Saturday, June 30, 2007

Why No Built-in Paternity Testing?

My previous attempt to use this blog as a tool for open source assistance in my current writing project was a great success; I have incorporated results from the query in my previous post into the draft of Future Imperfect. So I decided to try again.

Pair mated species, such as humans and many birds, follow a mating pattern of monogamy tempered by adultery. The female pairs with the best male who will pair with her then, given the opportunity, gets pregnant by the (genetically) best male available. Males spend time and effort attempting to engage in extra-pair copulations while preventing their mates from doing so.

A simple solution to the problem faced by males, and one now provided by modern technology, is paternity testing. If a male can tell which of his mate's children he fathered he can decline to help support the others, giving his mate a strong incentive not to cheat on him. My question is why that solution was not long ago implemented by Darwinian evolution. Why do males in such species not have some way of identifying their offspring?

One possible answer is that here, as elsewhere in evolutionary biology, we have an arms race. It is in the interest of males for them to be able to identify children born by their mates to other males but in the interest of females for them not to be able to do so; more precisely, it is in the interest of the female to be able to fool the male into thinking that another male's children are his. It is not clear to me why the females could be expected to win this particular conflict, but perhaps there are reasons that have not occurred to me.

Can anyone point me at relevant literature? Offer plausible answers?

31 comments:

Mike Huben said...

Two non-exclusive approaches that I'm making up off the top of my head:

(1) Males risk "punishing" their own offspring if their discernment is not 100% acurate. As this discernment evolves, we'd expect it to be rather fallible in the beginning: little better than chance. In that case, the false positives would heavily penalize the male's inclusive fitness. Would the savings compensate for this? In species with a timed breeding cycle, there's unlikely to be a chance to make up the loss.

(2) Females prefer gullible makes, and could abandon males that shirk duties to the female's offspring.

Mike Huben

Anonymous said...

I've read that there is a form of male paternity testing: apparently babies, in the first few days, look much more like their fathers than like their mothers.

Anonymous said...

Every organism of a sexually-reproducing species has a father and a mother. If the father is not the male who will be tricked into investing resources to raise the offspring, there are two competing male interests: the interest of the actual father, vs. the interest of the tricked male.

The mother presumably always has an interest in the child receiving investment by males in helping to raise the offspring.

I count two organisms interested in successfully disguising the paternity (both of which are the actual parents) and one interested in not being tricked -- and that one isn't a parent!

Lippard said...

There is definitely an arms race--check out Olivia Judson's excellent book, _Dr. Tatiana's Sex Advice for All Creation_.

Anonymous said...

A related question perhaps, why is it not the norm to have a paternity test done (perhaps using the same blood sample as the PKU screen drawn from the heel of the infant) immediately upon birth?
Clearly it would be in the father's interest, and in many ways, in the child's interest as well (as the child would have a far better idea what they have a 'family history' of from a genetic perspective). It'd be in society's interest also, as it would increase the perceived cost of infidelity and parental fraud.

Anonymous said...

Anonymous says:
"It'd be in society's interest also, as it would increase the perceived cost of infidelity and parental fraud."

How do you know that's in society's interest? Maybe the adulterous women are seeking better mates, thereby producing offspring with desirable genetic characteristics they don't find in their spouses. It's far from clear to me whether reliable detection and prevention of their fraud would have net benefits.

God Fearing Atheist said...

David,

Buss' "Evolutionary Psychology: The New Science of the Mind" textbook discusses ascertainment strategies on pages 197-200. He mentions two, both of which have been brought up already: information about his partner's sexual fidelity (and attempts to maximize it), and 2) perceptions of a child's resemblance to him.

Most of the data he cites comes from Daly, M. & Wilson, M. 1982. Whom are newborn babies said to resemble? Ethology and Sociobiology 3: 69-78.

Richard Y Chappell said...

It may also be worth noting that evolution can only select for an adaptive mutation that occurs in the first place. So the sheer contingent failure of the requisite mutation to ever occur is always a possible explanation for biological sub-optimality.

C.H. said...

Robin Baker's "Sperm Wars" is a great book written on this topic. It's one of the best books I have ever read and it completely changed the way I look at the human mating process. Your questions will be answered very thoroughly if you get a chance to read it.

Anonymous said...

This is a bit of a dilemma.

Men want to know whom they are investing in but they also want to commit cuckoldry on other couples.

If I can identify my own children the guy I betrayed can identify my children too.

Better to keep it secret.

Taylor Conant said...

Hi David,

The reason this paternity-testing didn't evolve naturally was because most male animals in the wild (and most human males prior to somewhat recent times on the history of humanity scale) did not stick around to help father the offspring. Most of the time, with most species, the male impregnates the female(s) and then goes and does his own thing... child-rearing being the exclusive domain of the female.

This practice is somewhat common amongst "primitive" tribes that exist even today, and was certainly true of most human interaction prior to settled populations which began to take on their own peculiar social mores as the groups begot more and more wealth.

Paternity-testing today is a necessity born of child support-legislation. There wouldn't be (as big of) an industry today without males being forced by law in many parts of the country (such as California) to support "their" offspring.

You've heard of the "gold-digger" phenomenon, right? One way a gold-digger gets her gold is to "have" the rich man's baby... it's up to the rich man now to prove the baby often isn't his.

Economic incentives are undeniable in all aspects of life, even areas as coarse as mating!

Tim Worstall said...

" A Natural History of Families" by Scott Forbes has some bits and pieces on this. N9ot as funny as Dr. Tatiana but makes the connections with human behaviour more explicit.

As to why females would not win this race: those who didn't would have no surviving children, so the gene for such incompetence would be bred out perhaps?

Anonymous said...

Men want to know whom they are investing in but they also want to commit cuckoldry on other couples.

If I can identify my own children the guy I betrayed can identify my children too.

Better to keep it secret.


That just means it isn't an evolutionary stable strategy - you would still have the situation that until the genes were in the general population, such a mutation would be beneficial in the short run, and therefore you'd expect to see the strategy to pop up when conditions were right, if such a mutation actually exists.

Anonymous said...

Almost said, but not explicitly said so far: it is in the interests of the children to get all the nurturing and support that they can get. It is never in the child's interest to develop a smell that encourages discrimination against the child.

Anonymous said...

it is in the interests of the children to get all the nurturing and support that they can get. It is never in the child's interest to develop a smell that encourages discrimination against the child.

Sorry, but that's not right. One case where it'd be in the interest of the child to develop a mutation that encouraged mutation against it is if said mutation also encouraged discrimination in favor of the child in some situations even more.

The fact that a father doesn't know with certainty the paternity of offspring means, from an ev. bio. standpoint, that he shouldn't spend quite so many resources on the child as he would if he did know. When making the calculation about whether such a mutation would be useful, this gain for the child should be compared to the loss when the dad turns out not to be the father.

There's the additional point that though the mutation might cause discrimination against the child while it's a child, it will then be an advantage for the child once it's an adult, which also needs to go into the mix.

And, as it has already been noted, at least for humans, it turns out to be useful for a person to actually know who his father is (currently), since that info is useful for worrying about genetically-linked diseases. Now we're something of a special case there, and it might be a temporary state of affairs - eventually doing analysis on our individual DNA may prove better than getting less accurate info from family histories.

William Newman said...

I think both in this case and in the anti-cancer case you discussed a few posts ago, you may be underestimating the practical difficulty of what you want (from the point of view of designing biochemical machinery which can comb through DNA as required). Look at how much trouble it seems to be for eukaryotic cells to tease their chromosomes apart for mitosis, or look at the rate (in bases per second) at which DNA gets transferred from cell to cell when bacteria exchange genes. There's a reasonable amount of selection pressure to make those processes smooth and fast, but still the result isn't all that fast. Now imagine doing detailed comparisons on the hundreds of millions of bases in eukaryotic genomes. Even if you didn't do anything complicated like looking for approximate matches, and even if you somehow didn't have to worry about error rate, it might be impractical to do what you want in any straightforward way using machinery built from proteins in aqueous solution.

As others have pointed out, various partial identity-testing mechanisms do exist in naturally evolved organisms. But perfect brute force DNA identity tests may be an approach better suited for human technology than for naturally evolved cells.

Alex J. said...

Male lions (and domestic cats) kill the cubs when they take over a new pride. This shows that the kill-the-other-male's-children strategy is viable in at least some situations.

Anonymous said...

men are at least trying to spread their attention over a set of options while women are focusing on one event?

GNZ

jimbino said...

Occam's simple answer is that the genes for stupidity and obliviousness are permanently linked to the gene for desperately seeking sex with a woman.

Anonymous said...

Men do marry women who already have kids. Carrying for the stepchild gives him the opportunity to have other kids with her.

Keep it secret advocates:
-Woman (provider<->lover)
-Children
-Husband (When beeing a lover himself)

Anonymous said...

two ideas

- it is not only in the mother's interest, it is in the child's interest as well. Human can evolve so that, as child they can easily conceal paternity.

- human evolution happens mostly as cultural evolution. The father is more interested in spreading its culture, rules, religion, allegiances etc than merely its genes.

Anonymous said...

My question is why that solution was not long ago implemented by Darwinian evolution. Why do males in such species not have some way of identifying their offspring?

A "solution" was "implemented" long ago, it's the basis for infanticide. In a variety of primates, and some birds as well, male conspecifics kill infants they did not wean. Improves their reproductive success, don't you know. The frequency of this behavior varies with species. Accordingly, these males must have the ability to distinguish their offspring from those of other males. I don't know for sure, but my guess is that this identification is accomplished via olfaction, at least in primates, don't know about birds (how do males know which eggs to destroy?).

Heterosexual human males don't follow this "strategy"; killing your date's kids is not generally considered a useful first step to get her to sleep with you.

Anonymous said...

I've read that there is a form of male paternity testing: apparently babies, in the first few days, look much more like their fathers than like their mothers.

How, when he's been dead for so long, is Winston Chruchill still fathering children?

Anonymous said...

As to why females would not win this race: those who didn't would have no surviving children, so the gene for such incompetence would be bred out perhaps?

I think it's worth unpacking this a bit. The assumption that she would have *no* surviving children suggests that the consequences of discovery include her *own* death. And there are certainly human cultures where that is true.

Even in cultures where adultery is not a capital offense, a known female adulterer is generally going to suffer a significant loss of status, and hence reproductive fitness. Discovered *male* adulterers (or victims thereof), however, are much more rarely subject to significant social pressure. Hence, the female has much more at stake.

Anonymous said...

Male lions and domestic cats don't have a sniff test or any other test for whether offspring are theirs, except for one based on recent history. On taking over a pride, they will attempt to kill all the kittens. Presumably these were sired by the previous alpha male. (The females will attempt to stop them, by hiding the kittens or blocking the male from the nest; at least in domestic cats, females may take turns guarding each other's litter.)

I'm not sure, but I suspect that the new alpha male can be quite sure the kittens are not his because he hasn't had a chance to mate with their mothers yet. "Adultery" should be rare among the harem-forming cat species, because females should endeavor to mate only with the best male specimen available, which is the alpha male by definition. If a better male shows up, he demonstrates his superiority by killing the alpha male and taking over...

GNZ, if I understand your question, mammalian females have a whole lot of nutrients and energy invested in each offspring by the time they are born. It limits the number of births that can occur. To maximize reproductive success, the females must follow up by concentrating on raising as many as possible of the offspring they have. (In addition, they can be sure they are raising their own offspring; in many species, the female will learn each newborn's unique scent right after birth, before any mixups can happen.)

OTOH, males have two possible strategies for maximizing reproductive success: work with one or a few females to protect and rear young that they have probably sired, or roam around and sire lots of offspring to lots of females, hoping that some of them will survive. It's quite possible to pursue both strategies at once - for example, King Henry I of England had a wife, two highly valued legitimate children, and is said to have had twenty acknowledged bastards.

Anonymous said...

"...why the females could be expected to win this particular conflict,..."

Possibly, if the female stays within her mate's genetic type, the results of her outside activities would not be as noticed. If this "habit" evolved during tribal/village living over generations, it would make sense to have a permanent union with a male who is gentle and willing to share the efforts of raising her children, and yet strive to mate with (acquire genetic material) a male of greater vigour (bigger herds? better hunter?). If her husband and her lover are of the same genetic type, her chances of getting caught are reduced, and the chances of her offspring being raised to adulthood are increased. Just a thought...
LJ

Anonymous said...

It just occurred to me one problem with the foolproof scheme is too many false negatives.

You can't put the genes in the Y chromosome because you could only tell your sons, nor in the X chromosome because you only tell your daughters.

If you put in some other chromosome, you either need to have it in both pairs to guarantee you can tell, or you need to preferentially choose the gene over others, which generally causes other disasters.

But unless your spouse also has the mutation, your offspring are only going to pass the mutation on half the time, and when they fail to do so, they'll get the answer wrong - i.e. they will think their offspring aren't their children. I think that might be a serious enough defect that such a mutation doesn't survive to make it into the general population.

Anonymous said...

It just occurred to me you might end up with too many false negatives to be able to reliably pass the mutation along.

1) You can't put it in the sex-based genes, because then you could only tell either your sons or your daughters apart.

2) If you put it in some other chromosome, either you need it in both parts of the pair, or you need to preferentially choose it, which would be bad.

But having it in both elements of a chromosome pair only gets you 1 generation of protection, unless the mutation is already widely spread. If your spouse doesn't have it, then your children will only have the mutation in one of their alleles, and only pass along the trait half the time, and therefore misidentify half their children as not their own.

Steve_Roberts said...

It's about the balance between the reliability of the paternity test (PT) and the incidence in the populaation of children-not-fathered-by-the-husband (ICNH). For example if PT is 90% accurate, and ICNH is 10%, then a positive PT is equally likely to be true or false. Acting on a false positive PT is of course much worse than not acting on a true positive PT. This implies that, to be useful, a PT must be very accurate, and ISTM evolution doesn't do stuff like that, evolution starts with something with a slight edge and hones it over generations. A PT with 55% accuracy may be biologically feasible but is operationally useless because it shows more false positives than true positives (on any reasonable assumption about ICNH).

Anonymous said...

Maybe it´s because natural selection is not a function optimization mechanism (it may be, contingently but not by necessity)...as long as paternity identification does not prevent reprotuction of male traits it won't be a signifcant filter, specially in aparently always-growing population..
(guess that's more Wallace's than Darwin's point of view, but it´s darwinians' burden to prove any kind of potential or plenum theory)

dnaattorney said...

paternity testing is only one essential use of DNA testing.