Rich and Strange Aeons

I'm reading a recent book about DNA and human migration, and it raised an interesting point: many of your ancestors didn't contribute DNA to you. This can be seen by first simplifying: imagine that crossing-over didn't happen, so that each of your chromosome came unaltered from one of your parents. You have 46 chromosomes, they each have 46 chromosomes, obviously half of the parental set got lost. But consider: after one generation you have two parents, after two generations you have four grandparents... after six generations you have 64 great^4-grandparents... and only 46 chromosomes. So at least 18 ancestors from back then failed to get any chromosomes into you: there's just not enough chromosome to be mapped to all of your ancestors!

In reality crossing-over does happen, but the book says that just means a linear increase in the number of splices, so that after 10 generations you have about 700 ancestral DNA segments but 1024 ancestors -- once again, 30% of them are left out of you.

In the extreme case, you have 30,000 genes, so ignoring splices inside genes, at 16 generations (only 320-640 years!) you have 65,536 ancestors and most ancestors than genes for them to contribute.

Of course, as you go back you start having nth-cousin ancestors and thus a smaller number of unrelated ancestors in a generation, but the principle still holds.

I'm currently reading this book, on recent human evolution. It's not an evolutionary psychology book, as that capsule description made someone thing, though it probably will be talking about evolution and the brain. It's by an anthropologist and a physicist, which is a bit odd, but nothing's obviously wrong about it so far.

Main idea isn't new to me: that human evolution, far from being halted by civilization, has in fact sped up. See John Hawks, who *is* an evolutionary biologist (longer piece, by him). Basic idea is that more people = more mutations and thus chances to adapt, and new environments, many of them created by us, means new things to adapt to. (Environment also includes new foods and diseases, and things like cities.)

The book structure is a bit odd; it spends most of chapter 2 talking about the possibility of incorporating adaptive Neanderthal genes, something for which in 2009 they had no real evidence apart from some intermediate skeletons in a few locations, not like the recent genetic evidence of Neanderthal and Denisovian genes. (Which might about as a verified prediction, then.) Speculation on incorporating genes, speculation on what we could have gotten out of them... not entirely divorced from evidence, but still odd.

But there's also math. Simple math. A neutral gene variant -- one with no adaptive effect -- lives or dies by chance, with a chance of taking over a population of 1/2N, N being the size of the population. But, per Haldane, an *adaptive* gene which grants one s% more children (a highly statistical measure, obviously) has probability 2s of sweeping a population. So a gene granting 1% more fitness has 2% chance of taking over. Not high for one such gene... but out of a 100, we'd expect 2 to take over.

(I suspect small number simplification; a 50% adaptive improvement can't be 100% likely to take over, though we can expect good things of it.)

Per Hawks, imagine that the frequency of 1% beneficial mutations is 1 per 10,000 people. Then a population of 10,000 will have 1 such mutation per generation, and it'll take on average 50 generations for one of those mutations to start taking over rather than withering away by chance.

But in a population of a million people, there'd be 100 mutations, and so we'd expect 2 mutations a generation to start being fixed. In 1000 years, 100 successful adaptations, rather than 1.

And we do in fact see lots of genes in the process of 'sweeping' populations. Mostly in metabolism and digestion, disease resistance, reproduction, DNA repair, and the central nervous system. The first two have obvious examples and aren't politically alarming, examples adult like lactose tolerance (less than 8000 years old in Europeans, 3000 years old in Tutsi) and malaria resistance (also 3000 years old.) They say skin color might be another one; hunter-gatherers can get vitamin D from meat, so even in the north don't have a big need for light skin, and several light-skin mutations are younger than agriculture. White people may not have existed 10,000 years ago. Europeans and Asians have more inactive variants of an African gene that promotes salt retention -- useful for high-sweating tropical dwellers, less so in cooler climates, and also leading to hypertension in a modern diet. East Asians apparently have various genes reducing the risk of alcoholism, and the authors speculate that the high rates of diabetes and alcoholism among 'indigenous' populations owe a lot to a near total lack of genetic adaptation to agricultural diets with lots of starch and alcohol.

But the central nervous system? That's the *brain*. Suggesting differences there is politically charged in the way that fire is hot. Yet if evolution happened in other parts of our biology, and it did, why should the brain be immune?

They agree that the amount of time we're talking about isn't enough to build up complex adaptations from scratch. But they point out that you don't need to; simple adaptations can still have big effects. There's huge variation in dogs (chihuahua, Great Dane; smart and friendly border collie, dumb basset hound and mean pitbull), just from shifting the balance of traits present in wolves, plus adding something that means dogs are far more attentive to humans (with the pinnacle of border collies, who can learn words in about 5 repetitions.) They don't mention the prairie and montane vole species, where a single gene change means monogamous pair bonding or not. And shifting the frequency of existing genes is far easier and faster than fixing a new mutation.

So even with all humans sharing the same basic mental mechanisms of intelligence and personality, the proportion of those mechanisms in various populations could easily have changed in the past 10,000 years. But you'll have to wait until I read the rest to hear what...