Tuesday, 15 December 2015

3rd genome of Christmas: the Denisovan little finger.

In the early 90s Svante Paabo, a charismatic, energetic innovator, made a bold proposal: that to study human origins one would do well to sequence the DNA of ancient hominids, in particular those species which had gone extinct. After all, DNA could be detected in their bones, provided they were not too old and kept dry and cold.

At first this seemed like science fiction, as people were just starting to get their heads around the sequencing of an entire human genome. But even in the 1990s it was in fact feasible, if technically challenging.

Svante and his team started with mitochondrial DNA from Neanderthal bones. (If you’re new to genomics, mitochondria are the power plants of the cell and have their own, tiny genome, inherited only from the maternal line.) Along the way, they discovered (much to their chagrin) all the myriad ways one could contaminate a DNA sequencing experiment with trace amounts of modern human DNA (there are many). They also figured out how to prevent them, for example by keeping bones cold and ‘dirty’ before extraction in the clean room. 

The mitochondrial genome they published in the late 90s confirmed that all human mitochondrial diversity lies outside of Neanderthal mitochondria – in other words, the out-of-Africa hypothesis of human origins was correct.

The next-generation sequencing revolution of the mid-2000s spurred Svante and his team to push for the next level: sequencing the Neanderthal genome (not just the mitochondrial one). But the ‘main’ genome, unlike the tiny mitochondrial one, couldn’t be amplified simply using a PCR reaction. It was far, far too big for that. Direct (random) shotgun sequencing of ancient bone will catch everything in the bone sample, which includes a lot of dead bacteria. They screened many, many bones, and finally found one in which 2% of the DNA sequence was hominid. That might not sound like much, but it’s enough. The Neanderthal genome team sequenced around one-fold coverage  of this ancient hominid (usually you need >10-fold coverage (redundancy) to get an accurate genome; one-fold coverage tells you a lot, but one has to accept gaps and errors).


But there was a huge surprise: a consistent statistical trend put the main Neanderthal genome slightly closer to European and Asian human genomes than to sub-Saharan African ones. This agonised the analysis team (on which I had the pleasure of playing a small part) and debates raged about all the complex ways in which trace contamination or complex analysis artefacts could have given rise to such a result. But at the end of the day, thanks in part to the innovation and persistence of Ed Green (on the experimental/processing side) and David Reich (on the population genetics side), it came down to a low level of inbreeding between Neanderthals and modern humans, presumably just after modern humans left Africa.

Needless to say, many people disputed this analysis, and put forward other, more complex hypotheses to explain what could have generated this weak signal.


Then there was a remarkable discovery: the tip of a little finger from what was thought to be a Neanderthal in a cave in the Urals, close to the small town Denisova. This turned out to be one of the best ancient bones ever sequenced, as around 70% of the DNA was hominid.

Another huge, compound surprise: this bone was not Neanderthal, rather a separate hominid species. Furthermore, this set of hominids had very clearly interbred with Papua New Guineans and Aboriginal Australians, presumably en route as they migrated from Africa towards the Australian continent.

This time, the signal for inbreeding was far larger. Arguments that this was some kind of complex contamination were clearly a big stretch. There are, sadly, not many native Papua New Guineans or Aboriginal Australians in research. On top of the new question of whether Neanderthals had interbred with the out-of-Africa migration, there was now clear-cut evidence of both a new hominid species (with minimal fossil evidence) and more extensive interbreeding.

But ancient hominid DNA sequencing doesn’t stop there. Another Neanderthal bone unearthed in the Urals showed evidence of a fourth hominid genome, originating somewhere in central Asia. And the rather amazing “cave of bones” in Spain has brought forth evidence of deep-lineaged, modern human mitochrondria.


Suddenly, the neat, ‘out-of-Africa’ theory, that a single explosive species replaced all previous hominids, was mainly right - but not quite so neat. There is a web of multiple hominid populations, all originating in east Africa as far as we can tell and setting forth around the globe, meeting each other and having occasional sex. It’s interesting to speculate how much communication and collaboration – or warfare – happened amongst us and our cousins.

Expect more surprises in the coming years – we are just starting to understand the amazing story of how we evolved.