This is fascinating, and I've had similar questions about pregnancy and eating seafood. You know, because for most of human history, seafood was a primary food source in many countries. I like the approach of sibling analysis, because it does seem a much for balanced approach. Great stuff here!
There is difference between sea-food and sea-fish. Sea-food is molluscs, octopi and the like. Sea fish is main source of protein if you live on the coast.
Definitely! There are a lot of good things about fish, however, in my experience, lots of new mothers are counselled to avoid fish in general (probably due to the mercury in some fish).
T1D is mostly about the immune system, which attacks the body's own cells, a typical auto-immune disorder. It is difficult but not impossible that it is microbially mediated in some way or another. About half of the genetic signal is in the HLA region alone, an immune system region. https://link.springer.com/article/10.1007/s11892-019-1235-1
I think type 1 diabetes is highly likely to be caused by microbes, for the reasons laid out in the Cochran paper. It strikes young, and up until a century ago, it was 100% fatal within months or years, so it absolutely decimates reproductive success. If it were primarily genetically caused, it would have been selected against too strongly to occur at the frequencies it does (barring a heterozygote advantage). Besides genetics, the other causes of disease are microbes and noninfectious environmental factors. Type 1 diabetes has been documented for at least 2,500 years, so if something in the environment causes it, it can't be super new, and it can't have been very common or else we would have evolved better defenses. Microbially caused diseases can persist for thousands of years, as there is an arms race where the host evolves immune defenses and the microbe evolves to evade these.
That is the logic by which Cochran argues that many common autoimmune disorders have a microbial cause. That paper's been pretty ignored, but the findings have been largely vindicated, as microbes are now implicated in many of the diseases he predicted they would be. The mechanisms aren't fully clear yet, but there are several possible ways that microbes could cause autoimmune disorders. Microbes could evolve to be similar to host cells, causing the immune system to be trained to attack these microbes and then accidentally attack its own cells (https://en.wikipedia.org/wiki/Molecular_mimicry). Certain microbes could cause the immune system to become chronically inflamed and overactive. Exposure to a sufficiently diverse set of microbes might be necessary for the immune system to develop properly, so a lack of certain microbes could cause immune dysfunction (https://en.wikipedia.org/wiki/Hygiene_hypothesis).
If type 1 diabetes is microbially caused, you would expect it to be moderately but not overwhelmingly heritable, with the genes involved most likely those related to the immune system (like the HLA genes). As you note, this is what we see. It will likely be hard to pin down the exact mechanism, as in the case of molecular mimicry, the offending microbe may be long gone by the time the disease develops, and in the other cases, many different microbes (or their absence) may be involved at the same time.
The heritability is about 80%, so the model has to claim it is mediated by modern day microbes. https://www.thelancet.com/journals/landia/article/PIIS2213-8587(24)00068-8/abstract (can someone post a copy?) Since it was described 2500 years ago, clearly, it is not just found in modern day due to modern day microbes. These kinds of evolutionary 'missing factor' arguments are risky. There are many traits with very high fitness cost, e.g. male homosexuality and schizophrenia, that are nevertheless not that uncommon.
Another issue for the microbial mediation model is that shared environment is too low. Presumably family members share microbial environments, which should result in non-genetic family clustering.
Have you checked whether familial patterns are consistent with just genetics, or do some relationships suggest special environments? E.g., adoption (may be impossible due to rarity), father vs. mother.
As especially the ADHD and autism diagnoses come along with advantages for children and (at least in Austria and some of what I know beyond Austria) the diagnostic procedure is ridiculous beyond believe, I would be cautiously inclined to suspect that children's ADHD and autism diagnoses of our days might have much more to do with the mindset of the parents than with the children themselves. More or less being that children whose parents went to get a diagnosis for them will have that diagnosis. Additionally, we have to look at the social environment, which, school!, has become highly dysfunctional itself. It is good to know that two of the early descriptions of autism in children, by Hans Asperger in Nazi Vienna and by Grunya Sukhareva in Soviet Union, were made within highly totalitarian states, which puts "lack of social ability" quite into a different light. So, I think the data on ADHD and autism are probably the least valid point in your data.
Yes, but regarding the weakness of the diagnostic construct and its cheap availability, the differences between siblings might not be such a relevant point. Parents also usually don't send all their kids to soccer. The parental differences I mean would be between parents who have ≥ 1 children diagnosed and parents who have none diagnosed. And the presence or absence of such diagnoses might not contain much information about the actual presence or absence of a pathology.
Psychiatric diagnoses fall in and out of fashion just like bellbottoms. In the past neuroses were in style. Currently, ADHD, autism, and PTSD are in fashion, and are way over-diagnosed.
For example, anyone who has had an unpleasant experience (everyone) and mentions it to a therapist is diagnosed with PTSD. In fact, the majority of people who go through a VERY bad experience experience post-traumatic growth - they become stronger. A minority experience post-traumatic stress disorder.
FAS seems to be caused by serious binge drinking, at the right (wrong?) stage of pregnancy, severe enough for teratogenesis but not severe enough to cause miscarriage or stillbirth. All these factors make FAS much less frequent than in some wild estimates which is why some FAS fans are trying to make it in a spectrum syndrome with much wider diagnosis criteria.
My first comment did not make it, so shorter version: Smoking moms will have been smoking usu. in all their pregnancies, so no big inter-siblings differences to be expected - but bigger inter-families. Similar with other "sins", even getting pain-killers. - Or how was that controlled for (in all studies)?
Presumably they compare siblings whose mom either quite smoking between pregnancies or picked it up. Then you have kids where the mom smoked during pregnancy for one but not the other. (I do wonder if they had data on whether the dad was the same between kids; I'd be a little worried that the reason mom quit smoking was that she remarried a guy who doesn't like it, etc.)
How the data of "mom smoking while pregnant" is collected? Reported by the kids? The mom? Some may first say "no" (fearing a sermon) and "yes" at the 2nd kid (as they first kid seemed ok / or caring less now) - some may have said "yes" first and after they noticed the social taboo lied the next time (Parents also lie about their kids screen-time. I am a father.) - Even those who quit after the first one: smoking has negative health effects for quite some time. - In short: If a study/analysis finds no negative effects for smoking: there is most likely something off with the study/analysis.
My guess is it is recorded at the hospital intake. You are right that it is probably under reported, but then people report that the take illegal drugs on those forms. Who the hell knows.
Thinking of the years though, if the study is using data from the 60’s-80’s people still smoked all the time then. I’ve had doctors that smoked during that period, so I wouldn’t think social desirability bias is messing things up too badly in that time period.
Yeah, still: I was born ca.1970. Expecting mothers were supposed to (mostly) stop drinking and smoking then. Our mom did. Dad did not. Things got more fanatic later, but I assume you can find too few reliable cases of "Same parents had one kid when teetotaler and one 5 years later when living as white trash" or "chain-smoking mom had Pete in 1980, went cold turkey in 1982, had great lungs etc. in 1987 when she got pregnant with Sam" AND not see significant differences between those siblings. - Also: nurses are still smoking more than your average Joe (both pot and cigs) - my impression from living near a housing for young nurses.
Then, the kids are only half-siblings? I guess they do check for full siblinghood (or whatever it's called, when one shares 50% and not just 25% of the DNA)... Full fratricity?
Agreed, and they probably did; I am just saying that if I were going to use the data that is one thing I would double check. If the data set was gathered for some other purpose it might be that they didn't include the father's name, but just a box for "Married: Y/N" kind of deal.
Of course there is always the problem that the man mom is married to didn't provide the baby batter, but presumably that points in the same direction as the other self selection issues, so if it is a problem it probably means you should more likely see an effect.
I recall a UK health study finding that 30-30% of kids dads were not in fact.
There is probably a huge range of variability across societies and times. Plus it is inherently a thought thing to pin down without genetics testing since people are likely to be reticent on the subject (or simply ignorant).
We know of the rates of falsely assigned paternity now from two sources - (A) comparison of genealogy trees with DNA trees, (B) routine analysis in organ donation between family members. In both, the rates are about 1-2%.
Of course, if we turn to "unsettled societies" like Himba pastoralists, we might find something different:
Here are the citations for low non-paternity (from that paper, references)
12.Larmuseau M. H., Matthijs K., Wenseleers T., Cuckolded fathers rare in human populations. Trends Ecol. Evol. 31, 327–329 (2016). [DOI] [PubMed] [Google Scholar]
13.Greeff J. M., Erasmus J. C., Three hundred years of low non-paternity in a human population. Heredity 115, 396–404 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Larmuseau M. H. D., Vanoverbeke J., Van Geystelen A., Defraene G., Vanderheyden N., Matthys K., Wenseleers T., Decorte R., Low historical rates of cuckoldry in a Western European human population traced by Y-chromosome and genealogical data. Proc. Biol. Sci. 280, 20132400 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Larmuseau M. H., Claerhout S., Gruyters L., Nivelle K., Vandenbosch M., Peeters A., van den Berg P., Wenseleers T., Decorte R., Genetic-genealogy approach reveals low rate of extrapair paternity in historical Dutch populations. Am. J. Hum. Biol. 29, e23046 (2017). [DOI] [PubMed] [Google Scholar]
16.Boattini A., Sarno S., Pedrini P., Medoro C., Carta M., Tucci S., Ferri G., Alù M., Luiselli D., Pettener D., Traces of medieval migrations in a socially stratified population from Northern Italy. Evidence from uniparental markers and deep-rooted pedigrees. Heredity 114, 155–162 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Solé-Morata N., Bertranpetit J., Comas D., Calafell F., Y-chromosome diversity in Catalan surname samples: Insights into surname origin and frequency. Eur. J. Hum. Genet. 23, 1549–1557 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Wolf M., Musch J., Enczmann J., Fischer J., Estimating the prevalence of nonpaternity in Germany. Hum. Nat. 23, 208–217 (2012). [DOI] [PubMed] [Google Scholar]
19.King T. E., Jobling M. A., What’s in a name? Y chromosomes, surnames and the genetic genealogy revolution. Trends Genet. 25, 351–360 (2009). [DOI] [PubMed] [Google Scholar]
But please, go ahead, show the link to the 30% misattribution in settled European societies using DNA :D
My own hypothesis is that the astronomically high 30% rate anecdotal notion was from the pre-DNA era based on facial features and the like.
This is fascinating, and I've had similar questions about pregnancy and eating seafood. You know, because for most of human history, seafood was a primary food source in many countries. I like the approach of sibling analysis, because it does seem a much for balanced approach. Great stuff here!
The major problem with fish is mercury, a problem which didn't exist prior to the industrial revolution.
There is difference between sea-food and sea-fish. Sea-food is molluscs, octopi and the like. Sea fish is main source of protein if you live on the coast.
Definitely! There are a lot of good things about fish, however, in my experience, lots of new mothers are counselled to avoid fish in general (probably due to the mercury in some fish).
Particularly in fatty mackarel. What a tasty fish :)
I can actually tentatively believe the smoking protects from type 1 diabetes finding. As Greg Cochran explained in his 2000 paper (https://www.researchgate.net/profile/Paul-Ewald/publication/12424950_Infectious_Causation_of_Disease_An_Evolutionary_Perspective/links/556441a308ae6f4dcc98e649/Infectious-Causation-of-Disease-An-Evolutionary-Perspective.pdf), type 1 diabetes is likely to be microbially caused. More recent studies have found that type 1 diabetes seems to be associated with more Bacteroidetes and less Firmicutes in the gut (https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(19)30412-8/fulltext?uuid=uuid%3A2aef375c-ac3b-445e-a36a-b4a807fee768). Smoking is associated with more Bacteroidetes and less Firmicutes in the gut (https://link.springer.com/article/10.1007/S00203-018-1506-2).
Congrats on the upcoming baby, by the way (I'm assuming that's what you're hinting at).
T1D is mostly about the immune system, which attacks the body's own cells, a typical auto-immune disorder. It is difficult but not impossible that it is microbially mediated in some way or another. About half of the genetic signal is in the HLA region alone, an immune system region. https://link.springer.com/article/10.1007/s11892-019-1235-1
I think type 1 diabetes is highly likely to be caused by microbes, for the reasons laid out in the Cochran paper. It strikes young, and up until a century ago, it was 100% fatal within months or years, so it absolutely decimates reproductive success. If it were primarily genetically caused, it would have been selected against too strongly to occur at the frequencies it does (barring a heterozygote advantage). Besides genetics, the other causes of disease are microbes and noninfectious environmental factors. Type 1 diabetes has been documented for at least 2,500 years, so if something in the environment causes it, it can't be super new, and it can't have been very common or else we would have evolved better defenses. Microbially caused diseases can persist for thousands of years, as there is an arms race where the host evolves immune defenses and the microbe evolves to evade these.
That is the logic by which Cochran argues that many common autoimmune disorders have a microbial cause. That paper's been pretty ignored, but the findings have been largely vindicated, as microbes are now implicated in many of the diseases he predicted they would be. The mechanisms aren't fully clear yet, but there are several possible ways that microbes could cause autoimmune disorders. Microbes could evolve to be similar to host cells, causing the immune system to be trained to attack these microbes and then accidentally attack its own cells (https://en.wikipedia.org/wiki/Molecular_mimicry). Certain microbes could cause the immune system to become chronically inflamed and overactive. Exposure to a sufficiently diverse set of microbes might be necessary for the immune system to develop properly, so a lack of certain microbes could cause immune dysfunction (https://en.wikipedia.org/wiki/Hygiene_hypothesis).
If type 1 diabetes is microbially caused, you would expect it to be moderately but not overwhelmingly heritable, with the genes involved most likely those related to the immune system (like the HLA genes). As you note, this is what we see. It will likely be hard to pin down the exact mechanism, as in the case of molecular mimicry, the offending microbe may be long gone by the time the disease develops, and in the other cases, many different microbes (or their absence) may be involved at the same time.
The heritability is about 80%, so the model has to claim it is mediated by modern day microbes. https://www.thelancet.com/journals/landia/article/PIIS2213-8587(24)00068-8/abstract (can someone post a copy?) Since it was described 2500 years ago, clearly, it is not just found in modern day due to modern day microbes. These kinds of evolutionary 'missing factor' arguments are risky. There are many traits with very high fitness cost, e.g. male homosexuality and schizophrenia, that are nevertheless not that uncommon.
Another issue for the microbial mediation model is that shared environment is too low. Presumably family members share microbial environments, which should result in non-genetic family clustering.
Have you checked whether familial patterns are consistent with just genetics, or do some relationships suggest special environments? E.g., adoption (may be impossible due to rarity), father vs. mother.
As especially the ADHD and autism diagnoses come along with advantages for children and (at least in Austria and some of what I know beyond Austria) the diagnostic procedure is ridiculous beyond believe, I would be cautiously inclined to suspect that children's ADHD and autism diagnoses of our days might have much more to do with the mindset of the parents than with the children themselves. More or less being that children whose parents went to get a diagnosis for them will have that diagnosis. Additionally, we have to look at the social environment, which, school!, has become highly dysfunctional itself. It is good to know that two of the early descriptions of autism in children, by Hans Asperger in Nazi Vienna and by Grunya Sukhareva in Soviet Union, were made within highly totalitarian states, which puts "lack of social ability" quite into a different light. So, I think the data on ADHD and autism are probably the least valid point in your data.
Parental differences cannot explain sibling control method results (except with extreme ad hoc interactions).
Yes, but regarding the weakness of the diagnostic construct and its cheap availability, the differences between siblings might not be such a relevant point. Parents also usually don't send all their kids to soccer. The parental differences I mean would be between parents who have ≥ 1 children diagnosed and parents who have none diagnosed. And the presence or absence of such diagnoses might not contain much information about the actual presence or absence of a pathology.
Asperger's primary objective was probably to protect his patients from being classified as "lebensunwertes leben".
Psychiatric diagnoses fall in and out of fashion just like bellbottoms. In the past neuroses were in style. Currently, ADHD, autism, and PTSD are in fashion, and are way over-diagnosed.
For example, anyone who has had an unpleasant experience (everyone) and mentions it to a therapist is diagnosed with PTSD. In fact, the majority of people who go through a VERY bad experience experience post-traumatic growth - they become stronger. A minority experience post-traumatic stress disorder.
FAS seems to be caused by serious binge drinking, at the right (wrong?) stage of pregnancy, severe enough for teratogenesis but not severe enough to cause miscarriage or stillbirth. All these factors make FAS much less frequent than in some wild estimates which is why some FAS fans are trying to make it in a spectrum syndrome with much wider diagnosis criteria.
My first comment did not make it, so shorter version: Smoking moms will have been smoking usu. in all their pregnancies, so no big inter-siblings differences to be expected - but bigger inter-families. Similar with other "sins", even getting pain-killers. - Or how was that controlled for (in all studies)?
Presumably they compare siblings whose mom either quite smoking between pregnancies or picked it up. Then you have kids where the mom smoked during pregnancy for one but not the other. (I do wonder if they had data on whether the dad was the same between kids; I'd be a little worried that the reason mom quit smoking was that she remarried a guy who doesn't like it, etc.)
Nordic register data can see who the father is.
Good to know! Thanks.
How the data of "mom smoking while pregnant" is collected? Reported by the kids? The mom? Some may first say "no" (fearing a sermon) and "yes" at the 2nd kid (as they first kid seemed ok / or caring less now) - some may have said "yes" first and after they noticed the social taboo lied the next time (Parents also lie about their kids screen-time. I am a father.) - Even those who quit after the first one: smoking has negative health effects for quite some time. - In short: If a study/analysis finds no negative effects for smoking: there is most likely something off with the study/analysis.
My guess is it is recorded at the hospital intake. You are right that it is probably under reported, but then people report that the take illegal drugs on those forms. Who the hell knows.
Thinking of the years though, if the study is using data from the 60’s-80’s people still smoked all the time then. I’ve had doctors that smoked during that period, so I wouldn’t think social desirability bias is messing things up too badly in that time period.
Yeah, still: I was born ca.1970. Expecting mothers were supposed to (mostly) stop drinking and smoking then. Our mom did. Dad did not. Things got more fanatic later, but I assume you can find too few reliable cases of "Same parents had one kid when teetotaler and one 5 years later when living as white trash" or "chain-smoking mom had Pete in 1980, went cold turkey in 1982, had great lungs etc. in 1987 when she got pregnant with Sam" AND not see significant differences between those siblings. - Also: nurses are still smoking more than your average Joe (both pot and cigs) - my impression from living near a housing for young nurses.
Then, the kids are only half-siblings? I guess they do check for full siblinghood (or whatever it's called, when one shares 50% and not just 25% of the DNA)... Full fratricity?
Agreed, and they probably did; I am just saying that if I were going to use the data that is one thing I would double check. If the data set was gathered for some other purpose it might be that they didn't include the father's name, but just a box for "Married: Y/N" kind of deal.
Of course there is always the problem that the man mom is married to didn't provide the baby batter, but presumably that points in the same direction as the other self selection issues, so if it is a problem it probably means you should more likely see an effect.
I remember reading that in settled, peaceful societies, the rate of falsely assigned paternity for legitimate children is like 1-2%.
I recall a UK health study finding that 30-30% of kids dads were not in fact.
There is probably a huge range of variability across societies and times. Plus it is inherently a thought thing to pin down without genetics testing since people are likely to be reticent on the subject (or simply ignorant).
We know of the rates of falsely assigned paternity now from two sources - (A) comparison of genealogy trees with DNA trees, (B) routine analysis in organ donation between family members. In both, the rates are about 1-2%.
Of course, if we turn to "unsettled societies" like Himba pastoralists, we might find something different:
https://pmc.ncbi.nlm.nih.gov/articles/PMC7030936/
Here are the citations for low non-paternity (from that paper, references)
12.Larmuseau M. H., Matthijs K., Wenseleers T., Cuckolded fathers rare in human populations. Trends Ecol. Evol. 31, 327–329 (2016). [DOI] [PubMed] [Google Scholar]
13.Greeff J. M., Erasmus J. C., Three hundred years of low non-paternity in a human population. Heredity 115, 396–404 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Larmuseau M. H. D., Vanoverbeke J., Van Geystelen A., Defraene G., Vanderheyden N., Matthys K., Wenseleers T., Decorte R., Low historical rates of cuckoldry in a Western European human population traced by Y-chromosome and genealogical data. Proc. Biol. Sci. 280, 20132400 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Larmuseau M. H., Claerhout S., Gruyters L., Nivelle K., Vandenbosch M., Peeters A., van den Berg P., Wenseleers T., Decorte R., Genetic-genealogy approach reveals low rate of extrapair paternity in historical Dutch populations. Am. J. Hum. Biol. 29, e23046 (2017). [DOI] [PubMed] [Google Scholar]
16.Boattini A., Sarno S., Pedrini P., Medoro C., Carta M., Tucci S., Ferri G., Alù M., Luiselli D., Pettener D., Traces of medieval migrations in a socially stratified population from Northern Italy. Evidence from uniparental markers and deep-rooted pedigrees. Heredity 114, 155–162 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Solé-Morata N., Bertranpetit J., Comas D., Calafell F., Y-chromosome diversity in Catalan surname samples: Insights into surname origin and frequency. Eur. J. Hum. Genet. 23, 1549–1557 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Wolf M., Musch J., Enczmann J., Fischer J., Estimating the prevalence of nonpaternity in Germany. Hum. Nat. 23, 208–217 (2012). [DOI] [PubMed] [Google Scholar]
19.King T. E., Jobling M. A., What’s in a name? Y chromosomes, surnames and the genetic genealogy revolution. Trends Genet. 25, 351–360 (2009). [DOI] [PubMed] [Google Scholar]
But please, go ahead, show the link to the 30% misattribution in settled European societies using DNA :D
My own hypothesis is that the astronomically high 30% rate anecdotal notion was from the pre-DNA era based on facial features and the like.