Other than the fact that if you oversleep your alarm and wake up at 11 am, you’re not fired, and, in fact, pretty much no one cares, the thing I will miss most about being a grad student is the chance to see talks by amazing researchers. And last week I saw two very good ones: Svante Pääbo and John Johnson. This post will focus on the first.
On Wednesday, a standing room only (well, there wasn’t actually standing room either, except behind a column that was blocking the entire view. I ended up sitting in the aisle) audience turned out to listen to Svante Pääbo discuss efforts to study and sequence the genes of archaic humans. I’ve already alluded to how amazing I think his group’s work is. They extract DNA from fossilized archaic humans, and can sequence this DNA and map the genome of the specimens they study. It is worth dwelling a little on how difficult and cool this is. First off, extracting DNA from a fossilized sample is a complicated feat of filtering. The majority of the DNA on the sample will be from microbes and fungi that lived and live on the bone. And the proportion of signal to noise here varies wildly. From the 40% original DNA in the Denisova cave sample, to 0.02% in an archaic modern sample from a cave outside Beijing that they attempted to sequence most recently. This is before concerns about contamination from modern human DNA of archaeologists, curators and experimenters. Or even things like horse DNA from horsehair brushes used in fine archaeological work. And this filtering is made all the more difficult because DNA breaks down over time. This isn’t like the Jurassic Park fantasy of a pristine specimen preserved in amber or permafrost. The DNA molecules have broken apart into tiny snippets of no more than 50-60 bases on average, and almost never anything over 200 bases. To map out a genome from this mess is like trying to reconstruct a book you’ve never seen from a pile of fragments of five words at a time, with parts randomly repeated and missing. Oh, and the book is in some language you don’t really understand.
But the fruits of this work are as rewarding as the labour is difficult (sorry for sounding like such a fanboy, but, well, I am). Comparing the mitochondrial DNA of modern humans with that of the Neanderthal and Denisova cave specimen, we were able to infer the split between our ancestors and the Neanderthals/Denisovans occurred around 650,000-850,000 years ago, and the split between Neanderthals and Denisovans afterwards, but only shortly afterwards.
We already talked about how Pääbo’s group’s research drove a nail in the coffin of the idea of total replacement. This is simply done by comparing the amount of difference between Neanderthal/Denisovan DNA and the DNA of various populations of modern humans. What is seen is that Neanderthals contributed up to 2.5% of the genetic material for all non-Africans, and that Denisovans (whose only specimen we have comes from Siberia) contributed up to 7.5% of the genes of Papuans and Micronesians. However, Pääbo emphasized that while this was evidence against total replacement, it was also evidence against any large-scale continuity – he called the emerging paradigm “leaky replacement”. The exciting part is that Pääbo and his collaborators can actually do much more than this. By sequencing the DNA of an archaic modern homo in northern Asia – and finding that they contain no sign of Denisovan admixture, they can actually infer a lot about the migration patterns, and locales of interbreeding among archaic hominins.
They can also hypothesise population history from when coalescence events occurred (I was confused on how you can do that with only one specimen? repeated areas of the genetic code?) to show that at one point, there was actually a larger overall population of Denisovans than archaic moderns, but by the time of our fossil, they were a small group. This is just absolutely astounding considering all we have of the Denisovan specimen is a little piece of hand bone – we have not the slightest material evidence for what the Denisovans were like or how they looked – and yet we know so much about them.
The exciting new work of Pääbo’s group is to actually zoom in on those genes that are different between Denisovans, Neanderthals and humans, single them out and attempt to figure out what these genes are related to and what the changes mean. Preliminarily, they find a lot of changes in genes that seem to be related to eye diseases.
I was surprised by Pääbo’s motivation for this latest work, though. He posited an oddly Richard Klein-ian idea of a snap revolutionary genetic change enabling art, better tools and what have you. It was strange to see the person whose work raised the “status” of the Neanderthal in the human family tree so much dismiss their humanity so nonchalantly. I am quite strongly predisposed against this idea to say the least. However, Neanderthal genetics already overturned my preconceptions once. I wouldn’t put it past them to do it again.
One interesting side-thought here is that maybe falsifiability of hypotheses isn’t all it’s cracked out to be. Richard Klein made a theory that, at the time he made it, was completely unfalsifiable (the revolutionary genetic change at the 40-50 kyr mark). And yet, and I would say partly as a result of his theory, here we are, many years (a couple decades?) later, on the verge of being able to test it. Perhaps UnlearningEcon and I are too hard on macro-economists? Perhaps Randall Munroe and Lee Smolin are too hard on string theorists? Perhaps the thing to do is to come up with theories even if you don’t think they’re falsifiable, and if these theories are compelling enough, someone will figure out a way to test them? I’m not persuaded, of course, but it’s a thought.