Deleted: Does Genetics Point to a Single Primal Couple?

This article was deleted without explanation from BioLogos in January 2020. This is a article of historical significance because it was the basis of an important article by Christianity Today in 2011, and it contains several major errors. The article was at this URL for nearly a decade:

Does Genetics Point to a Single Primal Couple?

By Dennis Venema and Darrel Falk on April 5, 2010

A Single Primal Couple?

Most Christians who have grappled with the science of genomics (the branch of biology that compares the DNA sequences of different organisms to one another) have done so with the question of common ancestry in mind: do humans share an ancestor with other forms of life, such as chimpanzees?

Here the evidence is very compelling, and reasonably accessible to non-specialists. For example, the human genome has numerous defective genes embedded it, and the vast majority of these defective genes are also present in the chimpanzee genome in the same relative positions with identical mutations. This sort of evidence is easily understood due to its qualitative nature.

A second question, and one that is less frequently explored even by Christians who accept common ancestry, is the issue of human/hominid population sizes during our evolutionary history. Specifically, is the human race descended from one ancestral pair in the recent past? Are we, as C.S. Lewis puts it in his Chronicles of Narnia , the “sons of Adam and daughters of Eve”? Is there genomic evidence to suggest that the human race is genetically derived from a primal pair? Here the evidence is more difficult for non-specialists to appreciate, because it is quantitative in nature.

Genomics can be used as an estimate of population sizes in the past by measuring genetic variation in the present. Genes come in different forms, or alleles: for example, the human ABO blood types are determined by three alleles of one gene. Some genes in human populations exist in hundreds of forms.

The catch, however, is that any individual person can only carry at most two different varieties of any one gene: one from mom, the other from dad. It therefore follows that a large population can pass on a large number of gene forms (alleles), but a population that passes through a population “bottleneck”—where only a small number individuals survive—will fail to pass on most of its genetic variation to future generations.

Attempting to square the Genesis account and common ancestry by positing a literal Adam and Eve who were the progenitors of the entire human race is, biologically speaking, looking for the most extreme population bottleneck a sexually reproducing species can experience: a reduction to one breeding pair.

Is there evidence that such a bottleneck has ever occurred? Dr. Peter Enns has been exploring whether this is even the right question to be asking from a biblical perspective ( here , here , here , and here ). Here we explore three independent ways of answering the question, this time from a biological point of view.

Method I:

The genetic consequences of a bottleneck required by a literal reading of Genesis 2-3 would be severe: at maximum, four gene-forms (two from each parent) would be passed on by Adam and Eve. Interbreeding in the (necessarily very small) population after the bottleneck would result in the further loss of some alleles due to chance alone. In short, the genetic impact of such an event would leave a stamp on the genome of that species that would persist for tens of thousands of generations as mutations slowly generated genetic diversity.

We can use this information, then, to estimate the minimum number of people that could have existed at any point in time. First we ask how many different alleles there are for a number of genes within the current population. Correcting for the rate at which we know new forms of genes appear (mutation), we can calculate the minimum number of people needed to generate the current amount of diversity. Numerous studies analyzing many different genes all point to a bottleneck. However, these studies are all clear: during the bottleneck, there were several thousand individuals, not two.

Method II:

In earlier posts, we have discussed the fact that DNA segments known as Alu repeats, can insert themselves at various locations in the genome. It turns out that the Alu sequence comes in various forms, like different makes of cars—Fords, Toyota, etc. There are several thousand families of Alu.

Consider just one family, which we will call Ya5. Members of this family have been inserted into human chromosomes at 57 mapped locations. If all humans descended from a single pair of individuals, all humans would have each of the 57 elements in pretty much the same locations, since individual members of the family almost never move. However, the human population consists of groups of people who share some insertion points but not others. The multiple shared categories make it clear that although a human population bottleneck occurred, it was definitely never as small as two. In fact, this line of evidence also indicates that there were at least several thousand people when the population was at its smallest.

This method is much different than Method I since it does not depend upon mutation rate, but the answer is similar.

Method III:

A third independent estimate makes use of a concerted research effort called the HapMap project. Humans have 3 billion bits of information in their genomes. (The official term for one bit is a “nucleotide.”) The bits between any two individuals differ at many sites, which is, of course, why we don’t all look the same.

In the HapMap project, one million of these differences have been analyzed by examining something called linkage disequilibrium The technical details are beyond the scope of our discussion, but to give you a feeling for how it works, imagine that you have a gene for blue eyes and a gene for a bent finger, both of which you inherited from your dad. Assume these genes reside in the same “neighborhood” on chromosome 2. Because these genes are close to one another, chances are that if your brother got the blue eye allele from your dad, he would have received the bent finger allele as well. After all they are neighboring genes, both on chromosome 2. Why? Blocks of genes in the same neighborhood on a chromosome are usually inherited together. Alleles that are very close together on chromosomes tend to stay together for many generations before they are “mixed and matched” through a process called recombination.

Now pretend that someone analyzes both your DNA and that of your brother in a double blind experiment. The investigator would, upon examining the results, be able to say, “I’ll bet these two people are related to each other.” And he would be right.

Now picture being able to do this, not for two differences, but for a million differences all at once and not just for two people, but for many people from all over the world. Using this approach, it is possible to tell how many people gave rise to all the prevalent combinations of differences. In short, we can tell if everyone came from just two people at any time in the last 200,000 years. So did we?


This third independent method tells us that everyone alive today is related, but not to a single pair of people. We are related to a population that consisted of several thousand people with their several thousand combinations of these million genetic differences.

Here’s the real point of this. When you have one way of doing a calculation and you get a certain answer, perhaps you are justified in being a little skeptical. Perhaps you made a mathematical mistake, or maybe you made a faulty assumption. However, when you do your calculation using two totally different approaches, using methods with completely different assumptions, and each method gives you the same answer, you become convinced it is correct. Three, of course is just icing on the cake.

So that’s the situation we are in with regard to the human population size in ancient history. There was a bottleneck. There were likely fewer people alive during that time than the number of fans attending a typical NHL hockey game. (We don’t know if they were all together in one village, of course, but the total was small.) However, it was not two people. Our species diverged as a population. The data are absolutely clear about that.

I’ve heard this assertion before, that Dennis Venema made some important errors. Did he? Where can I find a listing of those errors? I know that at one point I made a comment, that the methods he had cited were not as powerful as they could be. And I think that this comment was misinterpreted to mean that they were wrong. They weren’t, as far as I could see, just that with development of additional computational methods even more power could be squeezed out of the data, and his inferences could narrow down the possibilities even more. Also, is the argument against his conclusions that he could not rule out a special creation that loaded poor A&E down with massively excessive amounts of heterozygosity? Population geneticists estimating ancestral population sizes would never entertain such a possibility, and I would support them in that. Have I missed something important here?


@Joe_Felsenstein, I’ve explained some of them here: Three Stories on Adam, without attributing his quotes to him by name.

Yes, he made several claims outside mainstream population genetics. He still makes some of these claims.

1 Like

It seems to me that the evidence from amounts of heterozygosity and survival of haplotypes, evidence that is in papers by others, does rule out a recent (say less than 10kya) sole genetic ancestor who is the origin of all our genetic material. I am not talking about your GAE or any non-sole-ancestor in saying this. And any population geneticist would dismiss any argument that loaded sole-genetic-ancestors-A&E down with crazy amounts of heterozygosity. One infers past population sizes by assuming that those people were normal human beings.

1 Like

And thanks for the references.

1 Like

I totally agree that a sole genetic progenitor less than 10 kya is not consistent with the genetic evidence. However, we have to be able to show our work. The conclusion here, of course, we all should agree with you on. However we have to demonstrate that conclusion with accurate arguments.

One valid argument against this is the time to Y-adam. That rules out a single couple bottleneck more recent than about 150 kya.

Making the case based on estimates of effective population size, however, relies on equivocating minimum and average. We would need to do some work to demonstrate that a bottleneck of two can be picked up by these methods. For obvious reasons, that is not usually done in the literature. And when we did look at it we found that brief bottlenecks are difficult to pick up.

Venema also went so far as to claim we had strong evidence against a bottleneck going back 18 million years. That claim does not appear anywhere in the literature and fell apart very quickly.

1 Like

Methods I and III are valid approaches to demographic inference (even if the results are overstated here), but I don’t understand Method II. Alu insertions are just biallelic variants; I don’t see any reason they should be more informative about demography than 57 randomly chosen SNPs. Am I missing something?

1 Like

@glipsnort what references do you have to Method 1? It is not referenced in this article, and I’ve heard at least three different interpretations of what it means. Different interpretations came, in fact, from the two authors!

There was a paper somewhere that references using alu insertion sites and mutations, I think, to compute effective population size. It isn’t referenced here, though, so we can’t be sure from this alone.

The introduction also contains several consequential errors.

It’s written in broad enough terms that it could refer to a lot of different methods, and I don’t know which specifically he had in mind or what references would be appropriate.

1 Like

Wouldn’t “method 3” rule out the pre-loaded heterozygosity argument, as well?

Both Method 1 and Method 3 rely on population genetics models that deal with normal reproduction and normal mutation. They therefore do not even consider having bizarre individuals who have vastly more heterozygosity than can occur with ordinary models of reproducing and mutating populations.

One doesn’t need to posit exotic biology to see the problem with relying on effective population size estimates.

Method 1 is not well defined. It is also unreferenced. We don’t know from this document what Method 1 could be, and I’m not sure there is any method that fits this class that yields a minimum population size of 1000s.

The description of Method 1 is unclear, but I take it to be the use of standard population-genetic inference methods using SNP observations and the mathematics of coalescents. Coalescent population size estimates of past population sizes started in the 1990s. There are Markov Chain Monte Carlo methods and more approximate but more tractable approximation methods using Approximate Bayesian Computation. Here is a paper mentioning many of the methods in current use:
Inferring Past Effective Population Size from Distributions of Coalescent Times

Method 1

Yes, I am aware, but that isn’t what it says. There are no mention of coalescence.

Well, that is not true…

I do not know of any method that is based on "how many different alleles there are for a number of genes within the current population.”

One interpretation is that this refers to trans-species variation, which does in fact give a minimum bound. However, (1) that doesn’t match the methodology referred to here nor (2) does it produce a minimum estimate of thousands.

One interpretation is that alleles really means LD blocks, and LD linkage disequilibrium. Once again, we don’t get coalescence from counting alleles, so that doesn’t fit the methodology described here. Also, it doesn’t match the definition of allele given earlier in the paragraph.

One interpretation is that it matches the methodology from the abstract:

Once again, this does not match any published methodology of which I know. Discussed here is not coalescent rates but the number of alleles (not divergence) as a measure of variation.

It seems that the best case is that this is a phenomenally bad description of LD analysis. However, we actually saw one of the authors explicitly pressing it as a “there are too many alleles!!” argument. He did not discuss coalescent rates, and literally meant “too many alleles.” So I’m not sure how to read in manner consisten with anything I’ve read published in population genetics.

Perhaps @Joe_Felsenstein or @glipsnort know of some approaches of which I am unaware. I asked both authors to tell me, but they did not know of one.

The Larger Problem

Look at the question statement:

There is a conflation of three questions, and the difference is consequential.

  1. Specifically, is the human race descended from one ancestral pair in the recent past?

This is ambiguous, and could be yes or no, depending on what is meant. If we mean genealogically descend from single particular pair in the recent past of theological importance, along with others, then “yes.”

If we mean genetically derive all our DNA from a single pair (a claim that no creationist camp I know of insists upon), well then “no.”

  1. Are we, as C.S. Lewis puts it in his Chronicles of Narnia , the “sons of Adam and daughters of Eve”?

Well, as the GAE shows, the answer is “yes,” if Adam and Eve exist, we all descend from them, even if they are in the recent past. We are all the “sons of Adam and daughters of Eve,” if they are real people in a real past.

  1. Is there genomic evidence to suggest that the human race is genetically derived from a primal pair?

This makes two major equivocations.

Whether there is evidence for a genetic bottleneck is different from whether there is evidence against it. Perhaps the data does not tell us either way. The case in this article, however, is that there is strong evidence against a genetic progenitor pair. The evidence against such a recent genetic progenitor does exist, but you won’t find it in this article. That doesn’t mean that there is evidence for a recent genetic progenitor, which is a which is a separate question altogether.

Second, genetic progenitorship as understood by creationists does not rule out a large population size because it doesn’t rule out interbreeding injecting DNA into Adam and Eve’s lineage. What they mean might be roughly stated, “all of us get the DNA required for our humanness by direct descent from Adam and Eve.” This doesn’t rule out interbreeding, so it does not require a genetic bottleneck in the first place.

With those points in mind, an ancient genetic progenitor is entirely viable. It depends on the details of course, but this article created an immense amount of confusion by pressing an argument that was not valid.

Recent sole-progenitorship (genealogical, with people outside the garden) also is entirely viable.

The basic conclusion of the article ends up being incorrect.

Method II

Seems this is an example of the method referred to here. How reliable is this approach?

That’s not what anyone reading the question “is the human race descended from one ancestral pair in the recent past?” would think it meant. It’s not at all ambiguous, and your alternative verges on sophistry.

Not one? Not even Ken Ham?


Agree to disagree. But if you disallow my theologically informed reading, that accentuates my point nonetheless.

Remember the conflation at play:

This question is not about genetic descent, and it matches how it is normally a phrased in theology. If your reading of the first question is right, it is even more invalid to equivocate or conflated it with this second question. They are different questions in your reading, with different answers.

Ken Ham, as I’ve show both in the forum and the book, allows for (and even celebrates) the possibility of nephilim outside the garden. It’s in the Ark Encounter :slight_smile: .

Henry Morris openly speculates about people outside the garden too, albeit it disturbingly racist ways.

Well, now, that raises the question of how Lewis would have meant it. Are you sure?

But do the nephilim contribute to the human genome, and if so do they contribute more than a small fraction, the rest being from A&E? I’d consider that a minor variation on A&E as sole ancestors. Ditto for Morris, whatever he’s getting at. And wouldn’t all those folks have died in the Flood anyway?

1 Like

I think, whatever Ken Ham and Co. say, that a lot of people in the Judeo-Christian tradition think that what Genesis means is that A&E were the sole genetic ancestors of all humans. And that when they were created, there was not already a crowd of Nephilim hiding nearby. Pardon our misunderstanding of Genesis.