I think you’re thinking of “The Tale of the Twins Who Weren’t”, from “Time Enough for Love”, first published in 1973. So far as I know, mitochondrial DNA wasn’t that well-known in that time period, since it had only been discovered in the 1960s.
Much ado about mtDNA. Mitochondrial DNA is inherited matrilineally… but at about 16500base pairs it is a rounding error compared to the nuclear DNA. And it is so highly conserved that it is not really possible for it to be completely unlike one’s sibling.
I heartily endorse lazybratsche’s post about recombination. I attempted to estimate the probability once, and quickly ended up in numbers large enough to talk about things like the number of particles in the universe.
Another thing to consider is the difference between “identity by state” and “identity by descent”. IBS means that two DNA “letters” are the same - you get an A from dad, and an A from mom, and you consider them to be the same thing, because they’re both A. IBD means that you consider the A from dad to be different from the A from mom, because they came from different sources.
Now, when you compare the DNA you got from mom to the DNA you got from dad, they’re going to be identical in nearly every position. The calculations we’ve been doing so far are using the IBD definition of “different”. If you only look at IBS, the numbers change, because you’re only considering those nucleotides that are different between mother and father. Since those nucleotides are more or less evenly distributed across the genome, it doesn’t help a lot, but it does increase greatly the number of recombination events that can be considered the “same”.
And furthermore, mtDNA doesn’t determine any of the traits of the organism as a whole, just of the teeny tiny mitochondria within their cells.
Mostly, but not entirely. An egg has thousands of mitochondria, a sperm has only a few: just enough to power it for its single journey.
However, mtDNA has to “match” with nuclear DNA. That is, mitochondria have only a little DNA left, just the bits to control the processes but not enough to make the precursors (the protiens involved in mitochondrial construction and processes). The precursors (built by nuclear DNA) have to be compatible with the control DNA. If they’re not, the fertilized egg isn’t viable. This may account for a large portion of the roughly 30% of fertilizations that aren’t viable and self-terminate without much in the way of consequences.
However, in rare cases, the nuclear DNA is compatible with the father’s mtDNA, and in some portion of these cases, the father’s mitochondria thrive and the mother’s are selected out. (In some cases, this isn’t complete, yet the fetus is viable but lives life with both mitochondria.) Sometimes, the incompatibility isn’t bad enough, so the fetus is viable, and this leads to a number of different diseases.
Beside the point, but as usual in biology, something that seems simple (we inherit oru mother’s DNA) turns out to be a “mostly true” generalization with fascinating exceptions.
In any case, mtDNA doesn’t contribute much to a person’s noticeable characteristics, other than various diseases when they’re not working right. Practically speaking, I think it can be ignored for the purposes of this question. But if we’re being literal and require 100% DNA divergence, it’s STILL possible. However, I bet the odds against it (with or without mtDNA being relevant) are large enough that it has never happened and will never happen, not even if Homo turns out to be a long-lasting genus (which I doubt, in evolutionary timeframes).
Well, you’re making the blithe assumption that both mom and dad are human. (wink)
Seriously, good point. I had considered the commonality (due to both being human) but hadn’t quite hit this nice state-vs-descent distinction on the head. However, using forensic DNA analysis sidesteps the issue, being IBS. I still think the odds against two sibs being as unrelated as their parents is vanishingly small.
I haven’t thought this through fully, but could they be even less apparently related than their parents? (Ignoring test error, that is.) I think they could be, because for a given gene with two alleles, the parents could be AB AB but the sibs could be AA and BB. But I’m not sure how that would show up on today’s forensic tests.
Except that it’s not completely random. Every base pair (or even every allele) does not have an equal probability of being part of a cross-over. Cross-over happens in sections, and where cross-over occurs, the closer the two base pairs or alleles are to the center point, the more likely they will be part of the recombination. You still get a REALLY large number, but it’s not quite as large as if you assume completely random cross-over events.
Not really. Phenotype is a poor indicator of genetic unrelatedness. The odd-looking girl is still mostly genetically related to her siblings since many genes are not involved in superficial appearance.
ETA: There was a thread addressing this exact topic not that long ago. Don’t have the patience to do the search, but if the OP is interested, I think it was within the last 2 years.
Thanks for clarifying. Guess it would have been more accurate for me to say that in this particular case her two children were genetic brothers and her husband was the genetic father of both. But genetically the woman appeared to be the mother of only one of them. This in spite of the fact that she distinctly remembered carrying the other for nine months and then giving birth.
Doesn’t quite fit the subject but I thought it was an interesting aside.
Y’know, you’d think I’d have learned my lesson by now. I shoudn’t be surprised by anything in biology, any more… But I continually am.
Ain’t that the truth!
I got that bit from a very interesting book, “Power, Sex and Suicide” by Nick Lane. My wife saw me buying that book and gave me a questioning look, and when I said “It’s about mitochondria” she rolled her eyes.