Paternal Age

Decode Genetics has a new report out in Nature that shows how mutations increase with age, by sequencing family trios.  They found that women contribute about 15 de novo mutations, independent of age.

Men contribute more (55 on average) , and the number increases rapidly with age.   They found that the average 20-year old  father passed on 25 mutations, while the average 40-year old passed on 65, an increase of about two mutations per year of paternal age.

The researchers talked about the problems caused by these de novo mutations – things like schizophrenia and autism.  We already knew that such risks increased with paternal age, but this work quantifies the mutations responsible.

Stefánsson opines that the higher mutation rate with older fathers is not that worrisome, since the absolute risk for schizophrenia and/or autism is still small (~2%) and since mutations are our friends: “You could argue what is bad for the next generation is good for the future of our species. “

Well, not for the first time,  Kari Stefánsson is wrong.   If mutations with large effects are  more common with increased paternal age, mutations with small effects must also be more common.  Those small-effect mutations are removed slowly by natural selection, and so they accumulate with time. This eventually results in a population in which everyone has a higher genetic load, not just a few unfortunate kids out of each generation.

What this means is that modest differences in social structure, differences that cause changes in the average paternal age, are likely to have major effects on the mutation rate.  If those differences are maintained over time, say for a few thousand years, you would expect to see significantly different levels of mutational load in different populations. Genetic estimates of split times and such would also be wrong, but that’s a nit.

What do I mean by a modest difference?  Assume that population A has an average paternal age of 25: then the average number of new mutations per generation is 50.  Assume that population B has an average paternal age of 30: then the average number of new mutations per generation is 60, a 20% increase. As Kondrashov put it in his commentary, “It seems that multifactorial disorders that result from impaired brain function, such as autism, schizophrenia, dyslexia and reduced intelligence, are
particularly susceptible to the paternal-age effect. This is consistent with the fact that more genes are expressed in the brain than in any other organ, meaning that the fraction of new
mutations that will affect its functions is the highest. “

So, what is the likely consequence of a higher paternal age?   Population B will eventually be significantly dumber and crazier than population A.

I would guess that paternal age was not too high for most hunter-gatherers. It should have been pretty low in Neanderthals,  since they don’t seem to have lived to be very old, probably as a consequence of their high-risk hunting strategy.  I would also guess that it can sometimes be considerably higher in post-Neolithic societies with greater inequality of wealth, which often leads to unequal reproduction. I would guess that the average paternal age is generally higher in strongly polygamous societies.  But at this point I don’t have the numbers.  I think we can get some of them fairly soon.

Generally, analysts have treated the human mutation rate as if it were a fundamental, unchanging physical quantity, something like the gravitational constant.  It appears that this can only be the case if average paternal age never varies much. Since there may well be other factors that can affect the human mutation rater, even that might not be enough.

About these ads
This entry was posted in Uncategorized. Bookmark the permalink.

43 Responses to Paternal Age

  1. “multifactorial disorders that result from impaired brain function, such as autism, schizophrenia, dyslexia and reduced intelligence, are particularly susceptible to the paternal-age effect.”

    In other words: “multifactorial disorders that result from impaired brain function, such as [insert list of exceptionally imprecisely-defined and heterogeneous rag-bag diagnoses], are particularly susceptible to the paternal-age effect.”

  2. Your example does not appear so problematic if considered from the perspective of time rather than generations. If we fast forward 3000 years, there will have been 120 generations of the population that has average paternal age of 25 years, giving 6000 mutations (assuming they simply accumulate etc). However, we only have 100 generations of the population that has average paternal age of 30 years, also giving 6000 mutations.

    Of course, if you incorporate an increasing rate of mutations with age into the example, the problem returns.

    • gcochran9 says:

      You might be right for neutral mutations, but then they don’t do anything. Deleterious mutations are different. They come into existence at a certain rate per generation, while natural selection is always whittling away at them. Consider the class of mutations that cause a 1% reduction in fitness. If a population started out with zero mutations in that class, the number would immediately start growing. Eventually, though the creation rate (the # of new mutations in that class per generation) would be balanced by the destruction rate (1% of the existing mutations in that class). With a higher mutation rate, you get a higher equilibrium number.

      • observer says:

        It seems to follow from this that in a ‘dysgenic’ environment where impaired brain function does not reduce fitness, Mr. Collins’ point will still hold, and paternal age at conception therefore becomes irrelevant. And if that is true, prospective fathers in the West should not worry about their age, because the real problem is the number of mutations accumulating per year, not the number per generation.

      • gcochran9 says:

        Impaired brain function does reduce fitness, even today – at least schiz, and autism, and serious mental retardation do. If it were truly the case that there was no purifying selection at all today on mutations that impair brain function, civilization would fall in short order, judging from the results of highly relaxed selection in Drosophila etc. And since selection on such traits has certainly relaxed, even though it has not vanished, that might could happen. All the genetic trends are bad: selection against IQ within groups, higher fertility in lower-IQ groups, higher paternal ages, and relaxed selection. I’d love to hear Kondrashov talk about this with his hair down.

  3. Pingback: Beware of the ancient of days! | Gene Expression | Discover Magazine

  4. Sid says:

    Aristotle believed that men should stop having children around age 55 or so. In this matter, he may yet have been astute.

  5. Hank the Plant says:

    Is this a big enough problem to justify having sperm frozen in one’s 20s, for use in one’s 40s?

  6. Pingback: Beware of the ancient of days! | Gene Expression | BizNax

  7. Pingback: Beware of the ancient of days! | Biology News by Biologged

  8. TWs says:

    Didn’t the Saudi king have a big difference in age from when he fathered his oldest to youngest kids?

  9. Mo says:

    I have a layman knowledge of this sort of things. So here is my question? Does this mean that the much lamented teenage pregnancies (which most likely involve young fathers) could have good evolutionary effect even though they lead to bad social outcomes today?

    • The thing is, teen pregnancies are largely the providence of people who already have extra mutations accumulated.

      So, there’s nothing dysgenic about having kids in your teens, but if it happens to be the case that people with damaged genotypes have kids in their teens and (by correlation) thus have more kids than average, the net effect is probably dysgenic.

  10. harpend says:

    I think Jason is right Greg. Consider males a and b.

    Male a has a child at age 25 who in turn has a child at age 25, so the kid has accumulated 58 paternal mutations.

    Male b has a child at age 50 with 75 paternal mutations.

    Not such a huge difference, especially in a social system where, for example, wealth accumulation favors fitness. The 50 year old has been through selection’s sieve, the 25 year old not so much.

    • gcochran9 says:

      The mutation load would, in the long run, be about 30% higher in case b. I think that there is a lot more selection going on between conception and 25 than between 25 and 50 – judging from the mortality curve.

  11. JayMan says:

    There is one complication, at least with connecting advanced paternal age with increased risk of those mental disorders: men who carry genes for those disorders (particularly autism-spectrum disorders) might systematically breed later than men that don’t, creating an association between paternal age and the disorder without genetic load necessarily being involved (consistent with this is that it seems that maternal age is also correlated with autism risk).

    As well, it seems that Northwest Europeans have been breeding late for a long time. They don’t seem the dumber for it. Crazier, maybe, but not dumber…

    • billswift says:

      If you intended your last link as an example of craziness, it worked. It is garbage. Among other problems, autists are just as emotional as anyone else, and just as likely to have their reason over-ridden, they just don’t show their emotions as much.

  12. typal says:

    In population B the mutations in every generation (defined as average age a woman first gives birth in that population) mainly derive from polygynous men. Wouldn’t the men in population B end up being much more affected than the women by deleterious mutations (though selection’s sieve would- via female choice- remove the men with the worst mutations)?

  13. That Guy says:


    What do you think of this study, which seems to argue against teenage paternity in males, towards older (40+) paternity – though it’s not mutations per se:–18256111.html

    CONCLUSIONS: Teenage fathers carry an increased risk of adverse birth outcomes that is independent of maternal confounders, whereas advanced paternal age is not an independent risk factor for adverse birth outcomes.

  14. harpend says:

    Someone, or everyone, check me on this. Let us look at the total number of paternal mutations present in an individual after T years. We assume no selection, etc.

    If g is the male generation time, then in T years there are T/g generations.

    In each generation the mean number of mutations is 20+2(g-20)=2g-20 per generations.

    In T years, that is in T/g generations, The total accumulated number of mutations is

    (T/g) * (2g-20) = 2T-20/g

    In other words total mutation accumulation goes down, not up, with increasing male generation length.

    So what exactly is the problem?

    • Anonymous says:

      (T/g)*(2g-20)=2T-20T/g, which increases with male generation length.

      • harpend says:

        Oops, good thing this wasn’t on my final exam.

      • gcochran9 says:

        20-year old guys pass on 25 mutations, 40 years pass on 65. The accumulation of mutations is slower between birth and 20 than it is between 20 and 40, although the increase is close to linear with age between 20 and 40. Looks to me as if mutations are not accumulating much in pre-pubertal boys, presumably because the stem germinal cells aren’t dividing much yet. So.. older fatherhood increases the per-year mutation rate, because, for those older fathers, a larger fraction of their life before reproduction is post-puberty. So, if average paternal age has, over the long haul, been greater in humans than chimps, you would expect that the number of neutral mutations since the chimp-human split would be somewhat higher in the human branch than the chimp branch – all else equal. Not a lot higher, I should think.

    • albatross says:

      The fraction of the population descended from slow-reproducers also goes down over time, as they’re outbred by the fast-reproducers.

  15. tommy says:


    What do you think of the imprinted brain hypothesis for schizophrenia and autism? The idea is that the two disorders are polar opposites and that each is a result of differences in the optimal reproductive strategies of mothers and fathers. Here’s a Wikipedia article outlining the idea:

  16. j says:

    Older paternal age cannot be dysgenic per se. Men who are healthy and energetic to reproduce (also) when old, have passed the very fine meshed filter of natural aging and age related diseases. They have approved all the tests and earned the diploma of “Fit for life” signed by “Nature”. They may carry a lower number of harmful mutations. The human species is one of the animals that reproduces oldest and we are not being outcompeted by, say, monkeys that reproduce at age five. There is also a clear link between IQ and long life, at an average personal level as well as comparing different ethnic groups.

    • Matt says:

      This might be partially solved by considering Cochran’s comment above “Looks to me as if mutations are not accumulating much in pre-pubertal boys, presumably because the stem germinal cells aren’t dividing much yet.”.

      Humans have an extended development time compared to other animals – otherwise extended brain development couldn’t happen, and we wouldn’t be smart. This may be a factor within and between populations.

      It might be that it is the time from sexual maturity to reproduction that matters more than age (although, yes, I don’t think this is actually decreased in humans).

      As a side note, I think there is a fairly robust finding that richer people tend to sexually mature earlier than poorer people, in terms of age of first menarche amongst women….

  17. Anonymous says:

    On the other hand, transmitted telomere lengths increase with paternal and paternal grandfathers age at conception:

  18. panjoomby says:

    shouldn’t this mean that 2nd, 3rd, 4th born children (etc.) have steadily higher incidence rates of autism (tho never very high), b/c the dad is older each time?

    • JayMan says:

      An excellent idea! We could compare the prevalence of autism within a sibship and see if it increases from earlier-borns to later-borns. If parental age truly matters in and of itself, this would show such.

  19. Peter Frost says:

    When I was born, my mother was 40 and my father 46. There is in fact a recurring pattern of late childbearing in my family tree, as there is in many middle-class English families, yet there are no cases of mental illness (unless one considers freemasonry to be a mental illness).

    Late childbearing has been prevalent in Western Europe since at least the 1300s (cf. Hajnal line). If there are negative consequences, we should see some evidence by now.

  20. Steve Sailer says:

    “Well, not for the first time, Kari Stefánsson is wrong.”

    Is he the the guy who announced James D. Watson was 16% black and 9% Asian?

  21. Pingback: linkfest – 08/26/12 « hbd* chick

  22. Laban says:

    From Robert Graves’ “I, Claudius”:

    I quoted the Greek epigram: The old man weds, for Nature’s rule he scorns – ‘Father a weakly stock, or else wear horns’.

  23. Laban says:

    Correction – it’s from the sequel “Claudius The God”

  24. That Guy says:

    @Greg, @Henry,

    Any thoughts on this NY Times article – basically inflammation during pregnancy is the cause of Autism:

  25. JayMan says:

    Well, James Watson now seems to be making an argument along these lines, arguing that people start having children at age 15 – or at least preserve their gametes for future use – to reduce the occurrence of mental illness in the population.

    I gather that this is much more of an issue for men than it is for women.

    DNA pioneer James Watson’s genetic prescription: Have kids early – NBC

  26. gwern says:

    I’d be interested to see a discussion of this paper: “Paternal Age at Birth of First Child and Risk of Schizophrenia” Petersen et al 2011

    Objective: Greater paternal age is associated with increased risk of schizophrenia, and it has been hypothesized that de novo mutations in paternal germ cells are responsible for this association. An alternative hypothesis is that selection into late fatherhood accompanies a predisposition to schizophrenia. However, direct evidence of either hypothesis is lacking. If de novo mutations are responsible, greater paternal age at conception should increase the risk of schizophrenia. Conversely, if selection into late fatherhood is responsible for the association, greater age at which the father had his first child should increase the risk of schizophrenia. The authors aimed to distinguish between these two measures of paternal age.

    Method: A total of 2.2 million people born in Denmark between 1955 and 1992 were followed up until first diagnosis with schizophrenia. Incidence rate ratios were estimated in a Cox regression.

    Results: Among second- or later-born children, greater paternal age increased the risk of schizophrenia. However, when paternal age at the time of the father’s first child was accounted for, the risk of schizophrenia did not depend on paternal age at the birth of later children. In contrast, the risk of schizophrenia increased significantly with increasing paternal age at the time of the father’s first child.

    Conclusions: Factors related to greater paternal age when the father’s first child was born, and not the father’s age at conception of later children, are responsible for the association between paternal age and the risk of schizophrenia. These findings do not support the de novo mutation hypothesis.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s