Bloodless Bloodlines and Other Genealogical Paradoxes

Like a lot of families in the 1970s, mine got the genealogy bug. I remember my mother drawing out our tree, a branch for each parent, which then split into branches for four grandparents, eight great-grandparents, and so on. Some parts of the tree are still blank; others she filled in all the way back to the 1200s. As a kid I was fascinated by our genealogy, despite my vague sense that there was something absurd about it. If my mother kept drawing the tree back a few hundred generations, our ancestors would have grown exponentially until they outnumbered all the people who ever lived. Making matters worse, every now and then my mother found herself in an awkward situation when two distant cousins had married. How could she represent that in the tree? After all, in an ever-branching tree, each couple appear to be unrelated. But in the case of these cousins, their ancestry did not diverge but actually converged on a common ancestor.

It turns out that those awkward cousins are what keeps our family tree--along with everyone else's--from exploding into exponential absurdity. Go back far enough, and all the branches on your family converge on common ancestors. Go back further, and all family trees meet. How far back you have to go is the subject of a report in today's issue of Nature. The researchers put together a statistical model of today's population and worked their way back. To give the model some historical realism, they didn't let their virtual people randomly join together to have kids. Instead, they included some geographical structure. People in Indonesia, for example, have historically had children with other Indonesians. But they've also had a small amount of interchange with people in other regions--even with people as far away as Madagascar, which some Indonesians colonized 1600 years ago. (Douglas Rohde, the lead author, has posted a draft manuscript of a longer paper on this work on his MIT web site.)

Once the model was ready, Rohde and his colleagues used it to go back in time, in order to find a common ancestor of all living humans. They estimate that that person lived just 2,300 years ago. Of course, this doesn't mean that this person gave rise to all humans single-handedly. All it means is that, by one route or another, all family trees lead back to him or her. Those trees may belong to Australian aborigines, Arctic Inuits, or the residents of Easter Island.

Go back 5,000 years (160 generations) and things even weirder. According to the model, 80% of all people alive at the time are ancestors of every living person. The authors of the paper sum up these astonishing results with eloquence not usually found in the pages of Nature: "No matter the languages we speak or the colour of our skin, we share ancestors who planted rice on the banks of the Yangtze, who first domesticated horses on the steppes of the Ukraine, who hunted giant sloths in the forests of North and South America, and who laboured to build the Great Pyramid of Khufu."

This has a nice Family-of-Man feel to it, but it doesn't encompass the whole story. As Jotun Hein of Oxford points out in an accompanying commentary in Nature, we tend to think of genealogy as being the same thing as genetic continuity. But they aren't. If you have children, the chance that any one of your genes makes it into any one of them (instead of your spouse's) is 50-50. If one of your genes doesn't manage to get into any of your kids, it becomes extinct. The same lottery takes place when your children have kids of their own--will your genes make it into your grandchildren? The odds of one of your genes surviving for two generations are tougher than for one, because now it has to compete with the genes of your in-laws as well. By the time your first great-great-great-great-grandchild is born, he or she may not inherit any of your genes. Your bloodline stops dead, at least in that particular branch. (As a result, the new wave of DNA tests that promise to reveal your genealogy actually give a narrow view of your full heritage.) While 80% of people who lived 5,000 years ago are ancestors of all living people, 20% are the ancestors of no one at all.

The fact that genealogy doesn't equal genetic continuity has another strange consequence: the genealogy of our genes is different from our own genealogy as individuals. The most recent common ancestor of living humans may have lived 2,300 years ago, but the most recent common ancestor of any particular gene was carried by people who lived much earlier than that. The youngest of these genes may be about 60,000 years old, but others are 150,000 years old and others perhaps 500,000 years old or more.

It will not be easy to pull apart genealogy and genes; they seem so obviously united in our own experience. But then again, in our own experience the Earth doesn't seem to be speeding through space, and time seems to be constant. Things are not always what they seem. One of the best book-length treatments of this counterintuitive weirdness is The Ancestor's Tale by Richard Dawkins, which comes out in October. I'll have much more to say about that book soon.

UPDATE October 10, 2004: See also Posted by Carl at 12:59 PM