Our ancestors lost nearly 99% of their population, 900,000 years ago

Benjamin Thompson

Hi, Benjamin from the Nature Podcast here, slightly different show to normal this week, we're going to take a look at some of the recent stories that have appeared in the Nature Briefing. And joining me to do so are Dan Fox. Dan, how you doing?

Dan Fox

I'm doing great. Thanks, Ben.

Benjamin Thompson

And Shamini Bundell. Shamini, hi.

Shamini Bundell

Hi. I'm ready for some science. Yeah.

Benjamin Thompson

Excellent. Good to hear well, as I say, a few stories to get to this week. So Dan, why don't you go first, what have you got?

Dan Fox

So I've been reading an article in Nature about how human ancestors in Africa nearly went extinct around 900,000 years ago. So this is a paper that was published in Science. And it talks about how the entire population of our ancestors long before the emergence of Homo sapiens reduced down from around 100,000 individuals to just 1,280 breeding individuals, and it stays like that for around 117,000 years.

Shamini Bundell

Oh, is this stuff that they know from fossils? Presumably, it would be pretty hard to estimate numbers from that.

Dan Fox

So that's an interesting question that this research sort of answers in some ways. So this hasn't been worked out from fossils, or from archaeological evidence, but from a new model used to predict past human population sizes using 3,000 present-day human genomes.

Benjamin Thompson

Okay, so looking at the genomes then of current humans to look back all this way in time, right?

Dan Fox

Yeah. So, the method allows them to reconstruct ancient population dynamics using genetic data from present-day humans by constructing a complex family-tree of genes and then examining the finer branches of the tree with greater precision than before to identify significant evolutionary events.

Shamini Bundell

That's wild that they can do that. And so they've seen this, I guess it's like a bottleneck. Right? That lasted quite a long time. I mean, it obviously still is in our genes that they can see it, but has that had an impact on sort of subsequent hominin evolution?

Dan Fox

Yeah, so the researchers think this had a big impact on the evolution of humans. So they estimate around 98.7% of the human-ancestor population was lost.

Shamini Bundell

Ouch.

Dan Fox

And during this time, there were genetic changes that they've seen, there's chromosomes that merged together in this period. And so they think it's possible that this period led to the emergence of the last common ancestor of Denisovans, Neanderthals and modern humans.

Benjamin Thompson

Wow. Okay. And any idea what happened? Because, what, almost 100% reduction in the amounts of these ancient human-ancestors, obviously not quite 100%, because we're here, sort of chatting about it now. But what caused this? Do we think?

Shamini Bundell

I'm thinking meteors? I'm thinking biblical floods, do the genes mentioned any of that, or?

Dan Fox

Well, yeah, you know, I don't think that came out in the genetics, but this period is known as the Early-Middle Pleistocene transition. And it's a time when glacial cycles lasted longer and became more intense. And that in Africa, where these people were, resulted in longer, severe droughts. So it's possible that the answer is climate change, essentially.

Shamini Bundell

So this is before Neanderthals split from modern humans, Denisovans? Who exactly were this population? Like, do we actually know anything more about them?

Dan Fox

Well, the answer there is no, not really. So obviously, this research is done with modern-human genomes. There's no archaeological basis here. But there's not a lot of archaeological basis for anything during this period. There's a scarcity of fossils of humans between 950,000 and 650,000 years ago, and the researchers behind this study, actually think this might explain it, you know, maybe, you know, if there's only 1,000 living human ancestors around at any time, it's going to be quite hard to find evidence of them.

Benjamin Thompson

Well, an absolutely fascinating story. And I always enjoy our chats about ancient humans and ancient human-relatives and ancestors. But let's leap forward in time to today. And I've got two stories, both about Antarctica. And the first is one that I read about in Nature, and it's looking at, well, it's looking at Antarctica’s dirty secret, I guess. And this is based on some research published in the journal PLOS ONE in August. And I guess if you kind of picture Antarctica, you might think of it as this kind of unspoiled wilderness. But it appears that really isn't the case.

Shamini Bundell

I'm not sure I can imagine what could have been there to be spoiling the wilderness. I'm definitely picturing pristine white snow and it's all beautiful and shiny and sparkly.

Benjamin Thompson

Well, guess what, Shamini, it is humans yet again. Okay. And this is some research that's been looking at marine environments near coastal research stations on the continent, specifically is looking at the marine sediment near Australia's Casey research station, which is located apparently in the Windmill Islands on the east of the continent. And it's been surveying samples taken from around this area between 1997 and 2015. And it's found, well, it's found a lot of stuff to be honest with you. It's found high concentrations of hydrocarbons, metals like lead, copper, zinc. Apparently high concentrations of something called polychlorinated biphenyls — these were highly carcinogenic, and were banned in 2001. And comparing this data to that from other waterways, they found that these levels of lead, copper and zinc were in some cases similar to those seen in parts of Sydney Harbour and the harbour in Rio de Janeiro in the past two decades.

Dan Fox

Do they know why this has happened? Is it just that people in Antarctica are not careful about what they drop in the sea?

Benjamin Thompson

I mean, it seems like you might not be that far wrong. So one of the people in this article say that in the old days, waste was often just dumped a small distance from research stations, and it's not limited to the Casey research station, you know, it's likely that this could be the case for a lot of research stations there. For example, New Zealand's Scott Base is currently being redeveloped and contamination from past fuel-spills, and again, poor waste-management has been detected in the soil and marine sediments there. And this is stuff that doesn't just disappear, right. So it could be there for a very, very long time before it's cleaned up.

Shamini Bundell

I mean, how big can these research stations be? Is part of it, that they have been there so long? And you know, they don't have good waste-disposal facilities to take stuff away. So they've just been sort of dumping it in an ongoing cumulative manner.

Benjamin Thompson

Well, there are a lot of research stations there actually, and mostly built on ice-free areas. And this ice-free area makes up less than 1% of Antarctica, but supports the highest diversity of plants and animals. But it seems that a lot of this contamination came quite a long time ago, because in 1991, an international agreement was signed to ensure nations monitored their impacts on the continent. This was adopted, I think, in 1998. But of course, a lot of these stations came from quite a long time before then, when this contamination occurred.

Dan Fox

Are there any plans to clean up the mess? I mean, I guess it's not very easy to get a cleaning crew to the Antarctic.

Benjamin Thompson

Well, it is clearly an issue and future pollution could be an issue too. But each nation is responsible for its own monitoring around its research stations. And practices do vary. So in this article, they talk about how some researchers are trying to develop processes to help station managers set objectives for reducing the impact of their facilities. Other groups have been looking at bioremediation using bacteria to remove hydrocarbons from the soil, others upgrading their waste-water treatment systems. So things are being done. But what effect this historical contamination might have, or is having, I suppose, on biodiversity is unknown. And of course, we talked about climate change earlier. And as things warm, there is the chance that these things could leach out of sediment and soil and into waterways and so forth. So it is an issue that needs to be addressed. But I said I had two stories about Antarctica. And so let's move on to the second one. And it's something I read about in Science from a few weeks back, but it's an interesting one. It's kind of a sad one. And yet again, it is climate change related, and it's based on a report in Communications, Earth and Environment. And it's about emperor penguin colonies on the west of the continent and how they're faring in the face of well, sea-ice melting.

Shamini Bundell

Are they doing fine? Is everything fine, Ben?

Benjamin Thompson

I'm afraid to say they are not doing fine. So let's have a little bit of an emperor penguin kind of refresh, for those of you who don't study their life-cycle, emperor penguins arrive at breeding colonies on sea ice, kind of March, April time, chicks hatch August to September, the chicks grow, develop mature, waterproof feathers by the end of the year, right. But 2022 saw record-low sea-ice and it had a devastating effects on several of these colonies and satellite imaging showed that four colonies abandoned their breeding grounds before December. And it seems to be that the chicks either drowned or froze because they weren't prepared for the icy waters.

Dan Fox

Wow. So what is the loss of generation of chicks from four breeding colonies mean for the penguins?

Benjamin Thompson

It seems that emperor penguins aren't in dire straits just yet, it has to be said. So one colony did okay. And in this area, there's about 10,000 breeding pairs, which is just a fraction of the quarter-of-a-million breeding pairs in total. And colonies have failed to breed before, but these are quite long-lived animals, so I'm sure they'll have another go. But given the trends in sea-ice melt around Antarctica, it's kind of an open question what the long-term future is for these five colonies.

Shamini Bundell

Well, okay, double bad news from Antarctica there. I'm gonna bring us back away from some climate change stories. I've also got a story for you today. Read this article in Nature. There's paper published in Science, and it's sort of on the way to robots who can smell that's how I'm going to pitch this to you.

Benjamin Thompson

Finally, okay, right. Smelling robots.

Shamini Bundell

What we’ve all been waiting for, yeah.

Benjamin Thompson

Yeah. Okay, what's going on with the journey to smelling robots?

Shamini Bundell

Well, okay, so it's not quite robots going around sniffing roses or tasting wine yet, but a group of researchers have trained an AI to describe smells using the same kind of words that humans would describe them based on their chemical structure. So they got a sort of shortlist of words to use. The words are great. The examples given in this article are ‘fruity’, ‘grassy’, ‘winey’ and ‘fishy’. I kinda want to know the full list. And then they sort of got both humans and then trained up the AI to match those words with, you know, a particular odorant, particular smelly compound.

Dan Fox

So, is this the end of my perfume-based during test?

Shamini Bundell

It did pretty well, although, I mean, what is well, what is correct when it comes to smells? You know, it's a bit of a subjective thing. So, in one way, you could say that, yeah, this AI maybe did better than humans in that what it actually did was provide words that were very, very close to the average response, given by all the humans, so often closer than any one individual had sort of guessed and described.

Benjamin Thompson

And to what ends then? Why do researchers want to train an AI to identify or describe a smell?

Shamini Bundell

So this could potentially be something that sort of becomes useful in designing perfumes, food products, things like that. So because it's an AI, it's going off the chemical structure. So it's sort of looking for patterns in the chemical structure of a particular compound, and then sort of determining the smell from that. So it can do this with smells that don't exist, you know, it's not actually smelling anything, it is taking the chemistry.

Benjamin Thompson

So you can just show it a molecule and it will say this one probably smells like, what, grass or something like that?

Shamini Bundell

‘This is a particularly grassy smell’. Exactly, based on this particular chemical feature, perhaps. But sort of potential use for this, this is sort of a long way away, is the thing about smells is we don't really know much about how we smell and how those chemicals interact with the sort of odorant receptors in our nose. A smell has this thing where it bypasses a lot of the brain and goes straight to the sort of memory and emotion centres in the brain, which is, you know why they say that a smell can like evoke a particular memory very strongly. But there's so many mysteries and you know, lots of people studying smell, and so it’s sort of interesting, but this group have bypassed the mechanism of how do we interpret a particular chemical as being a particular smell, and sort of got the AI to link the two via probability, that kind of deep-learning stuff.

Dan Fox

So you've talked about identifying one molecule and what it smells like. But I guess most things we smell are made up of mixtures of molecules and lots of different odours that combine in interesting ways. I mean, how does the AI deal with that?

Shamini Bundell

Yeah, apparently not very well. Most things that we're smelling are these, sort of, complex interactions. Coffee, as the example given us an article, contains hundreds of different of these odorant chemicals that the AI is sort of looking at one at a time. And again, with this sort of mystery of how our noses work, essentially, you know, how do they combine and sort of compete with each other? And how does like a mixture of smells feel to us completely different from any of those chemicals sort of would separately and individually? So the AI cannot do that yet. And in fact, one of the researchers is quoted as saying this article, “Predicting what a mix smells like is the next frontier”.

Benjamin Thompson

Well, a terrific story, and I'm sure the researchers will be keeping their noses to the grindstone to get to that multi-smell future. Right? Well on that awful pun, let's call it a day for this week's extended Briefing Chat. Listeners, for more on the stories and where you can sign up for the Nature Briefing, head over to the show notes, we will find all the relevant links. We'll be back next week with a regular edition of the Nature Podcast, but for the time being. All that's left to do is say thank you very much. Shamini and Dan

Shamini Bundell

Cheers, Ben.

Dan Fox

Thanks, Ben.