Monday, May 01, 2006

Preliminary Evidence Of Higher Reproduction Rates in Autistic Families

In Autism, Genetics and Evolution and in the comments section of one of Wade's blog entries, Ian Parker makes some good points about the genetics of autism.

Autism is known to be extremely heritable. It is perhaps the most heritable of all psychiatric disorders. It is about as heritable as intelligence and a lot more heritable than personality. Even though autism is not 100% heritable, no behavior or skill is known to be 100% heritable or anything close to that.

Given the low concordance of autism in siblings and fraternal twins compared to the concordance in identical twins, it is likely that autism results from the combination of a number of alleles, 3 to 15 or even more according to researchers. Since the prevalence of ASD is somewhat high (and I'd argue the current prevalence is still an underestimate) it follows that the frequency of most of the alleles that result in autism must be relatively high.

Ian Parker asks why the frequency of these alleles remains high. Even if only 1% of individuals with one of these alleles became autistic, and assuming that autistics procreate much less often than the average person (which I believe is an undisputed assumption), then the frequency of the allele in question should gradually drop until it is close to zero (or whatever can be maintained through new mutations and limited inheritance), unless non-autistic individuals with the allele reproduce more often than the average individual.

The alleles that result in autism must therefore have a slight reproductive advantage by themselves, and be reproductively disadvantageous when combined.

A follow-up question is why alleles that are reproductively advantageous don't have a frequency that gradually increases until it reaches 100%. Obviously, if all autism alleles reached a frequency of 100%, then nearly every person in the planet would be autistic. So there has to be a sort of frequency equilibrium for each autism allele.

It also follows from the above that the frequency of slightly disadvantageous alleles can be high and stable.

I hypothesize that frequency equilibrium applies to alleles associated with most other spectrum disorders in the field of psychiatry. It also contributes significantly to the huge diversity of human behavior, and explains seemingly paradoxical combinations of creativity, intelligence and neurological impairment.

There are ways to test this hypothesis. The alleles associated with autism can be mapped, and reproduction rates of non-autistic individuals with these alleles can be measured.

For the time being, let us consider the reproduction rates in the families of autistics, taking into account that parents of autistics are usually non-autistic.

I will use data from Table 2 of the US Census Report for 2003. This gives 0.89 children per family, 1.82 children per family with at least one child, and 2.46 children per family with at least two children. (I assume the number of children per family with at least 4 children is 4.25, and I have verified that variations in this number don't produce large errors).

Note that these numbers do not vary a lot from state to state, according to a 2000 Census table. It also seems these proportions have not changed much since 1994.

Now consider Ghaziuddin (2005) which examined 58 subjects with AS, and found that they had a total of 64 siblings. This gives a proportion of 2.1 children per family with at least one AS child.

See also Deykin et al (1979) which is a somewhat older paper. It finds a proportion of 3.18 children per family with at least 2 children. Based on the 1980 data found in the 1994 Census report, I estimate the proportion of children in households with at least 2 children at 2.81 in 1980.

Piven et al (1990) results in a proportion of 2.81 children per family, but it is unclear if the sample is complete, and the authors only looked at adult autistics and their siblings.

By themselves, these findings are not statistically significant, but taken together they might be indicative of a trend. Unfortunately, no studies have looked at this question specifically, so there is little to go on. (I have contacted some authors of such studies but they have not been able to provide any additional useful info so far).

There are some additional caveats:


  • Census data looks at all children under 18. Autism studies look at children diagnosed with autism, which rules out very young children, but at the same time could result in a youth bias. Research specifically looking into this question would find control probands matched by age.
  • The autistic child in a family likely reduces the 2nd generation reproduction rates in that family. This is not relevant to the analysis, however, as the number of families with autistic alleles is probably much larger than the number of families that have an autistic child.
  • There could easily be sociological reasons to explain a higher number of children in families where one child has been diagnosed with a psychiatric disorder. But this could also work in the opposite direction.
  • It is not necessary for non-autistics with autistic alleles to have a preference to procreate. It might be sufficient that they form families more often.


Conclusions

Autism alleles are likely advantageous by themselves, but reproductively disadvantageous when combined. This results in a frequency equilibrium for each of the alleles. The frequency of most autism alleles is likely high and relatively stable. Most autism alleles should thus not be described as "disease genes". This has some implications on how researchers should look for autism loci.

There is an open question as to whether autistic families reproduce more often than average. Additionally, research on the reproduction rates of individuals with specific alleles could be carried out.

Finally, this analysis shows that eugenics is pragmatically misguided. If combinations of alleles that have a high likelihood of resulting in autism are identified, and parents abort pregnancies based on this information, it is likely that the frequency of all autism alleles will increase (as parents may simply attempt to procreate again).

37 comments:

  1. Autism spectrum people procreate, there's good evidence that the majority of parents of ASD kids are "BAP" broad autism phenotype.

    Girls who are "naive" which I think probably describes all girls on the spectrum, are more likely to get pregnant out of wedlock. I can't say if they'd be more likely to carry that baby to term.

    My ex's mother was like this. She had a large family, too, most if not all of her kids were on the spectrum or BAP, but I know that there is a genetic thing there... probably something in the PWACR on chromosome 15 because there is seizure activity in the extended family following this bunch who all look odd and act odd --all are descendents of an odd woman who's parents became mormon's and moved to the US from Denmark..

    hmmm.

    Anyway, starting in the 1960's in the West, the likelihood of an odd kind of guy reproducing increased, dramatically. Think of all those hippies and their "free love."

    My ex husband had a very shy great-uncle who stayed on the farm and didn't marry until he was 60 years old. He married a widower who probably proposed to him. They were very happy, from all accounts. I think those guys would have been more likely to marry in the 1960's and afterward, or if not marry, to have sex and end up with a child. There was less pressure in the 1960's for a a man a woman had sex with to also be a good provider. From the 60's onward, especially, men didn't have to pass the test of courting and a marriage ceremony if you know what I mean.

    I think the pro-autism alleles got a push in the 1960's so that they became more common.

    You can "see" autism alleles in the faces of old paintings and old photos, or you can see some of them, because they tend to create people with wide-spaced eyes and large heads....(though that is being disputed) a certain kind of ears, eyes, noses... the features associated with some forms of autism have been around for eons.
    Then you can go get Einstein's DNA... eventually, I think ... or check a hundred random antique DNA samples to look for frequencies.... maybe.

    We could check the Kennedy DNA and see if Rose Marie was likely to be autistic. Maybe she left some hair or something that could be checked.

    ReplyDelete
  2. Girls who are "naive" which I think probably describes all girls on the spectrum, are more likely to get pregnant out of wedlock. I can't say if they'd be more likely to carry that baby to term.

    Hmm. I didn't consider that. Naivete (in females particulary) could be a trait conducive to higher reproduction rates in the modern world.

    Anyway, starting in the 1960's in the West, the likelihood of an odd kind of guy reproducing increased, dramatically. Think of all those hippies and their "free love."

    Yes, I'm inclined to believe current frequencies have not always been the same, even though they were probably still there.

    I think the pro-autism alleles got a push in the 1960's so that they became more common.

    It's possible, but it's hard to tell. (No epidemiological studies in PubMed, with an abstract, before 1977 or so :) The 1990s increase seems to be mostly due to the introduction of the broad autism spectrum concept.

    ReplyDelete
  3. Hi Joseph,

    Some thoughts as follows:

    1. ”A follow-up question is why alleles that are reproductively advantageous don't have a frequency that gradually increases until it reaches 100%.”

    Natural selection does not automatically drive an allele frequency to 100%, especially if there are competing alleles that also convey advantage. The different and competing alleles wil presumably balance out depending on the environmental circumstances, although the balance is potentially precarious and could shift over time.

    2. The Census data used in the analysis does not necessarily provide a valid comparison for the research papers used. But here goes:

    a) As you stated here and I stated in Wade’s blog, children under 18 omits a significant number of older children either at home or temporarily outside of the home that do not yet count as distinct families. This underweights the Census data. All of the papers cited use actual numbers of sibblings, not sibblings under 18.

    b) The Ghaziuddin paper’s ASD sample has a mean age of 13.34. This suggests first that the ASD families are potentially ‘older’ than the average Census family. This matters as the older a family is, the more chance there is of the family size being complete or final. The Census families are ‘works in progress’, and many families will have more children. The proper measure to use for ‘complete’ families is the fertility rate, which in the U.S. case is currently 2.1. In this case, the two numbers are identical. Given ”assuming that autistics procreate much less often than the average person (which I believe is an undisputed assumption)” then this suggests that these families are not doing their part to maintain the allele frequencies.

    c) The Deykin sample appears to select for the existence of sibblings, and the Census data appears to have the same issues of i) under 18, and ii) current child counts vs. fertility rates, which would also apply for families with 2+ children (a count of children today within families with 2+ children is also not a final count). The Deykin families may also be works in progress, but without age data (and I’m too cheap to buy the paper) it is difficult to tell. But using the same 1.82 (<18 children) to 2.1 (fertility) ratio and applying it as a measure of understatement to the Census 2.81 (families with 2+ children) number, the suggested 2+ fertility rate is 3.24 (I’m not sure if this is a valid extrapolation, it probably overstates since 2+ children families are probably closer to completion than average families), which is higher than Deykin. But again, we do not have enough information.

    d) Plevin selected for the presence of adult sibblings. Using the 2.81 Census rate for Deykin (which again understates for 18+, and may not be historically comparable due to the 10 year gap) the ASD family reproduction is below the norm.

    So, in total, Ghaziuddin is probably a wash, and Deykin and Plevin potentially refute your hypothesis. Regardless of any weak directional findings, there are definitely too many assumptions and too little evidence to support your hypothesis or reject it from this data.

    3. From an evolutionary standpoint, current reproductive rates are definitely not indicative of long term historical trends. To be slightly tongue in cheek, the fact that an engineer and systems designer couple in Silicon Valley have above an average income and reproductive ‘fitness’ capacity today does not suggest that this ‘fitness’ also existed 20,000 years ago in a hunter-gatherer society. My blog write-up, if correct, suggests that based on mtDNA and Y-chromosome analysis that the alleles that contribute to autism must be at least 40,000 years old. This is the timespan over which these alleles would have to display an at least average level of ‘fitness’ to survive. To put this in perspective, ‘agriculture’ is only 10,000 years old, and civilization is considerably more recent (some may argue we’re still striving for it).

    4. You state that ” The autistic child in a family likely reduces the 2nd generation reproduction rates in that family. This is not relevant to the analysis, however, as the number of families with autistic alleles is probably much larger than the number of families that have an autistic child.

    This is relevant, in that this post-autistic reduction is the driver of the very downward trend for which the non-autistic persons with the alleles would need to compensate.

    5. You state that “The frequency of most autism alleles is likely high and relatively stable.”. I’d agree and suggest the following:

    If the chance of having a subsequent ASD child is between 2% and 10% (I’ve seen suggestions of 2%, 5% and 10%) then assuming giving birth to one autistic child does not increase the chance of subsequent autistic children (i.e. the underlying odds per child remain the same) then for a couple with the right combination of alleles between them the odds of any child being autistic are somewhere between 2% and 10%. So given current fertility rates, most couples with the potential of producing an ASD offspring will not do so, or even know that this is a possibility. Second, applying these percentages against the suggested ASD 0.60% prevalence rate suggests that somewhere between 6% (10% chance) and 30% (2% chance) of all couples carry a complete set of autism alleles between them. We can’t calculate the individual allele frequencies from this because we do not know which, how many, or which combinations of alleles are required to produce an autistic child. But we can assume that the presence of individual alleles is quite high.

    This also poses issues for calculating the frequency of these alleles over time. Since we don’t know which alleles to count, and need a good sample to estimate frequency within the population once we have identified the alleles, then estimating changes in allele frequency over time is moot.

    6. Camille’s comments – would these be - anecdotal? Or are they observations?

    ReplyDelete
  4. Regarding Camille's comment: "There was less pressure in the 1960's for a a man a woman had sex with to also be a good provider. From the 60's onward, especially, men didn't have to pass the test of courting and a marriage ceremony if you know what I mean."

    I came across an interesting write-up in The Economist recently (old write-up, new search) that demonstrated that the nuclear family of 1950s vintage is actually an abberation in modern (i.e. second millenium) times. I'd link to it, but that would require a) that I search for it again and b) that you have a subscription to follow the link. If there's any interest then let me know and I'll see what I can do.

    ReplyDelete
  5. Reading this reminded me of the alleged gene for homosexuality. How is that (probably more than one) passed on? There is an old joke, "My mother made me a homosexual," which elicits the riposte, "If I buy her the wool will she make one for me as well."

    The serious side to this is that women were supposed to carry the genes for homosexuality. If they were homosexual themselves they had less sexual autonomy than gay men and were more likely to be married and impregnated against their will. If they were straight they could still pass the genes on to their children.

    Another theory I read was that homosexuality could provide benefit to a tribal group in evolutionary terms because when times were hard they would be economically active contributors to the common wealth without placing additional demands on the group for child support.

    I do not know how valid these arguments are. I cannot remember the references. This is just more of my mental baggage. But it suggests to me that common sense arguments about genes and inheritance may miss more subtle actual mechanisms. I suspect that this is the case with autism.

    ReplyDelete
  6. Mike Stanton said:
    "But it suggests to me that common sense arguments about genes and inheritance may miss more subtle actual mechanisms. I suspect that this is the case with autism."

    Funny that, as that was where I was going with the original Autism, Genetics and Evolution post.

    ReplyDelete
  7. Ian, thanks for your long list of comments. I'll address them tomorrow.

    As to homosexuality, there is quite a bit of evidence of a genetic component, although it's not clear that multiple alleles are involved there.

    In any case, homosexuality is probably associated with other traits, such as high sex drive. In other words, non-homosexual individuals with homosexual alleles might reproduce more often than average, compensating for the lack of reproduction of homosexuals with the allele.

    ReplyDelete
  8. Fascinating Ian and Joseph - I've just gone and read Ian's post and seeing as the conversation over there seems to have dried up I thought it better to comment over here.

    As I believe I've said, I am not a scientist, it takes me awhile to process the terminology and thus the implications but I think you're definitely on to something.

    Like most I believe autism has a genetic basis and has the potential to be triggered either through genetic onset or via an environmental trigger. I say 'believe' as according to the dictates of my own way of processing these things I cannot state these things as heavily indicated.

    Forgive me if I'm going over old ground here but its how I can 'catch up' to you clever lot ;o)

    Ian, I agreed with your 'c' scenario - that development of the necessary gene(s) to manifest autism were present before any large-scale divergence took place, however, it could be that we lacked the environmental trigger at that stage - or the necessity depending on one's viewpoint. Autistic people are good with systems, possibly until the movement to a more structured form of society (post agricultural revolution) we had no real need to plan or systemise to an appreciably higher level. This is an appallingly unscientific theory but I think, given the right combination of alleles, we can produce an infinitely wide variety of human.

    Necessity is the mother of invention - can that be true on a genetic level? I have no idea. Is it possible that a societal need for people with increased abilities to be systems-oriented, with a very high level of attention to detail, drove the alleles that already existed in situ to become 'turned on'?

    Could it further be that as our age has become more and more systems driven, the need for hugely competent systems based people has led to a perceived (or actual) increase in the conditions most likely to produce such people - autism?

    I think Camille's points about the broader phenotype are very well made. One wouldn't necessarily have to be on the spectrum to pass on the genetic ability to be autistic. Just being part of this broader phenotype would be more than enough.

    ReplyDelete
  9. Ian wrote:

    Natural selection does not automatically drive an allele frequency to 100%, especially if there are competing alleles that also convey advantage. The different and competing alleles wil presumably balance out depending on the environmental circumstances, although the balance is potentially precarious and could shift over time.

    If a gene allows for alleles A and B, and individuals with A reproduce twice as often, then B is easily driven to extinction. Of course, the reproductive advantage of an allele is not an absolute. It depends on environmental circumstances. What I'm pointing out is that the frequency of alleles in other genes is also a kind of environmental circumstance.


    a) As you stated here and I stated in Wade’s blog, children under 18 omits a significant number of older children either at home or temporarily outside of the home that do not yet count as distinct families. This underweights the Census data. All of the papers cited use actual numbers of sibblings, not sibblings under 18.


    Yes, but except for Piven, the other studies seem to look at young autistics.

    b) The Ghaziuddin paper’s ASD sample has a mean age of 13.34. This suggests first that the ASD families are potentially ‘older’ than the average Census family. This matters as the older a family is, the more chance there is of the family size being complete or final.

    Yes, this is probably because of diagnosis onset. It's hard to tell how much of a difference the offset would make.

    The Census families are ‘works in progress’, and many families will have more children. The proper measure to use for ‘complete’ families is the fertility rate, which in the U.S. case is currently 2.1. In this case, the two numbers are identical.

    I disagree. The fertility rate is the total number of children per woman. It is unlikely that families that participate in autism studies are "complete" either. The best comparison available is to number of children per family with one or two children.

    Given ”assuming that autistics procreate much less often than the average person (which I believe is an undisputed assumption)” then this suggests that these families are not doing their part to maintain the allele frequencies.


    Note also that a factor that works in the opposite direction is higher divorce rates in the parents of autistics.

    c) The Deykin sample appears to select for the existence of sibblings, and the Census data appears to have the same issues of i) under 18, and ii) current child counts vs. fertility rates, which would also apply for families with 2+ children (a count of children today within families with 2+ children is also not a final count). The Deykin families may also be works in progress, but without age data (and I’m too cheap to buy the paper) it is difficult to tell. But using the same 1.82 (<18 children) to 2.1 (fertility) ratio and applying it as a measure of understatement to the Census 2.81 (families with 2+ children) number, the suggested 2+ fertility rate is 3.24 (I’m not sure if this is a valid extrapolation, it probably overstates since 2+ children families are probably closer to completion than average families), which is higher than Deykin. But again, we do not have enough information.

    Families with 2+ children would tend to be more "complete" already, which helps. Your extrapolation probably can't be applied in the manner you have.

    d) Plevin selected for the presence of adult sibblings. Using the 2.81 Census rate for Deykin (which again understates for 18+, and may not be historically comparable due to the 10 year gap) the ASD family reproduction is below the norm.

    Yes, the Piven study is probably not useful. I emailed the author and he didn't really remember if the sample was complete or how it was selected.

    So, in total, Ghaziuddin is probably a wash, and Deykin and Plevin potentially refute your hypothesis.

    I think Ghaziuddin is still of interest, as is Deykin. Piven is probably not useful.

    Regardless of any weak directional findings, there are definitely too many assumptions and too little evidence to support your hypothesis or reject it from this data.

    One problem is that the directional findings are likely to always be weak, considering that only a small portion of individuals with each allele are non-reproducing autistics. I think the data is suggestive, and warrants some mainstream research.

    3. From an evolutionary standpoint, current reproductive rates are definitely not indicative of long term historical trends. To be slightly tongue in cheek, the fact that an engineer and systems designer couple in Silicon Valley have above an average income and reproductive ‘fitness’ capacity today does not suggest that this ‘fitness’ also existed 20,000 years ago in a hunter-gatherer society.

    Sure.

    My blog write-up, if correct, suggests that based on mtDNA and Y-chromosome analysis that the alleles that contribute to autism must be at least 40,000 years old.


    That's also the case with the DRD4-7R allele of ADHD, which probably originated in humans that migrated to Europe. (Some would argue it originated with the Neanderthal).

    4. You state that ”The autistic child in a family likely reduces the 2nd generation reproduction rates in that family. This is not relevant to the analysis, however, as the number of families with autistic alleles is probably much larger than the number of families that have an autistic child." This is relevant, in that this post-autistic reduction is the driver of the very downward trend for which the non-autistic persons with the alleles would need to compensate.

    What I'm saying is that even though the reproduction rates of these families is reduced, this doesn't mean that the reproduction rates of all non-autistic individuals with the alleles is reduced.

    5. You state that “The frequency of most autism alleles is likely high and relatively stable.”. I’d agree and suggest the following:

    If the chance of having a subsequent ASD child is between 2% and 10% (I’ve seen suggestions of 2%, 5% and 10%) then assuming giving birth to one autistic child does not increase the chance of subsequent autistic children (i.e. the underlying odds per child remain the same) then for a couple with the right combination of alleles between them the odds of any child being autistic are somewhere between 2% and 10%. So given current fertility rates, most couples with the potential of producing an ASD offspring will not do so, or even know that this is a possibility. Second, applying these percentages against the suggested ASD 0.60% prevalence rate suggests that somewhere between 6% (10% chance) and 30% (2% chance) of all couples carry a complete set of autism alleles between them. We can’t calculate the individual allele frequencies from this because we do not know which, how many, or which combinations of alleles are required to produce an autistic child. But we can assume that the presence of individual alleles is quite high.


    Absolutely. It's probably rare for someone not to have any autism alleles.

    ReplyDelete
  10. Kev wrote:

    Ian, I agreed with your 'c' scenario - that development of the necessary gene(s) to manifest autism were present before any large-scale divergence took place, however, it could be that we lacked the environmental trigger at that stage - or the necessity depending on one's viewpoint. Autistic people are good with systems, possibly until the movement to a more structured form of society (post agricultural revolution) we had no real need to plan or systemise to an appreciably higher level. This is an appallingly unscientific theory but I think, given the right combination of alleles, we can produce an infinitely wide variety of human.

    It's possible these alleles have always been a bit advantageous by themselves, but not enough to allow them to take off. If they have become more advantageous today due to an environmental cirumstance, then their frequencies would increase until their interaction becomes too much of reproductive disadvantage (i.e. they result in too many non-reproducing individuals).

    Thinking about this model, though, it would seem that one of the alleles would slowly win out in the end. But I'm probably oversimplifying what is likely tremendous complexity of gene interaction.

    Environmental circumstance could be something as benign as a switch into the industrial revolution or the information age; or it could be a new liability, such as mercury in the environment. But if it's something like the latter, it would mean that the frequency of autism alleles should be on the decline currently.

    The prevalence of schizophrenia started to increase with the industrial revolution. It's currently believed to be declining, which is an unsolved puzzle. I'm inclined to believe this is nothing but new recognition of a disorder followed by diagnostic substitution. But an alternative is that environmental change results in the simultaneous increase in the frequency of alleles of multiple genes, which in turn results in an increase in the prevalence of disorders due to allele combinations. Finally, one or more of the alleles win out.

    ReplyDelete
  11. Hi Joseph,

    We’ll probably have to agree to disagree on some things:

    You said: ”Yes, but except for Piven, the other studies seem to look at young autistics.”

    I’d suggest that given the Ghaziuddin paper’s 13.34 avg ASD sample age that these are older families. While the range of first to last child can be huge, I’d suggest that a significant majority of families with a hypothetical 13.34 aged child have finished procreating.

    As for Deykin too, given that families take time to procreate (I believe the current estimate, excluding multiples, is 9 months per child, and we should probably allow time for at least a coffee break before the process resumes), more children suggests older families.

    You said: ”The fertility rate is the total number of children per woman. It is unlikely that families that participate in autism studies are "complete" either. The best comparison available is to number of children per family with one or two children.”

    I’ll have to disagree. Without further evidence to the contrary, any family has ‘stopped’ procreating until the next pregnancy, esp. families with multiple children, as per the logic above. They’re like Schrodinger’s cat – once you look in the box they no longer have an indeterminate state.

    The Census data, on the other hand, is clearly a work in progress (if it weren’t then the population of the U.S. would be collapsing, despite immigration). I’d agree that the extrapolation may overstate, but a) there would still be more children potentially ‘on the way’ to enable a valid comparison of ‘assumed complete’ to ‘statistically complete’ families, and b) we’re still missing those 18+ kids, who would be more likely than average - i.e. a larger number per family are being artificially excluded - in larger families due to family age if nothing else. So in your comparison you’re looking at actual families vs. averaged incomplete families with an artificial exclusion at 18, and suggesting that this supports your hypothesis. I disagree.

    Regardless, extrapolating conclusions from incomplete and potentially unrepresentative data leads to suspect results. You said: ”One problem is that the directional findings are likely to always be weak, considering that only a small portion of individuals with each allele are non-reproducing autistics. I think the data is suggestive, and warrants some mainstream research.” Based on the above, I disagree that the data is suggestive of your hypothesis, but would agree that there’s nothing wrong with further research.

    What I would ask though is, “What is the research seeking to determine?” From a sociological perspective it would be interesting to determine how family dynamics differ between families with autism vs. families without, but if it were determined that families with autism currently procreate more than average then the conclusion would be that that families with autism currently procreate more than average. And?

    From an allele frequency POV the issue is not the behavior of families with autism but the behavior of those who possess the alleles that contribute to autism, of which those within ‘families with autism’ are a subset. I’m assuming that we agree that most people who possess one or more ‘autism’ alleles are totally unaware of this fact. But it is the overall fecundity of those with the alleles (including the ‘unknown’) vs. the population that is at issue. Included within this number with the alleles is the population of autistics, and their assumed reduced rate of procreation rate provides the ‘gap’ that the rest of those with the alleles must overcome. A major issue with any analysis is that we do not know the relationship between those with the alleles and families with autism.

    As a hypothetical example, let’s assume that on average the families with autism each generate one additional child per ASD child to ‘compensate’ for the ASD child. But without knowing the background rate of procreation of those with the alleles, you have no way of determining this. Perhaps the allele rate is 2, vs. a population rate of 2.1. The family with autism has 3 children, the rest of the allele families have 2, and the population has 2.1+ (we have to arrive at 2.1 in total). Overall, those with the allele in total do not ‘keep up’. But you don’t know this because you don’t know who to count. Another hypothesis is that many ASD families practice ‘stoppage’ once they determine that one child is ASD (and this child may not be their ‘last’ child). Meanwhile, those with the allele keep right on procreating and have 2.5 children on average, vs. a population rate of 2.1. The allele is therefore – today- increasing in frequency. Again, without knowing who to count, we have no way of knowing this.

    My evolution argument was an attempt to apply the ‘rules’ of population genetics and evolution to the common wisdom that autism is genetic. The result is a certain set of requirements that the alleles that contribute to autism must theoretically adhere to. The results allow for some alternatives, including a) that those with the alleles procreate enough to overcome the ‘reduced’ procreation of autistics and b) that the alleles in autism combinations may not historically have resulted in the same level of ‘fitness’ impact that they convey today, suggesting the potential of an environmental trigger in at least some cases. Let’s be clear that the suggestion of the potential a more recent environmental trigger in no way suggests that the trigger might or might not be vaccine-related. The ‘answer’, which I believe will ultimately come from extrapolation, will be derived from a definitive understanding of the causes of autism.

    ReplyDelete
  12. Some thoughts on Kev's comments and Joseph's reply:

    Regarding recent changes in allele frequency, this is possible but would be very slow. Given a generational turnover of 30 years, we've only had a maybe 7 cycles (10 to be generous) since the start of the industrial revolution. That might seem like a lot, but in terms of a significant increase in allele frequency world-wide (as distinct from within a small population) that is not a lot. Also, remember that the alleles had to exist at the start of the process, so the assumed change would probably be gradual, not a step change, or a growth from zero to 'x'. As well, the 'competing' alleles would still exist within this period and would be providing a check against unbridled growth.

    Regarding the stability of 'competing' alleles, 'balancing selection' is the term for this, resulting in a balanced polymorphism, i.e. the maintenance of stable frequencies of more than one phenotypic form. As part of this, frequency-dependent selection is when the relative 'fitness' of a phenotype declines if its frequency becomes too high. Think of the 'balance' between systemizers and empathizers as an example.

    Another example of frequency-dependent selection is psychopathy, which is assumed to convey a natural advantage, but (luckily) only as long as the overall number of psychopaths within a population remain low. That's the trouble with population genetics - it is a very amoral process.

    Joseph said: "The prevalence of schizophrenia started to increase with the industrial revolution. It's currently believed to be declining, which is an unsolved puzzle. I'm inclined to believe this is nothing but new recognition of a disorder followed by diagnostic substitution. But an alternative is that environmental change results in the simultaneous increase in the frequency of alleles of multiple genes, which in turn results in an increase in the prevalence of disorders due to allele combinations. Finally, one or more of the alleles win out."

    Another alternative is that the alleles always existed and an environmental change resulted in previously benign alleles becoming deleterious. In this event the deleterious event - schizophrenia - would appear to increase, perhaps significantly, despite the absence of any underlying change in allele frequency. It would also be hereditary. But over time the effects of this change would presumably reduce the frequency of the alleles in question.

    ReplyDelete
  13. Ian wrote:

    We’ll probably have to agree to disagree on some things:

    Sure.

    I’d suggest that a significant majority of families with a hypothetical 13.34 aged child have finished procreating.

    I'd suggest that half of them could easily have more children.

    I’ll have to disagree. Without further evidence to the contrary, any family has ‘stopped’ procreating until the next pregnancy, esp. families with multiple children, as per the logic above. They’re like Schrodinger’s cat – once you look in the box they no longer have an indeterminate state.

    I fail to see the difference with families the Census looks at (except for the diagnosis onset offset).

    Regardless, extrapolating conclusions from incomplete and potentially unrepresentative data leads to suspect results.

    It's all there is to go on. At least it's clear that parents of autistics don't reproduce much less often than normal.

    If I were to design a test, I'd look at a random group of autistic children diagnosed within the last 9 months (to discount the divorce effect). Then I would find age-matched non-autistic controls. I would count the number of siblings. This would need to be repeated in several populations in order to determine a weak directional result.

    But again, even if you find no difference, it's still possible people with the alleles reproduce more often.

    What I would ask though is, “What is the research seeking to determine?” From a sociological perspective it would be interesting to determine how family dynamics differ between families with autism vs. families without, but if it were determined that families with autism currently procreate more than average then the conclusion would be that that families with autism currently procreate more than average. And?

    I think it would be an important part of the genetics vs. environment debate. It would also provide some additional knowledge on how the dynamics of psychiatric disorder prevalence might work. I think it would also discourage eugenics, if the dynamics are what I think they are. (They would see that eugenics could make things "worse").

    From an allele frequency POV the issue is not the behavior of families with autism but the behavior of those who possess the alleles that contribute to autism, of which those within ‘families with autism’ are a subset. I’m assuming that we agree that most people who possess one or more ‘autism’ alleles are totally unaware of this fact. But it is the overall fecundity of those with the alleles (including the ‘unknown’) vs. the population that is at issue. Included within this number with the alleles is the population of autistics, and their assumed reduced rate of procreation rate provides the ‘gap’ that the rest of those with the alleles must overcome. A major issue with any analysis is that we do not know the relationship between those with the alleles and families with autism.

    That's correct. In the absense of knowledge about the non-autistic population with autistic alleles, all we have to go on is undiagnosed parents of autistics (who might be "more autistic" than the average allele carrier). The fact that we're dealing with a spectrum also adds a layer of complexity. It's a simplification to separate autistics and non-autistics.

    Let’s be clear that the suggestion of the potential a more recent environmental trigger in no way suggests that the trigger might or might not be vaccine-related. The ‘answer’, which I believe will ultimately come from extrapolation, will be derived from a definitive understanding of the causes of autism.

    It is clear that genes are involved in most cases of autism, from twin studies. In a minority of cases, mutations are probably involved. Injuries and such probably contribute to some cases. Inherited alleles probably account for most cases, without guaranteeing a psychiatric diagnosis. A new time-dependent brain-damaging environmental trigger seems unlikely to me. Wouldn't it be a strange coincidence that right at the beginning of the information age, "geek" genes all of the sudden become a liability?

    ReplyDelete
  14. Ian wrote:

    Regarding recent changes in allele frequency, this is possible but would be very slow. Given a generational turnover of 30 years, we've only had a maybe 7 cycles (10 to be generous) since the start of the industrial revolution. That might seem like a lot, but in terms of a significant increase in allele frequency world-wide (as distinct from within a small population) that is not a lot. Also, remember that the alleles had to exist at the start of the process, so the assumed change would probably be gradual, not a step change, or a growth from zero to 'x'. As well, the 'competing' alleles would still exist within this period and would be providing a check against unbridled growth.

    That's right. If we're talking about alleles that have at least 20% prevalence in the general population, and the differences in reproduction rate are not that big, it might take 100s of generations to get from 0% to that frequency. The DRD4-7R allele is estimated to have reached 20% prevalence in Europe after 10,000 to 40,000 years.

    I'm inclined to believe that even among cave-men, there were guys who stayed in the cave perseverating about tools and fire.

    Joseph said: "The prevalence of schizophrenia started to increase with the industrial revolution. It's currently believed to be declining, which is an unsolved puzzle. I'm inclined to believe this is nothing but new recognition of a disorder followed by diagnostic substitution. But an alternative is that environmental change results in the simultaneous increase in the frequency of alleles of multiple genes, which in turn results in an increase in the prevalence of disorders due to allele combinations. Finally, one or more of the alleles win out."

    Another alternative is that the alleles always existed and an environmental change resulted in previously benign alleles becoming deleterious. In this event the deleterious event - schizophrenia - would appear to increase, perhaps significantly, despite the absence of any underlying change in allele frequency. It would also be hereditary. But over time the effects of this change would presumably reduce the frequency of the alleles in question.


    Right. That's the first explanation that comes up for something like that. An example of such a theory is here. The trouble with such theories is that they assume that diagnosed rate is the same as true prevalence, and do not consider the possibility of small changes in the frequencies of multiple alleles.

    ReplyDelete
  15. Hi Joseph
    I found this
    http://www.eurekalert.org/pub_releases/2006-03/plos-mth030206.php
    and this
    http://www.futurepundit.com/archives/cat_brain_evolution.html
    This page includes a lot of links to different studies.
    María Luján

    ReplyDelete
  16. Thanks for the links, María.

    in our gene ontology analysis, several other important brain genes also have signals of selection, including the primary inhibitory neurotransmitter GABRA4, an Alzheimer's susceptibility gene PSEN1, and SYT1 in Yoruba; the serotonin transporter SLC6A4 in Europeans and East Asians; and the dystrophin binding gene SNTG1 in all populations.

    GABRA4 and other GABA genes have been associated with autism by one study. The SLC6A4, a serotonin transporter gene, may also be linked to autism. Some studies have found a link, and others have not.

    ReplyDelete
  17. Hi Joseph
    Yes, I noticed these genes. Do you know about the most recent Dr Herbert´s paper -in press-about her search of potential genes in autism?
    María Luján

    ReplyDelete
  18. You mean the 'sky is falling' talk found here?

    http://www.gene-watch.org/genewatch/articles/16-1autism.pdf

    It does not present any new evidence that prevalence numbers represent a real increase. Frankly, this kind of analysis should be considered naive at this point, unless it does so.

    BTW, over the weekend I plan a new post on all the "hidden hordes" that have been found. There are plenty.

    ReplyDelete
  19. Hi joseph

    No, the manuscript I consider is
    Autism and environmental genomics, a review in Neurotoxicology.

    Please let me know if you are interested and I can send it to you.
    M. Luján

    ReplyDelete
  20. I can't say anything about higher reproductive rates but there is no reason that in a pre-industrial, rural world that autistic tendencies were considered abnormal or unusual. Artisans tend to work alone and in small communities and would have been considered highly successful and desirable by their community. Shepherds and farmers require few if any of the social skills that we hold as SO important in our modern society, again ability to deal with routine chores in a consistent manner are desirable traits and therefore would result in success and marriagability by community standards.
    It is our changing standards and our move to an information, urban society that makes it difficult for our "odd" children. It is not that autism is new to the gene pool, it is that we have moved into the whirlpool and they are having a hard time following.

    ReplyDelete
  21. Hawise wrote:

    I can't say anything about higher reproductive rates but there is no reason that in a pre-industrial, rural world that autistic tendencies were considered abnormal or unusual.

    You are correct. This is not really about whether autism is more common today than 1000 years ago. It may or may not be. It's about how the apparently high frequency of the alleles involved remains high and stable despite the reproductive liability that is observed in a portion of the carriers.

    ReplyDelete
  22. Just to be clear, the issue regarding autism, genetics and evolution is not whether the alleles that cause autism have been part of the gene pool for a long time. If the argument is correct then they must have existed for at least 40,000 years.

    Regarding Hawise's comment that "There is no reason that in a pre-industrial, rural world that autistic tendencies were considered abnormal or unusual" and following, this is true, but a couple of thoughts come to mind.

    First, the alleles have to have survived for the 30,000 years prior to the existence of agriculture too. And second, even if some characteristics of autism could potentially be advantageous in an agricultural or artisan environment, I'd suggest that not all autistics would necessarily be capable of demonstrating the required skills or capabilities. If they were then I'd suggest that autism as a diagnosis would not be something that we'd be blogging about.

    Those who did not demonstrate these skills would potentially not be as successful or eligible, although regardless, the alleles did continue to exist, so there must be some explanation for this. Perhaps those with some of the alleles but without autism maintained the allele frequency, perhaps with some help from those autistics who who possessed the capabilities to which you refer. Or perhaps the alleles in the 'required' combinations are not necessarily deleterious without a 'second hit'. Some claim to know the answer to this last point (and I'm sure that they'll provide the answer shortly), but I'd suggest that the question is still unanswered.

    ReplyDelete
  23. Just to be clear, the issue regarding autism, genetics and evolution is not whether the alleles that cause autism have been part of the gene pool for a long time. If the argument is correct then they must have existed for at least 40,000 years.

    Why 40,000 years? I don't see why it couldn't be 10,000 years.

    Regarding Hawise's comment that "There is no reason that in a pre-industrial, rural world that autistic tendencies were considered abnormal or unusual" and following, this is true, but a couple of thoughts come to mind.

    I think what he's saying is that the perception of autism as a pathology is in part cultural. Timimi (2004) explains this well.

    First, the alleles have to have survived for the 30,000 years prior to the existence of agriculture too. And second, even if some characteristics of autism could potentially be advantageous in an agricultural or artisan environment, I'd suggest that not all autistics would necessarily be capable of demonstrating the required skills or capabilities. If they were then I'd suggest that autism as a diagnosis would not be something that we'd be blogging about.

    I didn't follow that. Certainly, abilities such as visio-spatial intelligence, and certain types of problem solving ability, have probably been useful since the start. And this might have been enough to maintain a pretty good allele frequency even before technology.

    Doing some math, let's consider a 4-allele model where each allele has 30% frequency. Without considering assortative mathing, prevalence of autism according to this model should be 8 in 10,000. Now, suppose the frequency of each allele increases to 35%. Then the prevalence would be 15 in 10,000. If the frequency of each allele is only 10% (as might have been a long time ago) then the prevalence would be 1 in 10,000. More importantly, few allele carries would be fully autistic.

    Or perhaps the alleles in the 'required' combinations are not necessarily deleterious without a 'second hit'. Some claim to know the answer to this last point (and I'm sure that they'll provide the answer shortly), but I'd suggest that the question is still unanswered.

    It's obviously unanswered, but it's possible to provide models, and research further.

    ReplyDelete
  24. Joseph said: "Why 40,000 years? I don't see why it couldn't be 10,000 years."

    Apparently you didn't read my post that you cited at the beginning of your post. According to known haplogroup distribution based on mtDNA and Y chromosome analysis, the paths of the people who currently populate Japan and Western Europe diverged 40,000 years ago. Since this period there has not been any significant intermixing of these two populations (post-war U.S. servicemen don't count, unless you want to argue that Japanese autistics have significant and recent American ancestry), or any intermixing of a ‘third’ population with both.

    As such, unless you want to argue that Japanese and Western European autisms are ‘different’, the alleles that contribute to autism in both populations must have existed prior to this split (the option ‘c’ that Kev referred to above), i.e. 40,000 years.

    Joseph said: "I didn't follow that. Certainly, abilities such as visio-spatial intelligence, and certain types of problem solving ability, have probably been useful since the start. And this might have been enough to maintain a pretty good allele frequency even before technology."

    Yes, but I specifically referred to autistics - ” I'd suggest that not all autistics would necessarily be capable of demonstrating the required skills or capabilities” - i.e. those who possess the total required combination of alleles to be autistic. I was not talking about those who possess at least some of the alleles, who constitute a different, related, and much larger group. There is a big difference. If you still agree with what you said up front in your post, "assuming that autistics procreate much less often than the average person (which I believe is an undisputed assumption)", then again, 'autistics' as distinct from those possessing some of the alleles, can be assumed to be not procreating to the same degree as the population. That means those with some of the alleles who are not autistic, i.e. this other group, have to make up the difference. There is no disputing that they – non autistics who possess some of the alleles - might possess the characteristics to which you are referring.

    ReplyDelete
  25. As such, unless you want to argue that Japanese and Western European autisms are ‘different’, the alleles that contribute to autism in both populations must have existed prior to this split (the option ‘c’ that Kev referred to above), i.e. 40,000 years.

    Note that different alleles can result in similar phenotypes. The DRD4-7R allele (of ADHD) is prevalent in Europe, but not in Asia. This doesn't mean ADHD does not exist in Asia at about the same rate it does in Europe. In China, they apparently have the DRD4-2R allele. So we could be talking about something that started around 10,000 years ago in each population, with different alleles. ("Started" as in its frequency was marginal before). But again, I don't think it's necessary to start at zero 10,000 years ago.

    ReplyDelete
  26. That means those with some of the alleles who are not autistic, i.e. this other group, have to make up the difference.

    Yes, I'm not disputing that. But I am pointing out that with multiple alleles, if the frequency of each allele is low, the probability of full autism is low. In my (much) simplified model above, at 10% frequency for each allele, the probabilty of an allele carrier having full autism is 0.1%. I wouldn't think it's that hard to make up for this difference with marginal advantages in some areas.

    ReplyDelete
  27. If you believe that autism has a different genetic basis in Japan vs. Western Europe then the whole autism and evolution argument is moot, as autism could have arisen independently at any time and therefore may also not be universal across populations (allowing for founder effects or natural selection driving potential exceptions – e.g. Quebec Inuits). But you’re the first person I’ve encounterd who has made that argument - or at least, the first I've encountered who has not linked it to mercury.

    As for 0.1% prevalence, the full prevalence (ASD, which is obviously broader than autism but can also be assumed to carry the 'fitness' burden) is 0.6%, isn't it? And it is not the full population that needs to compensate for the 0.1% (or 0.6%) but rather those who carry the alleles, one by one. So in your 30% model, if the prevalence is 30% for allele 'A' and the prevalence is 81 per 10,000, then the prevalence for those with allele 'A' is 2.7%, not 0.8%. 2.7% from a 'fitness perspective is a big number. Modelling out the 0.6% on a four allele model with the same frequency per allele yields a 28% frequency, or a 2.2% ASD prevalence for each allele - also a big number.

    Finally, if you're arguing for an increase in prevalance, what years and what magnitude of increase would you be proposing?

    ReplyDelete
  28. I've also read that autistic kids are not only normal in appearance, but are generally considered attractive. (Google "autism appearance attractive"). This is likely anecdotal, not scientific. But it's another avenue of interest in this particular debate.

    ReplyDelete
  29. If you believe that autism has a different genetic basis in Japan vs. Western Europe then the whole autism and evolution argument is moot, as autism could have arisen independently at any time and therefore may also not be universal across populations (allowing for founder effects or natural selection driving potential exceptions – e.g. Quebec Inuits). But you’re the first person I’ve encounterd who has made that argument - or at least, the first I've encountered who has not linked it to mercury.

    Are you familiar with parallel evolution?

    And yes, I don't think there's a specific set of alleles that causes autism. There are probably multiple (dozens) sets of alleles that increase the probability of an autism diagnosis greatly. (After looking at what genome scans have come up with, I don't think it's going to be as simple as a set of alleles). Speaking of 3 or 4 alleles is really just a simplification to try to understand this. And I don't discount environmental influence altogether. You can have a gene for blue eyes, and a 2% chance that your eyes won't be blue. Twins reared appart have a 70% concordance of intelligence, whereas it's 80% for twins reared together. But this is complex stuff. It's not like we'd all be identical save for vaccines or something simple like that.

    I also think that autism works like mental retardation. There's a spectrum of IQ scores, if you will. About 3% of the population scores below 70. This is about the same in all human populations. And "mental retardation" in this sense is a cultural construct resulting from the normal distribution of IQ scores (not that biological causes don't correlate well with low IQ scores). This can change as time goes by (Flynn Effect). But no matter what you do, there will always be people who score below the 3 percentile.

    ReplyDelete
  30. Joseph said: "I've also read that autistic kids are not only normal in appearance, but are generally considered attractive."

    So you've met my daughter? ;-)

    I've heard this too, from several sources.

    ReplyDelete
  31. Regarding 3 to 4 alleles, I just used your example. I'm pretty sure I'm on record somewhere as agreeing with the view of multiple sets of alleles in multiple combinations.

    Parallel evolution is definitely a possibility, but I'm not sure that it fits the way that the genetic search is being conducted (e.g. an American genome scan would be missing the Japanese alleles), or the attempt by many to link prevalence data in different geographies to shoot down various hypotheses (e.g. the rate is the same in both the U.S. and U.K., despite different levels of thimerosal in each over time).

    ReplyDelete
  32. As for 0.1% prevalence, the full prevalence (ASD, which is obviously broader than autism but can also be assumed to carry the 'fitness' burden) is 0.6%, isn't it? And it is not the full population that needs to compensate for the 0.1% (or 0.6%) but rather those who carry the alleles, one by one. So in your 30% model, if the prevalence is 30% for allele 'A' and the prevalence is 81 per 10,000, then the prevalence for those with allele 'A' is 2.7%, not 0.8%. 2.7% from a 'fitness perspective is a big number. Modelling out the 0.6% on a four allele model with the same frequency per allele yields a 28% frequency, or a 2.2% ASD prevalence for each allele - also a big number.

    When you consider autistic allele non-carriers, you should also take into account the proportion of them who have a reproductive liability, particularly when you're considering the full spectrum. There might be people with schizophrenia, OCD and so on in the non-carrier population.

    It gets more fuzzy as you move up the spectrum. You can pick any dividing line between autistic and non-autistic and we end up with the same problem, i.e. non-autistic carriers must make up for lack of reproduction in autistic carriers. (Prevalence increases, but so does the reproductive potential of autistics).

    Finally, if you're arguing for an increase in prevalance, what years and what magnitude of increase would you be proposing?

    I'm really not arguing this. I'm ambivalent. I do think the industrial revolution and the information age introduce selective pressure relevant to autistic alleles. But on the other hand, their frequencies must have been pretty high already. Plus you'd expect autism to exist just from human diversity (per my "mental retardation" argument).

    ReplyDelete
  33. So you've met my daughter? ;-)

    My son, very handsome little boy as well :)

    ReplyDelete
  34. an American genome scan would be missing the Japanese alleles

    This is what has happened with ADHD, apparently.

    ReplyDelete
  35. "Plus you'd expect autism to exist just from human diversity (per my "mental retardation" argument)."

    I don't necessarily (but am open to the possibility). I'm closer to the 'second hit' camp myself, with the second hit (possibly auto-immune) being against a genetically predisposed individual.

    I pretty much agree with everything else in your 1:33 PM comment, but as above I'd suggest that allele frequencies don't change that quickly (at least not without a deliberate eugenics program, and I'm only aware of one of those of any scale in the 20th century, which thankfully terminated in 1945).

    ReplyDelete
  36. One of the most interesting posts and subsequent discussions I have read in a long time.
    Thank you.

    ReplyDelete