HYPOTHETICAL: “Jack” and “Jill” were born to an unwed mother who didn’t learn her lesson the first time. They were born (and put up for adoption) about 18 months apart. Jack was 25 and Jill was 23 when they met at the coffee shop, and developed a relationship, not knowing they were related. They have sex and she gets pregnant. Assume that they are both in perfect health.
What are the real-world odds that their baby is going to have significant birth defects?
This may be of interest:
http://www.jpeds.com/article/S0022-3476(82)80347-8/abstract
Small sample size though.
Are Jack and Jill siblings or half-siblings? Do they have the same father as well as the same mother? AFAIK that would make a big difference in their degree of genetic similarity and consequently in the chances of birth defects for their offspring.
I can’t do a search here for some reason, but I know there’s a Cecil column talking about this and, as I recall, he says the chances are quite low – nothing like those given in the small sample study cited.
I think the column you’re talking about is the one on cousin marriage.
First cousins marrying, as Cecil noted, is not substantially more risky in terms of genetic defects than more genetically distant matings. Sibling marriage, on the other hand, is:
Even so, that’s still less than the high rate of birth defects exhibited in Malthus’s study’s small sample.
It is curious that no such studies have been conducted.
No experimentation need be done, just observation through survey.
I know that incest is one of our more powerful social taboos, but I find it odd that nothing more than these small studies have been done.
Given the scenario in the OP, there isn’t even a taboo factor. The couple don’t know that they are siblings. I suppose it has much to do with the occurrence of the same negative recessive trait. But, I am not a geneticist. Others here probably could speak to this issue better than I.
Lesseenow, full siblings have all four grandparents in common, while half-siblings have only two grandparents in common.
First cousins also have two grandparents in common. So does that mean that half-sibling pairings are no more genetically close, and consequently no more genetically risky for offspring, than first-cousin pairings?
Or is having one shared parent automatically more risky in that regard? As you can probably tell, IANA geneticist.
I’ll take the opposite approach from Malthus’ cite, and estimate it based on genetics. Really, though, to get a good answer to the question you’d need lots of good empirical data, which probably won’t happen.
Anyways, to keep my estimate simple, I’m only going to consider autosomal recessive mutations. You’ve got two copies of each gene, so if one is simply “broken”, the other can often suffice. For each recessive mutation carried by the grandparents, there is a 1/4 chance that it will be carried by both parents, and ultimately a 1/16 chance that it will become homozygous (and have some mutant phenotype) for the child in question.
We can ignore spontaneous recessive mutations in the parents, since those won’t be homozygous in the child. And I’m going to ignore spontaneous parental dominant mutations, since those are much less common.
So, what are the odds that the grandparents are carriers of recessive mutations? Pretty high, actually. In each generation, there are a hundred mutations new mutations (give or take – this recent measurement says 60, this one says 100-200). Most of those mutations will be harmless, though it’s hard to say how many. The standard WAG by geneticists is that ~5 of those new mutations will be harmful recessive alleles. And that’s for each grandparent.
In total, the child has ~10 1/16 chances to inherit something bad, which works out to be a 47% chance of genetic defect. (Which is remarkably close to the 12/21 figure cited by Malthus.)
But so far I’ve neglected to consider what different sorts of mutations might do. A grandparental mutation in really fundamental biochemical machinery will kill the parent’s gametes (which have to function with only one copy of a gene). That would result in infertility for the parents. Similarly, there are lots of essential development-related genes. Break those, and the child will die early in development. If that’s before it implants in the uterus, the parents will just experience more infertility. Or, if development can proceed a little while longer, they could have a high rate of miscarriages. So, to end up with a child with a congenital defect, they have to express a mutation in a gene that’s important enough to cause problems when it’s broken, but not absolutely critical for eventual birth.
I have no way to estimate the fraction of possible mutations that are important but not lethal. It’s probably a very small fraction of those hundred new mutations per person, though, maybe just one or two on average. So, my ultimate WAG for the chance of birth defects? I’ll just bound it between 1/16 or 6% (one of the grandparents almost certainly carries something bad) and ~50% I figured above, assuming several disease alleles per grandparent.
It may be tricky to analyze data from existing cases because, er, how do I put this…sometimes incest runs in the family.
No, full siblings share the same biological parents. Half-siblings share one biological parent.
:smack: Doh, just read what i wrote, yes, you are correct.
If Bob mates with his cousin Alice, he has 1/4 chance of inheriting an allele that Alice inherits from a given common grandparent. But if Alice is Bob’s half-sibling the chance rises to 1/2 – the same allele was necessarily passed to the common parent.
File this under "Threads that would be even more interesting if they were tagged with ‘Need answer quick!’ "
Cool, my 6%-50% estimate is still in the same ballpark as the 7%-31% figure from actual doctors dealing with real people!
But the distinction you made is still a key point: Is sharing one parent essentially the same thing as sharing two grandparents, in terms of the likelihood of genetic defect?
Or does the closeness of the common ancestor (parent vs. grandparent) matter even if the total number of common grandparents is the same?
I’m going to have to reread lazyb’s post more closely to see if the answer is in there, but maybe somebody with better genetics chops will answer the question more simply in the meantime.
[On preview: yup, that happened. Thanks septimus!]
This article talks about a recent study of a large sample of human genomes that came up with a surprisingly high occurrence of defective genes. Hard data is out there, and more is probably on the way. Not all these genes would lead to serious diseases if paired, but still, it doesn’t look good.
We don’t even need to do that. Just ask any dog breeder what happens with that kind of inbreeding. We have centuries of data for that.
Kimstu’s cite says that most children of incest are given up for adoption. As a result, I don’t think a simple survey would work because those given up for adoption would not know that they were children of incest. Perhaps the adoption agencies themselves could provide data, but I’m not sure how much they follow the health history of the kids, or whether they’d be allowed to disclose information to researchers under any conditions.
I wouldn’t be surprised if the actual incidence of abnormalities isn’t significantly higher than the studies since they don’t seem to take into account miscarriages caused by these abnormalities. (Unless I missed it, which is always possible).
Dog breeding, and all domestic breeding, is a little different, because inbred strains are culled. Unfit offspring are killed, or at least not allowed to breed. This has the effect of removing harmful alleles. On the other hand, they’re usually inbred for many generations, unlike the human situation in the OP.
This is an excellent point.