ancestors: computing likely number of relatives from 15 generations ago ..

I was just watching one of those Who Do You Think You Are tv shows where people in the public eye trace their ancestory.

Not for the first time this weeks punter turned out to be a direct decendant of an English monarch: Edward III was a 13 x Great grandparent - which obv takes us all the way to William the Conquerer a couple of generations earlier. So say 15 generations to William I.

Statistically, there are an awful lot of people who should be able to do that - thousands, tens of thousands - but I can’t figure out how to do the maths.

Question: take a single ancestor 15 generations ago; how many living people would share a direct blood line to that person given the line would likely be protected from the ills of life through those generations (f.i. would enjoy a more successful birth survival rate) ?

Cecil’s article might be a good place to start 2, 4, 8, 16 … how can you always have MORE ancestors as you go back in time? - The Straight Dope

That’s brilliant - pretty much nails it. Thanks :slight_smile:

I’ve not seen the program but does ‘direct descendant’ have a specific meaning here? Does the program say that there is an unbroken male line of descent? IOW does DD have the same Y chromosome as Edward III?

Nails the general concept, but as the article suggests with example and rough estimates, there’s no rigorous way to compute the degree of ‘pedigree collapse’. You just know in most modern families there aren’t ‘duplicates’ among the generations which theoretically number 4, 8, 16, perhaps 32 or a few more, and on the other hand when the theoretical number 2^n, where n is the number of generations back, reaches the order of magnitude of the population of your country(ies) of origin at that time, there must be a lot ‘duplicates’ at that point, people who have more than one position in your family tree. And at some point back you and your dog have common ancestors.

But the degree of collapse would seem to depend where you are from and how mixed your ancestry is in fairly recent generations. My ancestry is said to nearly purely Irish. That’s surely not strictly true more than several generations back but could be mainly true much further back. The population of Ireland was less than a million in the middle ages, but 2^20=1mil, 20 generations, ~500-600 yrs, so at that point on the tree my pedigree is probably nearly fully ‘collapsed’. For people of more diverse recent background that boundary condition could be somewhat further back, though since it’s an exponential relationship not many times further back.

But maybe somebody has a reference which deals with the math of the overlaps more rigorously?

Take the generation backwards to the power of two. Google will do the math for you–just enter 2^15 (or 2^whatever) in a search and it’ll give you the answer–in this case 32768. Add one more generation and you are safely at the point where you have more ancestors than you have copies of genes, so you can’t have even a single gene from some of them. Go up to 32 or 33 generations and you have more ancestors (ignoring pedigree collapse, as I see from a quick skim has already been mentioned) than there are base pairs in your genome.

(I also have my ancestory traced back with named individuals to George III, but I don’t know how much I should trust the provenance of the research, and how realistic it is to expect there to be no cuckolding in that lineage.)

Wikipedia’s article on royal descent has a cited quote: “Statistically, most of the inhabitants of Western Europe are probably descended from William the Conqueror; they are equally likely to be descended from the man who groomed his charger.”

Will the C is about 30 generations ago. For 15 I’m getting on the order of 10s of thousands of descendants.

Maybe there’s a whole bunch of issues that aren’t possible to quantify like rates of immigration and emigration, diseases like plagues and ease of locale movement - in heavily agrarian areas surely duplicates were of a higher ratio …

Tricky.

This is a good and highly readable research paper:
Rohde, D. L. T., Olson, S. & Chang, J. T. Nature 431, 562–566 (2004)

The authors start from the simple math of an ideal panmictic unstructured population, then build a much more sophisticated model factoring in population structure, migration, isolated populations etc. They estimate that the MRCA of all humans lived around 3,400 years ago.

It’s interesting that if you go beyond that date, you reach a point of identical ancestry. Every person then alive is either an ancestor of everyone alive today, or an ancestor of no-one. The estimate for this is at 7,400 years ago.

To some extent, these dates are misleading, because genealogical ancestry is not genetic ancestry. These dates have no evolutionary significance, they derive from the fact that as a sexual species, we all have two parents; but from an evolutionary perspective, this doubling of genealogical ancestors is cancelled out by the fact that we only get half of each parent’s DNA.

When we measure genetic common ancestry, we treat each copy of each gene as a separate “individual” evolving independently in a population of genes that are carried in bodies only for convenience! (The assumption of independence is approximately valid because sexual recombination and chromosome assortment “shuffles” the genes around).

Coalescent theory is the study of genetic common ancestry. The expected time to coalescence of a gene in an ideal population turns out to equal to 4 times the effective population size, with time measured in generations. The effective population size for humans is obviously much smaller than the current census population; still, the average human gene has a common ancestor at a ballpark of one million years ago, much longer than the genealogical common ancestor.

No, he’s not a descendant of Edward III from an unbroken male line, not even the Queen of England is that. He’s one of millions of people descended from royalty, including myself for that matter (my son had our genealogy professionally traced and we’re descended from a daughter of Edward I’s. His mother, my ex, turns out to be descended from another daughter of the same king.)

In short nothing special about this at all.

Assume for simplicity that each man has one son and one daughter. He will have four grandchildren (son’s son, son’s daughter, daughter’s son, daughter’s daughter) of whom just one will carry that grandfather’s Y-chromosome.

After 10 generations, there will be 2[sup]10[/sup] = 1024 descendants, but only one will be an agnatic (Y-chromosome) descendant.

After 36 generations, just enough to take us from William the Conqueror to the present, there will be 2[sup]36[/sup] = 69 billion descendants with, again, just one of them carrying William’s Y-chromosome.

Of course that’s too simplified. The upper castes were much more productive than the lower castes, so William has more than 69 billion living descendant slots. How is that possible, with the total world population much less than 69 billion? We’re counting slots. The numbers would balance if he had just 69 million descendants, but appeared 1000 times in the fully expanded pedigree of each of those 69 million.

And indeed, William probably is the ancestor of most Englishmen. But as far is as known, nobody retains his Y-chromosome. Similarly, Charlemagne and Clovis the Great have billions of descendants but, as far as is known, no agnatic descendants. Most Y-chromosomes die out after a few generations.

But the Y-chromosomes that are successful are very successful. There are several hundred thousand Americans surnamed Stewart and, although most are not agnates of the Lord High Stewards of Scotland, many of them are.

Especially astounding, I think: It is now known that over half the present male population of Western Europe are agnates (Y-chromosome descendants) of a King (now known only as “R-L151”) who lived near Bavaria about 3000 BC. He founded a huge royal dynasty (now named for its Bell Beaker pottery style) that rapidly spread out and conquered Britain, France, Spain, etc. You can see his descendant clades here. In a few years time, expect these clades to be mapped to specific migrations and invasions.

There is an interesting review of a related question (what colonial American has the most descendants), and it makes reference to Gary Boyd Roberts’ work on the subject.

A sample snippet:

[QUOTE=Gary Boyd Roberts]
the ‘New England family’ - probably 100 million contemporary Americans descended from 5000 - 8000 Great Migration immigrants of 1620-50. If you have 50 or more sets (husbands and wives) of Great Migration immigrant forebears, you are probably related to almost all of the 100 million, within the range of 8th-12th cousins. The probability of kinship to notables is fully 100 percent, and the number of such ‘household name’ distant kin probably surpasses 500, possibly 1000.
[/quote]

One of the episodes dealt with two examples of (at least partly) Irish ancestry. Genetic testing determined they were distant cousins. Bill O’Reilly was amused; Bill Maher was dismayed…

Thanks. I was wondering if a mistress might have been involved.

FWIW, Edward III does have known living agnates(*), descendants of the Somerset Dukes of Beaufort. AFAIK, no illegitimacies are involved in this descent, at least if you accept the retroactive legitimization of John of Gaunt’s bastards (by Hugh Swynford’s widow) from Pope Boniface IX in 1396.

(* - based on official pedigrees. IIRC, DNA tests have raised the possibility of undocumented cuckolding somewhere in these lines.)

Don’t forget that fairly quickly, the reliability of descent begins to fall apart through false presumptions of parentage. You may be descended not from your grandmother’s husband, but your grandmother’s milkman, and by the time you go back four generations, to 16 linear ancestors, the probability gets high that there were shenanigans, and the family tree going back to George Washington is likely full of kudzu.

Like the adage that every Italian is descended from at least one pope, none of whom were married to anybody’s mother.

Especially as he had no children.