By Daniel Hubbard | September 4, 2011
To pick up where I left off last week, after discussing yDNA and mtDNA, I should go over autosomal DNA, the vast majority of our genetic material.
Autosomal DNA is a very different story from yDNA and mtDNA. On the plus side, autosomal DNA covers every line of our pedigrees, not just the paternal or maternal line. That gives the potential for your DNA to give you a handle on many lines that don’t share your father’s yDNA or mother’s mtDNA. While it is true that autosomal DNA is shuffled often, the shuffling is not so thorough and sections that are passed down intact give the possibility of matching patterns of DNA. To go back to my plaid cloth analogy, if you get a significant piece of the plaid cloth from the crime scene you can tell that it is plaid, you may learn how large the blue squares are and the width of a stripe or two. You will have a chance to get the order of the stripes. Showing that another piece of cloth matches that pattern should be possible. Too small a scrap and all you know is that part of the cloth was blue. Not so helpful.
With pluses can be expected minuses. One minus is that shuffling of genes adds some complexity. Both parents contribute exactly half of a child’s autosomal DNA but because of the random shuffling, once you go back more than one generation it isn’t possible to say how much any ancestor has contributed. For example, you would expect each grandparent to have contributed one quarter of your autosomal DNA and on average you would be correct. Nevertheless, after your father shuffles the DNA in his chromosomes, grandpa may have contributed a few more percent than grandma or visa versa. The farther back you go, the more the random shuffling comes into play.
Taking Ancestral Attendance
The way we inherit our genes makes a big difference to how we inherit our autosomal DNA. If we strictly inherited the same number of genes from each ancestor in each generation, interesting things would start to happen about 14 or 15 generations ago. We have about twenty-five thousand genes. and if we ignore the complexities of the X and Y chromosomes and the genes in the mitochondria, what happens as we go back through the generations? By the time we have thirty-two thousand ancestors in a single generation, that’s fifteen generations ago, we have more ancestors than genes. Obviously some ancestors appear multiple times in our family trees so that we have fewer ancestors than positions in our pedigrees but eventually we get to the point where it is impossible for each ancestor to have contributed at least one gene to our DNA.
One by one isn’t the way we inherit our genes. They are grouped on chromosomes. If recombination, the shuffling of the genetic deck, didn’t happen then we could have at most forty-four ancestors contributing to our autosomal DNA because we have twenty-two pairs of autosomes. We go past that number of ancestors between the 5th generation (32 ancestors) and the 6th generation (64 ancestors). That is a surprisingly small number of ancestors contributing to our autosomal DNA compared to our first attempt at a calculation. However, this isn’t the correct answer either. Our genes do get shuffled.
So if we add recombination, how much difference to this calculation does it make? It certainly makes the numbers harder to derive. Now probability functions come into play and to get an answer requires doing a simulation. After searching high and low for someone who had the necessary numbers, I found not the probabilities but the actual results of the required simulation from a geneticist. The answer turns out to be that almost all of us have DNA from all 16 ancestors that are 4 generations back but that is the last generation where we can make that statement. Go back one more generation and only about half of us have DNA from every ancestor in that generation.That is, about half of us inherit DNA from 31 or fewer of our 32 ancestors that are 5 generations back.
So Many Ancestors, So Many No-Shows
In fact the simulation showed that as we move backwards through the generations, at first the number of ancestors that have contributed genes, we could call them “genetic ancestors,” grows very fast. Early on the number of genetic ancestors per generation doubles every generation, just as you would expect. Then the growth slows. Here and there in our pedigrees there is an ancestor that does not contribute to our DNA and, of course, none of the ancestors of those people contributes either. The number of genetic ancestors doesn’t quite double with each generation any longer. Finally, the number of genetic ancestors per generation virtually stops increasing. Most of us have DNA from no more than 125 ancestors in any generation recent enough to be involved in our research. A very interesting result.
For example, six generations back, where we have 64 ancestors, most of us get our autosomal DNA from only about 54 ancestors. Eight generations back we have received DNA from a bit fewer than 100 of our 256 ancestors. By eleven generations back we have reached that virtual plateau and have DNA from about 125 of our 2048 ancestors. That is just 1 of every 16 ancestors in that generation. From there on, the fractions are easy. They go down by half with every additional generation. At 12 generations only 1/32 of our ancestors contributed DNA to our autosomes. So, very often the answer the question in the post title is “not to inherit.”
So, the broadness of autosomal DNA testing comes with a price or two. The time depth cannot very deep. I’ve read differing opinions. Some people talk about 10th cousins while others, including Family Tree DNA say that beyond five generations and you are really pushing it. That would seem to be closer to the truth. Thinking about this type of DNA takes a different mindset from yDNA or mtDNA. Because the amount of autosomal DNA we inherit from ancestors has an element of randomness, we can’t predict how much we will inherit from any one ancestor. Because of the way that randomness occurs, we can’t tell in any particualr case, what ancestors will have contributed no DNA at all.
This also means that if you compare autosomal DNA with a moderately distant cousin that you have tracked down, there is a chance that even without a mistake in your research—when you really are distant cousins—you will share no DNA. It isn’t even that the test fails to detect the relationship. It is that there is nothing to detect. Put another way, a negative result in an autosomal DNA comparison can disprove a close relationship but not a more distant one. Even third cousins have a small chance of not matching.
On the other hand, the more distant the cousin relationship, the greater the number of cousins you are likely to have. So while your chance of finding a match with any one person drops rapidly as the distance between you increases, that is at least partially compensated by the fact that if you are simply looking for any match, the number of people you might match increases with more distant relationships. In other words, you might not have any possibility to match that new fifth cousin you just found but your chances of matching one of your fifth cousins are improved because you probably have lots of cousins that distant.
Playing with Matches
Another difference between autosomal DNA on the one hand and yDNA and mtDNA on the other, is that if you have an autosomal match, you won’t necessarily know why. You can’t say what line in your pedigree is the one that contains the common ancestor. Your paper trail may give you a very good idea. Other DNA tests may also help. If you are able to have your parents tested and you match someone who matches your mother but not your father, then at least you know what side of the family to look on. Be careful though, there is no rule that says that you can’t match someone for more than one reason.
For better and for worse, autosomal DNA is not “directional.” That can be an advantage when you have no way to use yDNA or mtDNA to get a handle on your problem. Because it is so broad, it is probably a good way of simply finding relatives who happen to get themselves tested. If you don’t have a problem per se but simply want to be tested and wait for unknown distant relatives to match you, then the broadness of autosomal DNA lets you cast a wide net.
The narrowness of yDNA and mtDNA can be a disadvantage if you don’t have the right person to test but it also allows for precision. They allow specific hypotheses to be tested. For instance, what do you do if your yDNA test doesn’t match people that the paper trail implies that you should match? How do you find where along that chain of fathers did things go wrong? You find a distant male relative with a Y chromosome that should also match and you triangulate on the common ancestor.
Years ago some distant cousins tracked me down because when one of them had her father tested, a problem appeared. There was a group of men who had consistent yDNA tests confirming the paper trail that connected them to a man who arrived in America in 1630. My cousin’s father did not match the test results of those men, though the paper trail says he should have. I was in Sweden at the time, but my father took the test for me. He matched my cousin’s father perfectly.
Following the paper trail and doing two yDNA tests allowed us to check the Y chromosome of a specific man who was born about two hundred years ago. When the tests matched, we knew what his Y chromosome was like and, by implication, the Y chromosomes of all the men between the two test takers and their common ancestor.
In practical terms, the results eliminated all the generations between our fathers and their common ancestor as the source of the problem. An incorrect father that could have occurred anywhere in our pedigrees over a space of eight generations, now was narrowed down to the most distant four generations. We’re still looking for the man that could help narrow it down even more.
There You Have It
So there you have it. Different DNA tests have different strengths and different weaknesses. If you need precision or great time depth, yDNA and mtDNA win. If you want to cast a broad net and time depth isn’t a concern then autosomal DNA might be interesting. If you don’t have any way of getting a yDNA or mtDNA sample and you have a problem that isn’t too many generations back, then autosomal DNA might help you out. It is all a matter of fitting the tool to the situation.