In this series, we explore the genetic evidence that indicates humans became a separate species as a substantial population, rather than descending uniquely from an ancestral pair.
In the last few posts in this series, we’ve explored Vern Poythress’s arguments in his short book Did Adam Exist? – specifically, his arguments against human – chimpanzee common ancestry. While we now turn to his arguments against the conclusion of human population genetics that our species does not descend uniquely from an ancestral pair, it’s worth noting, in passing, that Poythress’s arguments do not even attempt to refute the most obvious and straightforward evidence for common ancestry. For example, we see mutations in genes that form nested hierarchies in the human, chimpanzee, gorilla, and orangutan genomes. By far and away the simplest reason for this pattern is common ancestry, and as it turns out it’s the exact same pattern of common ancestry predicted by DNA sequence identity between these primates. We also see the remains of a gene in the human genome devoted to large-scale egg yolk production – something placental mammals such as humans simply do not need. If Poythress is to build a convincing case against human common ancestry with primates, these and other similar lines of evidence are the issues to tackle. As it stands, he has not even attempted to address them.
Having dealt with common ancestry to his satisfaction, Poythress then shifts to discussing human population genetics. As I see it, the three main arguments Poythress puts forward in this area are as follows:
Population genomics methods report only long-term average population sizes. These methods could not detect a bottleneck to two individuals, even if one existed, since they report only long-term averages.
Population genomics methods report sizes for more recent human populations, and do not address more distant history. As such, Adam and Eve may have lived further back in time than these analyses measure. This view is consistent with Scripture since the biblical genealogies may have gaps.
The findings of population genomics are based on uniformitarian assumptions that may not in fact be true if allowance is made for miracles.
While the third argument requires us to delve into the philosophy of science and how science might interact with theology, the first two claims are strictly scientific. As we will see, these scientific claims – like Poythress’s claims relevant to common ancestry – also fail to stand up under scrutiny.
Population bottlenecks and genetic variability over time
Poythress’s first claim, that population size estimates are long-term averages only and thus could not detect an original pair, appears a few times in the book:
Another paper uses genetic diversity among humans today to estimate average population size over the remote past, and offers nine different estimates in the region of 10,000. But these numbers depend on models that assume a constant population size through many generations. The figures are in fact giving us rough averages over long periods of time, so they say nothing about the possibility of two original individuals.
… the analysis always results in figures that represent a rough average over many generations in the human population. Consequently, the principal figures, like 3,100 for non-African populations and 7,500 for the African population, represent average populations over many generations. They say nothing one way or the other about whether the size decreased rapidly to two individuals in the more distant past.
This argument fails to understand how genetics works within a population. A population bottleneck to two individuals (the most extreme bottleneck that a mammalian population can experience) would not merely affect one or two generations, nor even a few hundred generations, but rather mark the genome of a species for hundreds of thousands of years. As such, it would easily be detected by modern population genomics methods. Let’s examine how such an event would shape genetic variability in a species.
The first effect an extreme bottleneck would have would be to greatly reduce genetic variability – the number of alleles that the species in question would possess. Recall that an allele is an alternative version of a DNA sequence. In mammals, each member of the species can have a maximum of only two DNA variants for any given DNA sequence – one allele inherited from mom, and one from dad. For a population to maintain high genetic diversity, you need a large number of individuals that have alternative DNA sequences to reproduce. If a population is reduced to only two individuals (or, as Poythress argues, was specially created starting with only two individuals) then the maximum possible DNA variation for any location in the genome is only four different alleles (two each in the male and female, with all four alleles being different from the others). Also, all alleles in this species genomes would coalesce to approximately the same time point, since all genetic diversity above this baseline (that would later accumulate through mutations) would be effectively the same age. However, as we have seen, humans are highly diverse genetically, and our alleles do not uniformly coalesce to a specific point in our history, but rather to a wide range of times, consistent with our lineage maintaining a large population over time.
A second effect of a reduction to (or start from) two individuals would be to place the original alleles into four chromosomal linkage patterns. These linkage patterns would slowly be broken up by recombination over successive generations. As with allele coalescence, we would see a pattern in present-day genomes that could be tracked back to four ancestral linkage patterns indicative of two original ancestors. These patterns in the human genome, however, indicate thousands of ancestors, as we have seen.
(Now it is of course possible - to jump ahead to Poythress’s third argument – to state that the patterns we see in the present-day human genome are the results of God’s miraculous intervention superimposed on what is in fact descent from an ancestral pair. I will address this possibility in a future post, but for now I’ll merely state that I do not find this argument plausible.)
So, if indeed humans descended from two individuals we would expect our present-day genetic diversity to be low, and the pattern of diversity to reflect an origin from a maximum of four original chromosomal arrangements. We do not, however, observe anything like this for present-day human genetic diversity.
Perhaps one way to appreciate that population genetics methods can (and do) detect extreme population bottlenecks would be to examine other mammalian species that do have the features we would expect for such an event. Cheetahs are one well-known example of a species with greatly reduced genetic diversity. Population genetics models for cheetahs support the hypothesis that they experienced a strong population bottleneck, which has recently been estimated at about 42,000 – 67,000 years ago, nearly eliminating their genetic variability. This signature remains with them to this day, since there has not been adequate time for new variation to arise through mutation. A second example would be the more recent bottleneck in Hawaiian monk seals that occurred in the 1800s, which is easily (and dramatically) seen in their extreme lack of present-day genetic variability. So, contrary to Poythress’s claim, the methods of population genetics are fully capable of detecting a reduction to (or start from) two individuals, since such an event would shape the genetic variability of that species for a very long time, well within the detection limit of current methods. The fact that our genome does not show these features thus remains a problem for his argument.
In the next post in this series, we’ll discuss Poythress’s second claim – that perhaps Adam and Eve lived long ago enough to avoid detection by current methods.