Jennifer Raff has written an informative summary of what you can actually learn about ancient ancestors from those DNA tests.
- Your DNA is not a good snapshot of your whole family tree more than a few generations back. You have many more genealogical ancestors than you have genetic ancestors.
- Any given individual in the past (including all of the ancient people referenced in the Primeval DNA test) is extremely unlikely to have passed along their DNA to anyone, including you.
- Any person in the distant past—be they anonymous peasant or famous monarch—who passed on their DNA into present times might be your ancestor, but he or she will also likely be the ancestor of everyone else in the world. In other words, as geneticist Dr. Adam Rutherford explained in his post on the subject for The Guardian, “we are all special, which means none of us are.”
I like to turn all that around, and consider our descendants. 1) You’re going to be genetically disintegrated and your bits scattered among your descendants, if any; 2) most of us will not have our specific DNA represented in any distant descendants; and 3) if you do leave descendants, your genes will be dispersed among huge numbers of individuals. Sorry, everyone, there really isn’t any such thing as genetic immortality. I blame sex.
Enjoy your life now.
I feel once again vindicated I my choice to leave no genetic trace behind me. I just hope my memetic legacy will be equally insignificant.
It seems we live in a four dimensional (at least!) universe. Why are we not worried about having boundaries in three dimensions but terribly upset about having boundaries in the fourth?
@Ed Seedhouse, I think we are worried about having boundaries in three dimensions. That the whole point of movement – to overcome having boundaries in three dimensions. Wanting to live longer is no different from wanting to fly.
Charly:
Who are “we”? Speak for yourself. Pretty sure I don’t agree with any of that.
I don’t get why you would be worried about not filling all of space. What would there be for you to do, if that were the case and this worry were alleviated? And where would your friends be, if you were everywhere? It doesn’t sound like you could have any, except maybe imaginary ones. Isn’t that worrying?
I guess you may not be worried (about anything) if you were the entire universe (or coextensive with it), but really, it’s just hard to make sense of that idea in the first place, never mind why you might consider it desirable.
Anyway, motion doesn’t actually offer that, so it’s not actually overcome in that way. The phrase “the whole point of movement” is just teleological and silly. Flying doesn’t seem at all the same to me. And a desire to “live longer” is not a desire for immortality.
You touched on some of that in the ‘Chance in evolution’ talk up at Eschaton in Ottawa back in 2012; at least the fact that over time, people end up being the ancestors of pretty much everybody or nobody.
Isn’t this something like, we are all related to Ghengis Khan?
Sure. We’re all homeopathic Ghengis Khans.
I’m apparently drinking a few molecules of the poison that Socrates drank. After having passed through my urinary tract, they’ll be off somewhere else.
Huh. that’s not what I would have expected. I mean, with gene swapping, etc. I wouldn’t expect long chains of DNA to be passed down to genetic descendants past a couple generations, but I would have expected some DNA passed down from lots of different folks in the past. I’m very interested to see the math and to hear how this works in real life.
Is it possible that this used to be true but is no longer? For instance, if populations used to interbreed less often then migration could easily replace loss of population when one group borders on another group whose cultural toolkit becomes less fit over time. You’d end up with the same population in the region, but one group dies out, losing all its ancestors’ contributions to the gene pool, while another group ventures in.
I could imagine, say, the Altai region or the Brazilian Espinhaço Mountains featuring lots of different relatively isolated but not entirely cut off fertile regions that hosted populations that later died out to be quickly replaced by new populations. In this way you could have several hundred years worth of ancestors providing genetic contributions only to a population that eventually goes extinct. With wars and territorial conquest, I can also see any region effectively cleansed (whether or not it was isolated, indeed it seems more likely when highly interconnected as crossroads-lands have long been battlegrounds) by conscious human action, wiping out the genetic contributions of many ancestors. For example, think of all the genetic contributions of European Jews that must have been lost in the Holocaust.
But modern populations seem too interconnected for this, and as many genocides as occurred in the 20th century, I’m hopeful that such slaughters are going to be less and less likely in the future. I wonder if this will continue to be true or not. It just seems counterintuitive to me – which of course doesn’t mean it’s wrong, but does have me searching for an explanation.
@Charly:
Is wanting to live longer no different from wanting to swim, dive, & surface?
CD @9: Someone will probably come along and tell me I’m full of shit, but I think it goes like this:
You’re a result of 23 coin flips by your parents, deciding which chromosome of a pair you get from each parent. So, you could have anywhere from 0 to 23 chromosomes from any grandparent, with the probability peaking between 11 and 13. Go back five generations, and there are (in most cases) fewer chromosomes than ancestors. Some of them would inevitably not have contributed anything to your makeup.
Oops, go back six generations.
That’s why it always annoys me in science fiction where a person has somehow travelled through time to meet their descendants, who are spitting images of them or their spouses.
I would not have wanted to be whoever applied the first round of succussion. On the other hand, by now any one of us could presumably fight off the original GK with a flick of one fingernail.
Huh. You know all those moments you have, when you’re perfectly happy, but then you remember that thing you said 8 years ago and you’re suddenly sad, and very ashamed of yourself?
Think how meaningless all those memories are, when we know our whole lives are just another ephemeral blob of the brief bit of human static.
#11: No, that’s roughly right. It’s more than 23 because of crossovers, but yeah, it’s a ‘chunky’ process with a 50/50 chance any one chunk will be lost or retained. Someone’s ‘chunk’ in the population may eventually go to fixation, but probably not yours.
@PZ:
You’re responding to Rob at #11 and/or #12? I ask because comment #23 hasn’t happened in this thread yet. I’m assuming you just had the number of human chromosomes stuck in your brain.
I’ve been wondering about this as we’ve heard more about people being caught by the police due to others who share their DNA. There are so many amazing stories — like how they got Richard III’s DNA, and then used it to figure out that his great-grandmother’s kid was not from her husband. Or how they’re able to trace most of the population of Ireland back to one prolific king.
Suppose you managed to get a full DNA sequencing from everyone who is currently alive, and fed it all into a ginormous computer. Would you be able to use it to construct a family tree of the entire human race? How far back could you go?
Of course, you’d need actual genealogists to come and try to put names on everyone. What an amazing project, though.
For those who like to see the numbers:
Part 1
The DNA ancestry companies base their model on standard Mendelian autosomal inheritance. That is, each parent gives a randomly shuffled half of their genome to each child. This means that the proportion p of shared genome for people n generations apart is given by
.
For any given index person, n = 0 so p = 1. Yes, you share 100% of your own genome!
By some estimates Genghis Khan is the direct ancestor of 1 in 200 males alive today. But Genghis died in 1227. Using an estimated average human generation time of 30 years, that puts him about 26 generations ago. Which means his direct descendants share
of his autosomal genome. Since the human genome is 3 billion base pairs long, that makes it about 20 base pairs remaining from old Genghis.
But wait! The ancestry company from the article says it can look for descent from the original Israelites. That was 500-600 BC, so roughly 83 generations ago which means there is a
shared genome in descendants (I’m rounding to the nearest power of ten from here on). Eagle-eyed readers might notice that this number is much smaller than the 3 billion of base pairs in your genome. It works out that anyone descended from the ancient Israelites has only one trillionth of a base pair to show for it. Which is a physically meaningless statement. Homeopathic indeed.
But why stop there? The company claims to look for “Ice Age European” DNA. The last Ice Age ended 11,700 years ago or 390 generations ago. Our share from the glacial era is at best
. Which makes our remaining share of DNA many orders of magnitude smaller than Planck length. Now we’re reaching dilutions homeopaths can only dream of.
Okay, but no. If literally all your ancestors were descended from ancient Israelites, your DNA would be 3 billion pairs of Israelite DNA even though you don’t have 10^25 separate ancestors at that age – obviously, as that would exceed the population of the world even today by many orders of magnitude. You can’t in a straightforward way use that formula to computer how many base pairs are going to come from an ancestor sufficiently remote or you’re going to end up with that problem for everyone: we should all have less than a single base pair of DNA total, since we inherited less than 1/3Bth of 1/3Bth of our DNA from each ancestor of sufficient age, yet the population of the earth wasn’t 3 Billion strong.
Any one ancestor of sufficient age, then, likely contributed 0 DNA to you, yet as PZ was saying, there must be many ancestors who contributed a highly disproportionate amount of your DNA since even 1 base pair is “disproportionate” in any case of enough generations distance.
Part 2
Of course the DNA ancestry companies don’t believe in fractional subquantum nucleotide inheritance. What they are really doing is searching a customer’s genome for genetic sequences called STRs that are correlated with ethnic groups.
STRs are simple repeats of DNA, usually only 4-6 letters long. A common one is GATA. These STRs are repeated a number of times in many parts of our genome. What the DNA ancestry companies are doing is choosing gene locations and looking to see how many of these repeats their customers have. So Customer A might have 10 repeats of GATA at locus X. Remember that we have two alleles for each locus, so this means Customer A has 10,10. Customer B could have 10,12. Customer C could have 12,15.
The DNA company checks against a database that says 10 repeats of GATA is associated with, say, the ancient Romans. Which means Customer A is told they have strong Roman inheritance, Customer B is told they have weak Roman inheritance, and Customer C is told they have no evidence of Roman inheritance.
The companies test around 15 STR loci, and for each ethnicity they use four loci (called, conveniently, names like “Jewish I” to “Jewish IV”).
To be fair to them, the companies do say in their materials that the test is not proof of ancestry for several reasons. But that’s in the fine print. Their product pages include testimonials like “BOOM! THERE’S MY NATIVE AMERICAN!!!” or about how the test “validated” a suspected genealogy.
What the companies don’t point out is that the correlations drawn from genome databases come from large studies of modern populations and very small studies of ancient populations, where there are mummified or frozen bodies from which to extract DNA. The reliability of this data is very hard to estimate, especially for the ancient genomics.
They also fail to disclose that STRs have a very high mutation rate. Most coding DNA experiences single-point mutation at the rate of 1 per billion base pairs per gamete. STRs mutate 1,000 to 100,000 times faster. Or to put it another way, those STRs that they are identifying or excluding from people are very poor markers of deep inheritance. 300 generations means that even if you are directly descended from a museum-quality frozen corpse, there is a very good chance that very few of your selected STRs will be the same as your ancestors’. Forensic scientists have even looked into using STR mutations to tell apart direct descent from shared family genes over time frames of 1-2 generations.
Essentially the DNA ancestry companies are in the business of selling romantic fantasies. They test for many markers, but their marketing is especially heavy on groups like Vikings, Romans, original Israelites, and Native Americans. They’re selling a gallon of snake oil with a few teaspoons of real medicine stirred in so they can claim it contains real science.
Crip Dyke–
I think I addressed that point in my next entry. But I’m not trying to say that people actually have fractional base pair inheritances. I’m pointing out that the model they are using is based on flawed assumptions.
Part 3: Grandpa Genghis
The story of Genghis Khan’s genetic heritage is an interesting one. The common version is that Genghis is the direct ancestor of 1 in 200 males alive today. There is a good (albeit uncritical) summary here and a more skeptical take here.
Notice the “male” in sentence? That’s because the researchers used STR analysis of Y chromosomes only. This makes it easier to track lineage because there is almost no recombination of the Y chromosome. (Actually a tiny bit of recombination can happen between X and Y chromosomes right at the tips — only 29 genes have been observed. And we’ve also seen Y chromosomes auto-recombine at palindrome sequences.)
So the Y chromosome evolves by mutation and direct descent, as in an asexual organism. Makes it easier for tracking male lineages.
But what is missing from the story is that Genghis would have shared his Y-chromosome pattern with his three brothers and two half-brothers (and god knows how many undeclared half-brothers, cousins, uncles, and so on). So we don’t actually know how many of the purported descendants of Genghis are really directly lineal to him. It also means that there is a person who we know has more descendants alive today than Genghis. His father, Yesügei. And his grandfather had even more. And so on.
The further back we go, the greater the number of descendants. If we go back far enough, then all of us alive today share DNA from the same ancestors. Which is why non-African humans inherit ~2% of their genes from the Neanderthal genome.
@chrislawson, #22:
You absolutely addressed my concern. If I’d know that concern was going to be addressed, i wouldn’t have said anything. I should have paid more attention to the fact that you’d labeled your earlier entry “Part 1”.
Excellent work. Carry on, and thanks for the education.
@consciousnes razor
I was evidently not talking about “filling all spaces”, that is just silly and I do not know how anyone could interpret my post as such, apart from deliberate strawman to knock down. I meant that whilst our bodies are indeed limited in space, they are capable of moving through said space and we (collective “we” as in the human species) are trying to reach more and more of it with ease. As evidence of this I propose the millenia of inventors who tried to make travel easier and eventualy even flight, submarines and travel to space possible.
People usually do not like being restricted in their movement through space. When there is a barrier, someone somewhere will inevitably try and overcome it. Sure, some will be content with whatever limitations are thrust upon them by fate, but some will not. And we as a species are working on making more and more space accesible to us on both collective and personal level. Many people today would not feel at all content if you took all our modern travel options away from them and took them back to how things were a few thousand years ago, where all the space you could reach was the extent of your daily walk on foot. And it was all made only possible because we were not content with what we have and we tried to expand it.
Analogously, there are multiple barriers that limit the extent of our life, and our species is actively working on trying to knock them down or back. We will never manage to knock them all, just as it seems we will never manage intergalactic flight.
And we, just as we dream about unlimited travel through space, however unrealistic it is, dream about immortality, although it also is unrealistic. As evidence of this I propose all the legends, mythos, sci-fi and fantasy and religions through history, which all are often, if not mostly, in one way or another about entity(ies) who are not bound by ordinary space and/or time limitations and our aspirations at reaching the same.
And yes, of course, individuals vary a lot. But would you accept an argument against the fact that we (collective we) like to produce, make and listen to music for example if a tone-deaf person told you “speak for yourself, to me it is just nonsensical and irksome noise”?
The analogy between a desire to live longer and a desire for flight (and analogy between immortality and, say, unlimited movement through teleportation) is not perfect, I agree, but no analogy ever is. Maybe I could work on it more, I agree too. But why should I, here and now? I am not writing a PhD thesis in philosophy here.
So I hope this comment clears the misunderstanding and is enough. If not, I won’t suffer and I hope neither will you.
This is a very odd way to put this. It sounds like they’re saying parents are unlikely to pass on their DNA even to their own children. The qualifying “in the past” and “ancient people” don’t help since the mechanics of inheritance haven’t changed over the millennia. Either they mean they don’t pass along their entire genome even to their children or, that recognizable parts of people’s genomes do not survive the passage to distant descendants, which judging from the comments, is what they were trying to say.
Can’t comment, too busy crying rn.
Charly:
You were responding to Ed Seedhouse, who was talking about that, because it’s analogous to immortality. Motion isn’t.
Yes, I do accept that there are some (fairly rare) individuals who don’t like music. And much more common are deaf people, who don’t experience it, so we don’t know whether they would like the experience if they had it. What do you expect me to do? I won’t pretend that they don’t exist.
@consciousnes razor.
I disagree with that. Living a life of certain lenght is not analogous to occupying certain space. It might be in physics (I am not a physicist), but not in how time and space are perceived.
During our lives we travel through time from our birth to our death. Our actual existence is always limited to the present moment, which I think does not bother us in general. What bothers us is the extent of the travel we can achieve (and/or the fixed direction – wish for time travel is also present in many cultures). To wish for immortality means to wish to extend that time travel into eternity, not to wish for the present to expand to include every moment of our life ever simultaneously.
Analogously, we travel through space whilst still being confined to our bodies. The size ouf our bodily existence generally does not bother us, but our travel through space is being limited to our ability to traverse said space by means available to us and that does bother us and we, very obviously and evidently, do try to expand that ability to travel.
I expect you to admit that anyones personal experiences and preferences are irrelevant to whether it is possible to use the word “we” when talking about human species or humanity in general. I have given you the clarification you asked for:
Which, btw, is the same usage of “we” that I understood Ed Seedhouse to use. And many others. It is pretty common usage.
What I did not expect you to do was to completely miss the point of that explanation.
Charly @29:
No, in physics we’d talk about your worldline, which only occupies more space if you gain weight ;-)
FWIW, I thought your #3 simply meant that we have striven to increase the spatial and temporal domains accessible to us. Which is true.
Charly @29:
Well, all I can say is that you appear to me to have missed the whole point of my post. Of course we all risk that every time we say anything so I guess I should have no complaint.
That ignores the fact that one individual can be your nth generation ancestor via multiple lines of descent – and in fact, almost certainly will be, if they are your ancestor at all, if you go back far enough. So if the maths used has not corrected for that (I assume that in fact, it has), then it’s faulty.
Charly:
No. When you move around, you still have a spatial boundary, which is similar to the temporal boundary that corresponds to your death (or birth). Your body is in fact limited in both respects, and that doesn’t go away in any sense (perceptually or otherwise) because you can change where your body is located.
The length of your life is a duration. That’s a magnitude which is similar to spatial extent (e.g, length, width, depth, the size of an area or volume, etc.). A specific moment in time is like a specific spatial location, the value of a coordinate if you pick a coordinate system. Those are two different concepts, despite your arguments, and they aren’t (normally) confused in human perception.
The direct ancestor in the male line. If you remove that qualification, I should think it’s a lot more than 1 in 200.
KG@32–
I agree. There are two key problems with the naive application of Mendelian inheritance that I was reducing ad absurdam.
One is that nucleotides are shuffled in a granular way. It’s not just that they are grouped in genes, but also that the genes themselves are recombined in huge chunks rather than one by one. Mendel himself described a Law of Independent Assortment (a Law!), that is, heritable traits are randomly assigned independent of each other like independent coin tosses. We learned that he was completely wrong about this in 1905 when Bateson, Saunders, and Punnet performed an experiment very similar to Mendel’s and discovered linkages…although it still took a long time to work out why. In humans, the average recombination rate is 1.6 per chromosome per generation. To use an analogy of shuffling a deck of cards by cutting it at random points repeatedly (a perfectly acceptable method mathematically but incredibly tedious), each human chromosome is shuffled by only one or two cuts per offspring. That is not much shuffling at all. Most genes are inherited along with a huge chunk of their chromosomal neighbours.
The other problem is, as you say, that populations are finite. We don’t descend from an unbroken line of ancestors with zero shared genome. For most of us, our genetics is a reshuffle of the available alleles in the community we are born into — although there is a constant influx of new genetic material from social mixing leading to interbreeding.
There are basically no isolated human genetic populations. The most isolated humans we know of are the Rapa Nui of Easter Island. For a while it was believed that the Rapa Nui had intermixed with South Americans based on population genetic studies (similar to the techniques used for consumer DNA ancestry testing). The problem was that this data was based on DNA sampling from modern Easter Island inhabitants, whose ancestral isolation was utterly pulverised by European contact in 1722 and the subsequent colonisation/abduction for slavery, and dates of admixture were mathematically deduced from there. An interesting 2017 Cell paper (seriously worth reading — it’s open access!) showed that this hypothesis is probably wrong. Using ribs exhumed by archeologists in the 1980s, the researchers found that there was no evidence of South American admixture in autosomal or mitochondrial DNA from pre-contact individuals (I’m simplifying but that’s the gist). They found instead that the Rapa Nui genome belonged squarely in the Polynesian. They also showed that they could find spurious hits for native South American DNA in European populations (highest in Norway, with up to 5%!) and even more weirdly, a 0% rate for the Botocudo, an actual South American native population — clearly there is a problem with the modelling, especially with small sample sizes. And most consumer DNA tests have a sample size of one.
It is impossible to say how much genetic mixing there was between Easter Island and other Polynesians, but it shows that even the most isolated population on the planet might have interbred with other groups, and even if they were truly isolated, that isolation only lasted around 500 years before the European empires came along and ruined it.
KG@34–
Yep, that data comes from Y chromosome analysis which is why they limit their interpretation to males. But I would expect that if 1 in 200 males is directly descended from Genghis Khan (please note my objections to this above!), then there would also be 1 in 200 females descended from him…but we don’t have a genetic tool to examine this. Is there some reason I am missing to think the world has more female than male descendants of Genghis?
@Ed Seedhouse, fair enough. I do not think I misunderstood your post. I think I understand it correctly but I disagree with it, but since you did not clarify how you think I misunderstand /shrugs.
@consciousness razor – at this point we only can agree to disagree, because I agree with this part, it is actualy one of my points.:
But I disagree very obviously with this:
To me, similar to spatial extent is our subjective perception of “now”, of the momentary perception of flow of time, not the length of our lives.
My analogy:
actual moment in time = location
perceived now = spatial extent (body size)
life duration = spatial reach
where = means analogous, not mathematicaly or otherwise absolutely literally idetnical and interchangeable.
I agree with this:
because when I said they are identical, I meant to say they are analogous. I did not mean it literally and I did not expect it to be taken literally. Poor formulation on my side in that first post, but you did not convince me I am wrong. Maybe we just perceive subjectively these things differently and arguing who is right is therefore like arguing about the true nature of beauty.
Oh, I get what you mean now. Yes, that Y chromosome line will not include male children of female descendants.
I just checked the 2003 Cell paper that estimated 1 in 200 males today are descended directly from Genghis Khan, and yes, it is the raw calculation with no adjustment for descendants who are lost from the Y-line. So, yes indeed, the actual proportion of descendants is probably considerably higher than 1 in 200.
WMDKItty — you OK? Can we help?
chrislawson
I’m okay, it’s just that song gets me every time.
@26 ridana
No, everyone passes DNA to their children. But back from there it’s all about probabilities. Look at what happens to a single gene, A.
Your mother has two copies of A – Am from her mother and Af from her father. Your dad has two copies as well – am from his mother and af from his father. Your mom can only pass along one of her two copies as can your dad. So two of the copies of the A gene that came from your grandparents will be lost. Because your genome is so large and genes get mixed together at each generation you will still, on average inherit 25% of each of your grandparents genes but the farther back you go, the more likely it is that a particular ancestor’s contribution is completely lost.
chris61@42,
This is true, but there are two exceptional cases: your mother’s mother’s, mother’s… mother, because mitochondrial DNA is (almost if not quite always – this is disputed) inherited only from the mother, and (for a man) their father’s father’s father’s… father, because the Y-chromosome is inherited father to son.