Biocodicology - A Deeper Dive

[Mark Knowles]

One could find the closest living relatives on the database. This will also give an indication of region of origin. One can reconstruct the family tree.

A person has 2 parents, 4 grandparents, 8 grand-grandparents, ectc. Guessing 25 years average between generations, 600 years is 24 generations.  Thus a person living today has about 2^24 = ~4'000'000 potential ancestors who were alive in ~1430.

The number of actual distinct ancestors will be much smaller, of course, because there is a large chance of one marrying one's cousin N-removed.  So your ancestor along a specific 24-step father/mother chain may be the same person as that along another 24-step father/mother chain.  But even so the number of actual distinct ancestors must typically be in the tens of thousands.  

The curious thing is that, the way recombination works, each person's DNA has a relatively small number N(g) of segments, such that each segment is inherited whole from an ancestor g generations back.  While the number of potential ancestors grows exponentially with g, the number of segments N(g) grows linearly with g.  The consequence is that, for g large enough, one's DNA carries pieces from only a relatively small subset of one's ancestors g generations back.  

So a descendant of Mr. X, the VMS author, today may have zero (not just a small amount) of his DNA.

One can reason in reverse, and conclude that a person alive in the 1400s has ~4'000'000 potential descendants today.  Again, many of those descendants will be the same person.  But even so the number may be tens of thousands.  And each of them has tens of thousands ancestors who were alive in the 1400s.  

Thus, even if we find that José Schmidt in Uruguay has a bit of DNA that matches a bit of DNA extracted from the book, how can we determine which of José's 55'672 actual 15th century ancestors is the one who had that DNA?  Can we even get the names -- forget the DNA -- of a few thousands of those ancestors?

This is only a small sample of the complications of DNA mixing and propagation. I have a very limited understanding of it myself -- only enough to know that it is quite complicated, and that details like those, by themselves, may make the idea unfeasible.  No matter how sensitive the technology becomes.

All the best, --stolfi

Genetic comparison leads to the identification of multiple descendents not only one.

If you read up on the examples I mentioned such as the Golden State Killer where multiple relatives were identified and so a common family tree produced and the case of the identification of Richard III for whom the DNA of known descendents was compared that might make it clearer.