In 2023, a controversial study introduced a dramatic twist to the story of human history by suggesting that our species came to within a whisker of extinction almost a million years ago. Published in the prestigious journal Science, the research made headlines around the world, yet more recent examinations of the data suggest that our ancient ancestors’ dance with death may never have occurred.
To conduct their analysis, the authors of the original study developed a computational model called FitCoal, which they used to analyze the genetic history of more than 3,000 present-day individuals by tracing their mutations back in time in order to determine the population dynamics that could have given rise to the current distribution of these genetic variants. Among all African genomes included in the sample, the researchers detected signs of a population “bottleneck” some 900,000 years ago.
Based on their findings, the authors suggest that around 98.7 percent of the ancestral human community died out during this period, leaving global population numbers dwindling at just 1,280. The fact that this same bottleneck was not detected in non-African genomes, however, has never been satisfactorily explained, leaving many experts questioning whether the supposed crash actually occurred.
The reality of what happened was a very, very complicated thing – much more complicated than our simple models are going to be able to represent.
Dr Aylwyn Scally
Among those to challenge the findings is Dr Aylwyn Scally from Cambridge University, who told IFLScience that while FitCoal may have indicated an ancient population collapse, other models failed to replicate this signal. “All of these methods rely on differences between present-day people and differences in ancestry in order to infer things about the past, but actually by the time you go back hundreds of thousands of years, most of those differences have coalesced into a single ancestor,” he said.
“So present-day genetic differences are very uninformative about anything beyond about 200,000 years ago.”
Among the many statistical tools that have been created to try and reconstruct past population sizes are the likes of mutation spectrum history inference – aka mushi – as well as the multiple sequentially Markovian coalescent (MSMC) and various others. Despite all being slightly different, these models each rely heavily on assumptions and simplifications relating to mutation rates in order to calculate when and how this historical coalescence occurred.
According to Scally, these unavoidable presumptions limit the reliability of all such tools, since “the reality of what happened was a very, very complicated thing – much more complicated than our simple models are going to be able to represent. So there are lots of ways you can go wrong, lots of ways you can be misled.”
Suspecting that FitCoal may indeed have made a fundamental calculation error, several scholars have conducted their own analyses in the last few months, with damning results. For instance, one study – which has yet to be peer-reviewed – revealed that the proposed bottleneck is in fact impossible to detect using mushi, while another found that other established models also fail to pick up this population crash.
It’s currently unclear exactly why this is so, although the authors of the first paper say it may have something to do with FitCoal using a different coalescence rate to the other tools.
Yet another paper, published this week in the journal Genetics, concluded that FitCoal is in fact too simplistic to accurately model ancient population dynamics, which may explain why all other models have produced contradictory findings. “FitCoal artifactually tends to infer a sharp bottleneck when there in fact is none. In other words, the reported severe bottleneck is likely a statistical artifact,” write the study authors.
The researchers also say that if our ancestors really did come so close to extinction, then a similar trace should be present in the genomes of all modern non-African populations. The fact that FitCoal didn’t see this only adds to the suspicion that its findings may be erroneous.
What’s more probable for most people in the field, is that their method is responding to some aspect of the data in a particular or unusual way, which is probably an artifact rather than the truth.
Dr Aylwyn Scally
Discussing the work of the researchers who first proposed the population crash, Scally says that “if your method is the only one that detects a particular signal, you need to have some explanation for why that might be the case. You need to explain why it is that your method is able to see this and nobody else is. But they don’t manage to do that in their paper.”
It’s this lack of clarification that has left so many researchers doubting the reliability of FitCoal and therefore suspecting that the supposed population bottleneck may simply be a statistical artifact – or error.
“What’s more probable for most people in the field, is that their method is responding to some aspect of the data in a particular or unusual way, which is probably an artifact rather than the truth,” says Scally.
Source Link: Did Humans Almost Go Extinct 900,000 Years Ago?
