Because I wanted to follow my DNA back another generation to my mother’s mother, I did the same thing for her. I discarded one chromosome from each of my mother’s pairs, which left me with the twenty-three chromosomes she got from her mother. Again I pushed them a little farther up the table, back through time and to the spot in the tree where my maternal grandmother perched. Then I added twenty-three blue strips of paper to pair up with each of the chromosomes that my grandmother gave to my mother, to represent all the genetic material that my grandmother had but that was never passed down to my mother. In order to disassemble those twenty-three chromosome pairs and rebuild their original state in my grandmother’s cells, I chopped them up and painstakingly pushed tiny little pieces of paper between the strips into alignment with others.
Some of the motley strips that represented my grandmother’s original chromosomes had four different segments with three different colors. Because I was using specific colors to indicate where the DNA would end up, I could see which segments of my grandmother’s DNA came down through the generations to me, which went to my mother but not to me, and which were not replicated at all. If I continued to do this for one hundred generations, all the pieces of my personal genome would become smaller and smaller and be dispersed further and further throughout my genealogical tree.
Yet as I dispersed and reconnected the chunks back through the time span of just a few generations, I could see they were not just different colors but also slightly different lengths. At first it was easy to divide the genomes up and push them back through the generations, but it wasn’t too long before the chunks reached a size where they were not divided at all. They moved from one generation to another without changing in size. This was what Ralph wanted me to see: It is often the case that these segments of DNA will be passed on whole to the next generation and then, still whole, to the next. It may be many generations before they are once more chopped down.
A common misconception of the way DNA is passed on is that with each new generation any one section of DNA is divided in half before it is passed down. We have learned, of course, that there are exceptions. The genome isn’t a perfectly smooth collection of equal bits that break up and come back together in exactly the same way. Because Y DNA and mtDNA don’t get reshuffled with other DNA, they can be used to learn something about an individual in your family tree who lived 10,000, 50,000, or 100,000 years ago. That person is still there, in a sense, in you in a completely disproportionate way to the rest of your grandparents. The X chromosome is different too, as it recombines in an uneven way across the generations.
Still, despite the fact that we have come to accept the idiosyncrasies of Y and X chromosomes, as well as mtDNA, we have assumed that the rest of the genome followed the basic pattern of being halved, piece by piece, and halved again. As we have seen, the most impassioned antigenealogists like to cite the principle that if you go back ten generations, you have 1,024 ancestors in that generation alone, which means that the amount of genetic material you are getting from any one person in that group is 1/1,024 of their genome, so small as to be virtually meaningless. The implication is that you are a genetic soup, and it is pointless to look for patterns in soup. This is a powerful argument, but I have long wondered how much of its force comes not from its math but from its confident, almost cartoonish absolutism.
Indeed, according to Ralph and Coop, it turns out that many segments of autosomal DNA are moved as is from one generation to another and another without being reduced further. The reproduction of a genome is not a smooth process of disintegration into ever-smaller pieces; it’s a lot more uneven than we suspected. As students Ralph and Coop were taught the standard notion that because DNA is halved at each generation, it means you must get a quarter of your DNA from each grandparent, an eighth from each great-grandparent, and so on. When they first realized this could not be true, Ralph told me, “It took us a while to get used to the idea, because we were used to thinking in terms of a half, a quarter, an eighth, because at first everything gets cut in half. But if you take a segment of DNA and you cut it in half and you put it in two bags and the next time you cut in half again, pretty soon the chunks that you have are so short that the chance of them getting cut at all is really small. So the whole chunk moves into one of the bags.”
If every chunk of DNA were halved with every generation, the result would be a rather neat picture of proportionately shrinking segments that matched an expanding fan of cousins. But if the cut and shuffle of DNA down through the generations is not a smooth, even process and relatively large chunks of DNA may be passed on through generations more or less unchanged, it has some interesting implications for what DNA can tell us about the past. For example, if you share chunks of DNA with a fifth or, let’s say, a twentieth cousin that have been inherited from a common ancestor, those chunks may be the same size as each other. Ralph suggests the window of relatedness is about 500 years long, meaning that if you share a chunk of a certain length with someone, you probably had a common ancestor somewhere within the last 500 years. A smaller chunk could have been inherited from an ancestor who lived anywhere from 500 to 1,000 years ago. An even smaller chunk could be from a common ancestor who lived between 1,000 and 1,500 years ago.
This has implications for the genetic databases at companies like Family Tree DNA and 23andMe. Users may find relatives who are first, second, or third cousins based on how much DNA they have in common. They will also find they belong to a massive group of more distant cousins. According to Ralph, based on the DNA you have in common, these distant cousins may be fifth cousins, but they might also be fifteenth cousins, and you may share a common ancestor much further back than you think. In fact, when you consider any pedigree, said Ralph, you soon stop getting many new chunks. We may be made anew in each generation, but it’s out of blocks, not powder.
There are other ways that the transmission of DNA over many generations is irregular and that may have a big effect. For instance, women have a higher rate of recombination than men. The size of the population that your ancestors came from and how stable it was affect what is passed down. Keep in mind too that thirty-two is just the average number of times the chromosomes break and recombine at each generation, and that can vary. Geneticists also talk about “hot spots” on chromosomes, places that may be more inclined to recombine over and over.
When you take all of this into account, you can use segments of identical DNA from common ancestors to unearth connections from the past three thousand to four thousand years.
• • •
Curiously, even if some long-past relatives are more prominently represented in our genome than we imagined (relatively speaking), there are many more who have simply evaporated. This is because our personal genetic tree is not equivalent to our genealogical tree, which is to say that not every one of our direct ancestors has contributed to our genome.
If you look back ten generations, everyone in your lineage is by definition a member of your genealogical tree. The fact that they paired up with the spouse they did and then had a child who became your ancestor is an unchangeable, existential bond. When you first start looking back into your lineage, the genetic and genealogical trees are, of course, exactly the same: Your genealogical parents are your biological parents, and so on. But there is a point in time at which the genetic connection between you and most of your many generations of grandparents vanishes away to nothing, despite the fact that your genealogical tree keeps growing.
Beginning around eight to ten generations back, geneticists agree, there were so many people contributing their DNA to your family line that by the time you came along, a lot of it has simply dropped out. If you could trace every piece of your DNA back through time, it would follow the branches of your genealogy, but it would trace fewer and fewer of those lines as it went on, leaving your genetic tree much smaller than your genealogical tree and leaving you in the odd position of being biologically unrelated to ma
ny of your blood relatives.
Estimates of how many people drop out of the genetic tree vary. By the time you go back ten generations, it may be that you have completely lost the genetics of at least one and likely many more grandparents. By the time you go back sixteen generations, you will have 65,336 ancestors (not counting possible repeats where an ancestor may appear on more than one lineage). According to Nick Patterson at the Broad Institute, this is far more than the number of distinct chunks of ancestry in your genome from that generation; there will only be around one thousand such chunks contributed by a sixteenth-generation ancestor. So most of your sixteenth-generation ancestors will have contributed nothing to you genetically. If you built a time machine and traveled back four hundred years and, let’s say for the sake of argument, found yourself in a romance with one of your sixteenth-grade grandmothers, the good news is that you can feel fine about having children together. However morally bizarre that might be, it would not be genetically problematic.
But wait, doesn’t this mean the antigenealogists have a point—isn’t this the same thing as saying that the human genome is quickly chopped down into meaningless dust over the generations? While it’s true that a lot of our personal genome eventually dissolves, it does so in a patterned way and there is meaning in the trail it leaves. If some people disappear from our genome while others remain, that is one more pattern shaped by history.
• • •
If you take any two modern Europeans, no matter how far apart they live, they will likely share millions of genealogical, if not genetic, ancestors within the last thousand years. Still, even though the genealogical tree is much bigger than the genetic tree, it doesn’t fan out forever either. Because the number of ancestors in a tree doubles with each generation, any person’s tree of possible ancestors grows exponentially and quickly reaches a point where it is greater than the population of the world at the time.
For example, geneticists and historians vary somewhat in their estimate of how long a generation is. Some say twenty years and some say thirty, so let’s assume it’s around twenty-five years. We’ll also assume that there is just one person in the world today, let’s say you. In order to arrive at you, there must have been more than a billion people on the planet approximately 750 years ago. Or, to put it another way, about thirty generations ago there would have to have been roughly 2 billion people in the world, whose children and grandchildren and so forth met one another and married until one day you appeared. But in fact there were only 400 million people in the world 750 years ago. What this means is that your thirtieth-great-grandparent along one line is probably your thirtieth-great-grandparent along many other lines too. Genealogists call this pedigree collapse.
Pedigree collapse is common in family trees that go back to the nineteenth century and earlier. In many different cultures marriage between cousins was not the exception but the rule. (For more on the demographics and implications of cousin marriage, see chapter 14.) Some scientists, like Ralph and Coop, argue that by two thousand to three thousand years ago, everyone who was alive at that time across the globe was actually a genealogical ancestor of everyone alive today. The argument is a purely mathematical one based on the idea that, because your theoretical genealogical tree would be so massive two thousand to three thousand years ago (over 120,000 trillion ancestors), it must surely include all of the much smaller number of individuals (50 million to 170 million) who were alive in the world at that time. Many population geneticists I spoke to found this to be a completely noncontroversial theory. It wouldn’t take long for the networks of relationships in different countries to be altered by the introduction of just one or two travelers from a distant area, connecting all of the populations of the world. Others whose work is more engaged with the events of history found the idea implausible: While it is technically possible that the connecting events happened, they believe it is more likely that the world’s populations were more completely isolated from one another for a longer period.
If the hyperconnectedness of humanity is true, it would mean that everyone alive today—you, your neighbor, Vladimir Putin, and the emperor of Japan—could count the same Egyptian pharaoh, as well as everyone else alive at the time, as a distant grandparent. A set of common genealogical ancestors doesn’t mean, of course, that people today don’t have genetic differences (which can be seen in any number of experiments, such as those exploring the genetics of Britain described in the previous chapter). It doesn’t mean either that once you go back three thousand years everyone’s family history is effectively the same. Even if both you and the emperor of Japan can count the same pharaoh in your family tree, the pharaoh may appear many more times in your tree than the emperor’s. Genetic history would not be possible if everyone’s genealogical tree were identical a few thousand years ago. If you imagine a great network of ancestors stretching from now back through time, with every person a node in the network, all the people alive in the world three thousand years ago who left any descendants would be significant nodes, because each person alive today could trace a path back to them—but it wouldn’t be the same path. Some people would trace thousands of paths back to the same node, while others would trace many fewer.
Curiously, it is often the case that when someone says, “We are all related to Confucius or Boadicea or Erik the Red,” the implication is that there is no texture in this history, that if we look back far enough, everyone in our family was the same, so there is little of interest to say about the connection that one person or group of people may have with people from the past. This is not the case. The topology of the human network, in which we are all nodes, is incredibly complicated. While there are points of sameness—perhaps we can all trace at least one connection back to everyone from three thousand years ago—that sameness does not mean that all the other pathways we trace back to shared ancestors do not have great significance. If you start tracking the ancestry of segments of the genome, like the Y chromosome and mtDNA, the picture of shared ancestry becomes much more complicated again. The patterns of all the paths tell us things about history that we might otherwise never know.
Recall that in the British genetics project people in Cornwall and Devon and other regions carried a signature of their local region from a certain time. The fact that they had a Cornish or Devonian flavor to their genome doesn’t mean that they didn’t also have a few ancestors from Scotland or France or that there weren’t Catholics or Vikings or even a Chinese philosopher in the genealogical mix. Genetically they were not purely one thing; the same is true of their genealogy. Their heritage was a matter of dilution—enough of their ancestry had come from the one time and place that it was still recognizable. This will be true whether we look back one thousand years or five thousand years.
• • •
In addition to some general rules of relatedness in Europe, Ralph and Coop identified some tantalizing variations. People in southeastern Europe seem to share many more genetic common ancestors than people in other parts of Europe. The period when they share this large cluster of common ancestors dates to around 1,500 years ago, when there were significant Slavic and Hunnic expansions.
In dramatic contrast to the rest of Europe, most of the common ancestors of Italians seem to have lived around 2,500 years ago, dating to the time of the Roman Republic—the period that preceded the Roman Empire. Modern Italians certainly share ancestors within the last 2,500 years, but far fewer of them. In fact, Italians from different regions of Italy today share about the same number of common ancestors with one another as they do with people of other countries. It is as if other European nations are more genetically homogeneous, while Italy is composed of many smaller countries. How do we make sense of this?
To understand it we have to understand not only the way that the living material of families is passed on but also the world in which the families were formed. While DNA was spread across Italy by all the normal processes of reproduction, those processes
were shaped by the culture and geography in which they took place. Genomics by itself won’t give us the whole picture of European history.
There are many stories that might explain the peculiar pattern of Italian common ancestry. For example, Ralph and Coop suggest that the region may have been less affected by the population expansions that spread across the rest of Europe in the last two thousand years. To test the possible scenarios, said Ralph, they would have to look at demographic, linguistic, and other historical patterns. Many factors would have to be taken into account, such as whether the difference in common ancestors was true for people whose families had lived in Italy for centuries (which would confirm that the effect did not exist simply because an entirely new population moved in from a different country five hundred years ago).
I was intrigued by the way that, in addition to the larger group of common ancestors from 2,500 years ago, other genetic analyses of Italy show that many small different genetic groups are related to each other in a way that forms a gradient from north to south in the country. I asked Guido Tabellini, the economist who measured the difference in social capital between the north and south of Italy, if he could think of a reason why there were fewer common ancestors between some groups in the most recent 2,500 years. He pointed out that the political entities that ruled northern and southern Italy were separate until the middle of the nineteenth century, and that separation reduced economic and social integration within the country as a whole. In addition, he said, the north and south were ruled in very different ways: “Northern Italy had the tradition of more independent and enlightened government, whereas southern Italy had the tradition of being dominated by external influence.
“A second important difference which has not been studied much,” said Tabellini, “is the role of the family. If you look at the family traditions within Italy, they are also very different, and so are attitudes towards females. The attitude towards people outside of the family can also be explained by these different family traditions.”
The Invisible History of the Human Race Page 26