I guess that's just captured within the 50% heritability? And if you could confidently untangle them, you might find heritability is higher than 50%?
Heritability is a pretty funky concept because it's contextual to a certain point in time, environment, and population, effectively.
An example I like is that if you measured the heritability of depression in 2015, and then you measured the heritability of depression in 2021, you would likely see changes due to environmental effects (namely, there's the pandemic/lockdowns and this could conceivably cause more people to experience depressive symptoms). Let's assume we make those measurements and the rate of depression did increase, and we could tie it causally to the pandemic or related events.
In that scenario, the heritability of depression would have decreased. I don't think anyone would argue there were massive genetic changes in that 6 year time period on a population scale, but the environment changed in a way that affected the population as a whole, so the proportion of the effect on the trait which is genetically explained decreased.
For something like lifespan in the above example, you can imagine that in a period of wartime, famine, or widespread disease the heritability would also decrease in many scenarios (if random chance is ending a lot of lives early, how long the tail of lifespan is influenced genetically is much less important).
Given that note, it's generally tricky to talk about whether heritability increases or decreases, but with more accurate estimates of how genetic predispositions form you could see the heritability of certain traits increase with the environment held stable, as there's certainly ones that may be underestimated or genetic factors that aren't currently accounted for in many traits.
*edit: I realized I never mentioned the other thing I wanted to mention writing this! since you mentioned what the percent heritability means here, I think the best way to think of it is just "the proportion of phenotypic variation for this trait in a measured population which is explained by genetic variation." So it's dependent on the amount of variation in several aspects (environmental, genetic, phenotypic).
Some epigenetic effects are semi-heritable too, eg maternal exposure can be transmitted. That's in addition to environmental effects like you mentioned. Two otherwise identical cohorts can inherit the same genetic predisposition for depression where one manifests and another does not entirely due to their circumstances.
Evolution is just super super messy.
Environmental effects are not necessarily modifiable. It includes randomness, background radiation, unknown risk factors, anything which is not genetic.
> the standard deviation of lifespan is ~12-15 years in developed countries.
That seems rather higher than I would have expected, at least if one corrects for preventable accidents and other such things (that I would expect to shift the results away from a normal distribution).
> at least if one corrects for preventable accidents and other such things
You can't really correct for these. Yes there are genuine accidents that will kill you under any circumstances, but for a lot of things both your odds of having an accident and the odds of surviving it are strongly linked to age. As a simple example, despite driving significantly less, the elderly get into more car accidents and suffer worse injuries in those accidents than people earlier in life. Only the age range of 15-24 has higher car accident fatality rates.
There is no such thing as death by old age. At most there are deaths in the elderly that don't get attributed to a specific cause (typically because of so many different things going on at once and no desire to cut up grandma after the fact to see which straw broke her back) which we tend to refer to as "died of old age" but it's not a recognized medical cause of death. People die of diseases, injuries, and various other things, many of which are strongly influenced by age but also heavily influenced by other factors.
You can set a cutoff point and say these things don't count as age related deaths whereas these others do. As long as you're consistent with these choices, you can learn something useful. But a wide enough net that is widely agreed to cover what we think of as aging is going to include a lot of other maladies, whereas a narrower selection criteria is probably going to yield wildly different results from one analysis to the next.
Lifespan is a quite skewed distribution, so the SD looks large because it is in fact a poor summary of the bulk of the distribution. The actual part we care about for age-related mortality is narrower than such an SD would imply if we had a normal distribution (simple image example: https://biology.stackexchange.com/a/87851).
Lifespan isn't as important as healthy lifespan. Lifestyle can mean the difference between being able to complete an Ironman triathlon at age 80 vs being bedbound.
The 50% number is a bit mysterious, but if I understand the text of the article correctly, it essentially means that if we do not account for the noise added by accidents and such, we get a Pearson correlation of life expectancies of monozygotic twins of ~0.23. If we correct for accidents, the correlation rises to 0.5, hence 50% (with some further analysis they go up to 0.55, hence "above 50%" in the abstract). Now, in practical terms, this means that, given a MZ twin who died recently of natural causes, we could obtain an estimate for ourselves, but only if we make additional assumptions. A correlation coefficient alone is not very informative.
>Now, in practical terms, this means that, given a MZ twin who died recently of natural causes, we could obtain an estimate for ourselves,
Uh... am I misreading your comment, or are you suggesting that when your identical twin dies of non-accidental death, you can be pretty sure you're about to croak in the next wee days or weeks yourself? Very difficult to engineer that alarm bell (you either have a twin, or not), and too damned late to matter.
Keep in mind this research is based on correcting twin study heritability estimates for confounding effects. However, new research shows that heritability estimates derived from twin studies are themselves dramatically inflated: https://open.substack.com/pub/theinfinitesimal/p/the-missing...
For a counterpoint to Sasha’s view, you should probably check out https://open.substack.com/pub/astralcodexten/p/the-good-news...
Anyone who cites Jordan Lasker and Kierkegaard on heritability discussions is not someone you should take seriously.
I work in the field and I think he is quite good at providing a fair layman’s overview. He’s also famous (or notorious) for having few barriers in terms of who he reads/responds to.
One of these guys he’s citing has identified himself as a Nazi. Maybe Scott should start exercising a little discernment here. Regardless, he’s making a pretty weak argument against Sasha’s interpretation of the study.
There's a lot of genes that impact lifespan, both good and bad. For example my father has hereditary hemochromatosis due to 2 copies of the HFE C282Y mutation. He was diagnosed in his 50's, so I'd expect the damage it did to his body to impact lifespan.
In my case I don't have it (I'm just a genetic carrier). If I did have the genotype and took the necessary dietary measures to avoid the phenotype, then it likely wouldn't impact lifespan.
On one hand you can argue a heritable disease like HHC has an impact on lifespan, but with genetic testing and treatment you can argue it doesn't impact lifespan (or it's impact is significantly mitigated).
In the absence of other evidence, isn't it the case that any given trait is 50% heritable and 50% environmental?
Being 100% uncertain is not the same as either option being equally likely. More like the probability of either option is undefined.
The problem with twin studies:
1. There are genetic mutations that make you immune to HIV.
2. Monozygotic twins will both be immune, or not immune, while dizygotic twins may be either, one can be immune, while the other one could get AIDS.
3. Thus, a twin study would likely show that AIDS is a genetic defect.
Geneticist would say it's intended design of twin studies. Your example doesn't show AIDS is a genetic defect, but that genetics may predispose/protect against it.
There's serious issues with heritability research in general, it's observability stuff -- not experimentation, so imv, its at best proto-science, and in many cases plainly pseudoscience. "Heritability" itself has little to do with whether something is inherited, and speaks only to correlation with genes. Since we have a vast amount of genes which are shared for all sorts of reasons (ie., mating is based on shared culture, wealth, geography, etc.) -- the metric is mostly useless.
Accents are highly heritable, since they always correlated with location which is always correlated with genes.
Even if you do these twin studies, you have to assume a model of how genes and the environment interact, and all such models are obviously false.
Thus even if you grant that heritability measures on high quality twin studies are 'sign correct', in the sense that they show P(genetic effect) > P(no genetic effect) -- any magnitude of this effect, or any theory of is, is more or less pseudoscience (unless there are experimental studies showing gene-trait mechanism).
For example, it is "obvious" that P(genetic effect) > P(none) for intelligence, since genes control the structure of the brain and body. But there is no evidence (I'm aware of...) that beyond provision of a functioning brain, our genetics play any role in intelligence stratification. ie., all correlation with task performance and IQ can be explained by correlations in the metal retardation / mental deficiency range.
This doesn't mean intelligence is very malleable beyond a certain age. My own views is that genes are basically providing functioning hardware to the womb, and after that point its early development (both pre-birth and probs up to at most 3yo) which locks in a lot of the observed intelligence stratification. This is a very different story than popularisers of IQ research communicate though, but be aware, none are very good scientists and most of this research is methodologically unfit
Genetics may predispose for nicotine addiction, obesity, alcoholism, etc. This is intended design of genetic studies which look at multiple genes, like twin studies, GWAS studies, etc.
I think the studies which find a single gene variant which would have large impact on lifespan would be interesting. Not sure if variants like that exist though.
This finding rectified my mental model of longevity after a long, perplexing period where longevity was estimated to be much less heritable than expected when comparing to other studied traits.
I know enough about heritability to know that the science people use words differently than I expected, but not enough to explain that so here's someone's article about it:
https://dynomight.net/heritable/
Heritability IMO falls into the same bin as "standardized/relative effect sizes" (e.g. correlation coefficients, Cohen's d, odds-ratios, "explained variance", relative risk, etc), in that a division / re-scaling is introduced to supposedly increase interpretability, but, in reality, this has precisely the opposite effect.
Heritability is a bit worse though because the variance is partitioned into three giant piles of mush, at least two of which piles are very poorly measured / controlled at all.
How is heritabiltity of life span useful if by the time the lifespan becomes known (eg at 80yrs old) the inheritance is not possible anymore (eg menopause)?
Heritability acts on lineages, not individuals (in general, not always) - a good rule of thumb is that traits that benefit 3 or more generations of a family have a good shot at being propagated. In this case, the advantage (of both menopause and longevity) is increased well-being of the tribe, ampliyfing the positive effects of culture and stability. Wisdom of the elders is implicit to the genetics. This is a tradeoff with the cost in resources; at some point the cost to keep someone around might exceed the benefit, but from an evolutionary standpoint, the accounting is over a lifetime; in a relatively stable environment, genes that improve longevity and healthspan will be reinforced by the positive feedback loops of culture and nurture and civilization and technology. Menopause is also prevalent in orcas and a handful of other mammals - and older females help rearing and protecting babies, and so forth, with a protoculture providing that feedback loop.
It's not useful. Indeed that's likely why we die of old age - there is no selective pressure to remove harmful mutations that don't reduce your ability to pass on your genes so such harmful mutations just accumulate over many generations. You might have a mutation that will cause your heart to rupture at age 150, but you'll never know it because you'll die of something else first.
It is possible though to selectively breed animals like flies for long lifespan. You wait to see how long one generation lives and cull the descendants of those that died early. It's inefficient but lifespan extensions of 50-60% have been demonstrated. One could imagine through gene editing that a species might be able to reap the benefits without the culling.
https://pubmed.ncbi.nlm.nih.gov/3113991/
Note that evolution isn't about individuals, it's about genes (which we should further note is more than just DNA, but that's a different discussion). If it weren't valuable for humans to live at all without being fertile, then the average age of menopause and the average age of death would likely be much closer together. As it turns out, the human genes that were best able to pass themselves along were those that kept old people around despite being infertile, presumably to the benefit of helping to raise grandchildren, among other things.
Explanation I've heard in popscience books:
Healthy grandparents that are around to support their children and take care of grandchildren increase the fitness of the entire lineage by helping their children have more children and those grandchildren to be healthier/safer.
You can make it broader and simplify:
If you are interacting with a carrier of your genes at all while they still might reproduce, you are having an impact on their fitness and thus evolutionary pressure exists.
We're social animals. Anything that benefits the extended tribal unit is advantageous. Adults beyond child-bearing age contribute significantly to child raising, education, leadership etc of the entire tribe.
Similar effects are seen in other species
https://www.nature.com/articles/s41467-018-05515-8
In (quantitative) genetics literature, heritability is usually defined (simplifying a bit) as the proportion of variance of a trait (lifespan, height, etc), in a population, that can be explained by genetics. The rest, by environmental factors, or error.
If height were a 100% heritability means that all differences in height between individuals would be explainable by genetics.
* correlated with genetics.
What is the question you are asking? What does "useful" mean, in other words? How does it contribute to the reproductive success of the offspring?
the more little old ladies around, the easier it is to raise kids.
It's probably not that useful (evolutionarily) beyond some age. Old people consuming resources without adding anything or holding back societies.
Those old people used to be young and helped pay for their parents and grandparents to live into old age. Part of being young and productive is helping take care of those less able than you, including the elderly.
Unless you're volunteering to work for 40 years then be executed on retirement, I think you should delete that comment and that thought from your mind.
Given the opportunity, one wonders what you'd like to do with 'old people consuming resources without adding anything and supposedly 'holding back societies'. Earlier in the 20th century a significant cohort of intellectuals had decided ideas on this and the earlier generations.
Related somewhat to this: 'The Intellectuals and the Masses' by John Carey makes for truly shocking reading.
Who? https://www.theguardian.com/books/2025/dec/14/john-carey-obi...
It's interesting on my mother's side of the family, most everyone lived well into their 80's and 90's. The execution being for my Mom and her sisters who smoked heavily. Her brothers both died in their 60's but were in the Vietnam war and were definitely exposed to Agent Orange and both had brain cancer. My dad lived until nearly 80 after smoking since he was 12 years old and 2-3 packs per day.
Rats. I have ancestors that died at 97, others at 81. Some even younger. So, no telling.
Still that is above average. I'm average, some ancestors died at 65, others 97, averaging out to around 80.
do you know what they died of? car accidents are probably less heritable, unless they're caused by heritable rash behaviour...
Unfortunately all of my male relatives suffered from hit by a bus itis.
Cancer mostly. Except Mom. She died of heart failure at 97. That's mostly, tired of living so long. She gave up.
Wait. They studied twins, removed accidents etc. But wouldn’t this lead to overestimation of heritability due to shared environment?
FTA: “We use mathematical modeling and analyses of twin cohorts raised together and apart”
So, take one cohort of twins raised together and see how well their life spans correlate.
Take another cohort of twins separated at or near birth and do the same.
Then, do some math magic with both to estimate heritability.
This is incorrect. Twin studies typically compare MZ twin similarity against (same sex, usually) DZ twin similarity. Assuming that there is nothing special about MZs for the trait (e.g. in this case if MZ twins lived longer by virtue of being MZ twins), you can estimate heritability free of shared environments.
The obvious problem you run into with twins raised apart is that there in fact aren't many twins who are raised apart.
From tfa
> We estimated uncorrected heritability (uncorrected for extrinsic mortality) (materials and methods) in three independent ways: (i) MZ twins reared apart (n = 150), (ii) DZ twins reared apart (n = 371), and (iii) MZ versus DZ twins reared together (196 MZ, 325 DZ)
This is from _one_ of the datasets they examined, but there were also two others. n=150 twins reared apart in their small category, or n=520 twins reared apart total is the lower bound of data they had, and even that is not too shabby imo
I don't have an opinion to offer here other than the intrinsic limitations of studies that depend on twins raised apart (that there aren't many of them). It's an unusual instance of a stat where the obvious concern with the premise is underappreciated rather than overappreciated. I've got nothing on MZ/DZ controls.
ah I see, you're commenting on the general difficulty, not necessarily saying this study's results are bad due to the limitation. My apologies, I don't think we disagree.
Yes but those aren’t random samples. Children not raised with their birth parents had different circumstances. As did children who got split up, and families adopting children is also a selection bias.
Yeah I’d take this study with a spoon of salt. As with many human studies, it’s hard to control for all factors.
Seemingly due to reduction in extrinsic factors affecting lifespan.
tangentially, readers may be interested in this paper: https://stateofutopia.com/papers/1/evolving-brains-cull-long...
(you can reproduce its results yourself in a few minutes).
Shorter lifespans drive faster evolution. That was taught in basic biology and we, as a society, know it all too well (infectious diseases).
It’s difficult to square obsession with a long life with a healthy humanity.
Faster evolution does not necessarily translate to better outcomes. Exhibit A: the respiratory capabilities of the brachycephalic pug. Exhibit B: the rabbit fear response – they can get so terrified that they break their own spines trying to escape. Exhibit C: every creature with a hybrid r/K reproductive strategy involving child- or sibling-cannibalism.
> Faster evolution does not necessarily translate to better outcomes.
For individuals, of course yes. But for populations? Also yes, but temporarily as dead-ends (A), or inconsequential stopovers (B), or distasteful (C).
There's also some wisdom in that if you make kids later in life, you pass them the genes to survive (with 50% probability it seems) up to that age.
So if you're in the kind of family that dies of cancer at 30, and make kids at 25, perspectives don't look great.
Now, not to these people shouldn't make kids but perhaps, choose a spouse whose family dies on average at 60+?
Marry "up", not "down" :)