<link rel="stylesheet" href="//fonts.googleapis.com/css?family=Roboto%3A300%2C400%2C500%2C700%7CRoboto+Slab%3A400%2C700">DNA study confirms Noah because it gets the science wrong

Answers in Genesis recently published two studies by Dr Jeanson in their scientific journal. Both focus on how evolution since Noah and his ark. Both are titled “On the Origin. . . “. And the results of both – they claim – support the creationist point of view.

The first investigated whether evolution could have happened rapidly enough to explain the emergence of all the species since Noah. This revealed it couldn’t; a fact which was explained away with miracles. The second paper is what interests me today. It focuses on humans and whether the diversity present in our species is consistent with the Noah story. 

This research has two main conclusions. The first is that all humans’ mtDNA fall into three groups. This matches Noah’s three children – each of whom is thought to have been the ancestor of one group of people (explaining why non-whites should be slaves). The second conclusion is that the number of mutations present in our mtDNA is consistent with the Biblical timeline.

Both of these conclusions would seem to be fairly major victories for the creationists. And several creationist organisations are already touting them as such. The problem is that they’re based on bad science. And yet this all got through creationist peer review.

Unrooting a tree

The first major contention of Dr Jeanson’s latest paper on Noah is that our mtDNA is consistent with the Biblical narrative. This was calculated simply enough; by drawing a family tree of 369 humans’ mtDNA.

The resulting tree has three major groups. Noah brought his three kids (and their wives) onto the ark. Under the creationist narrative, each is thought to have given rise to one branch of humanity. Thus the presence of three groups – matching this story – would vindicate the creationist model.

Dr Jeanson's claim that these results are consistent with Noah

Dr Jeanson’s claim that these results are consistent with Noah

Already things are starting to go wrong for Dr Jeanson as he doesn’t seem to have done anything to vindicate this conclusion. There’s no statistics demonstrating that the mtDNA he’s studying falls into these clusters. He seems to just be eye-balling it from the tree. “Eh, it looks like there’s three significant nodes”. Hardly the most reliable methodology for vindicating genesis.

The arrows point to the three clusters of Noah's wives

The arrows point to the three clusters of Noah’s wives

Now, some of you might have noticed this tree doesn’t look exactly like the traditional family tree you might be familiar with. This is because it’s unrooted. To “root” it you have to add in a branch that isn’t part of the group you’re studying (so for humans, the outgroup might be chimps). This puts the entire tree in context and allows you to construct the rooted family tree you’re more familiar with.

An unrooted and rooted tree for the apes

An unrooted and rooted tree for the apes

This is where the first major problem with Dr Jeanson’s paper is: you need to root a tree to study ancestry. An unrooted tree simply tells you how similar the individuals in it are. It can’t tell you what they’re ancestry was; who descended from whom and so forth. For instance, under the unrooted ape tree all of the apes could have evolved from humans. There’s no information about the direction of evolution in an unrooted tree. Or for Dr Jeanson’s tree, one group of women could have given rise to the other two; rather than all three co-existing as Noah’s wives.

Yet Dr Jeanson is trying to study human ancestry and whether it matches with the three wives of Noah. That’s something that can’t be done from this tree. At best he could say humans fall into three similar clusters. Except as his chart shows, there’s more than three clusters involved (with no statistics to narrow it down to three).

What’s more, I’m really curious as to whether or not it would be even possible for a creationist to vindicate the Noah story using this method. After all, you need an outgroup. A distantly related population to put your tree in context. Yet they claim humans are unique, distinct from the rest of the world. Is there even an outgroup for them to compare this data to?

So not only does this not show what Dr Jeanson thinks it shows, it’s impossible for it to do so.

Mutations since Noah

Basing conclusions on the wrong type of chart seems bad. To make matters worse, it’s a chart that can’t support your conclusions at all. However, things go off the rails even more when it comes to Dr Jeanson’s second hypothesis.


At least I’m finding this out having just started my PhD. I won’t have wasted too much time then.

This is the claim that the number of mutations in human mtDNA is more consistent with the Biblical narrative than the evolutionary one. This was calculated by examining the number of differences between our mtDNA. The number of mutations per generation was then used to figure out how many generations were needed to produce this number of differences. This was then calculated into years by using UN data from Africa (in the 70s) on how long generations are.

It turns out that if humans have been reproducing like Africans from the 1970s then the number of differences between individuals is consistent with the flood timeline. Unlike the evolutionary estimates.


Because creationist science is real science, it doesn’t need error bars

But before you give all your money (or your house) to Answers in Genesis, we need to dig a little deeper. It turns out that the diversity present in a genome isn’t just the result of time x mutations as Dr Jeanson would have you believe. Several other factors are involved.

One of the biggest being evolution. Kind of ironic, isn’t it? For example, think about what would happen if one of these mutations was harmful to the mtDNA. The individual with it wouldn’t survive as well, eventually being out-competed (and that mutation lost). So you might expect a lineage to have accumulated 10 differences from another, based on the rate of mutation. But in reality the actual number might be much lower if some of those changes were harmful and purged from the genome. A similar effect would happen if the mutation was beneficial. It would outcompete others, driving them extinct and reducing the overall diversity in a population.

Some mutations die off, decreasing diversity in the population. In this initial population there are 3 different types, but by the end there are only 2

Another big factor is population size. The general idea is that the more reproduction that goes on in a population, the more mutations will be passed on to each subsequent generation. A larger population can also “store” more diversity. For example, using Jeanson’s numbers, if you were to compare the mtDNA of me and Bob, there would be 123 differences between us. And if you compared Bob and Sally, there would be 123 differences between them. The key issue is that they could be different differences. There might be a separate set of 123 differences between Bob and Sally than there are between Bob and me. This would produce up to 246 mutations separating all of us.

nucleotide diversity

To simplify it, here’s a bit of all our mtDNA. Each person has 1 difference. Yet there are 3 mutations contained in the population, because everyone’s difference is different.

Ultimately there are upwards of 7,000 mutations in human mtDNA; with each person having a maximum of 123 of them. Is this consistent with Dr Jeanson’s predictions? That all depends on how big the population was. Different sized populations (with different levels of reproduction) would take different lengths of time to accumulate the 7,000 mutations and the resulting 123 differences between people. Now, to be fair to Dr Jeanson, there is some debate over how big a role population size plays in genome diversity, with there being some evidence for it and some evidence against it. Ultimately, it seems like it does play a role but a relatively small one. Nevertheless, that’s a factor that should be investigated yet another factor ignored by the creationist.

And those aren’t the only two factors involved. There are countless others. All of this mean there’s often relatively little correlation between the raw rate of mutation and the resulting diversity in the population. For example (in the nuclear DNA), flies and humans have a similar mutation rate. But the resulting diversity in our genomes differs by an order of magnitude. Yet the entire conclusion of this paper relies on their being a link between these two values.

Or – in simpler terms – the creationist model is nonsense. And this paper is too.


A creationist claim DNA supports the Noah story, but it only does so because he uses the wrong figures and data to support his conclusions.

Related posts

Categories: Creationism


Natural Historian · 17th May 2016 at 2:52 am

Thanks for looking into this. Yes, Jeanson and Tomkins are making it a habit of just plain horrible analyses. To top it off he seems to completely forget that his organization insists that Neanderthals are post-flood humans and so he needs to include their differences as well but makes no mention of them and doesn’t include them in his unrooted tree! Neanderthals would be around 200 to 250 pairwise differences and Denisovans 350. Are they no descendants of Noah and his family?

    Adam Benton · 17th May 2016 at 9:26 am

    Another interesting contradiction with AiG’s position is the implication that all mutations in the mtDNA since Noah are still present there. Hence why the prediction and “observation” line up so nicely. Which would mean that there have been no deleterious mutations natural selection has purged. Every mtDNA mutation must have either been neutral or beneficial to ensure they would all be preserved in the present. I’m not exactly sure that’s a view of mutations Ken Ham would agree with.

    I have heard some creationists (including a few from AiG iirc) try to explain Neanderthals and Noah via that same heterozygote stuff they were using to explain rapid evolution. Noah’s wives had a human chromosome, but some also had a Denisovan or Neanderthal one thrown into the mix. This made the wives “human” whilst allowing the Neanderthal lineage to survive the flood.

cynicalme · 17th May 2016 at 1:29 pm

I’ve been pondering Jeanson’s results for months but I haven’t had the time to look into them properly.

How does he come up with the red bar (the predicted diversity based on the evolutionary model)? Doesn’t he look at the current, modern day mutation rate in mtDNA and then use a formula like this to calculate the time that has passed since mitochondrial eve?

If he has done this then his method looks correct. He should be looking at the nucleotide diversity between the two most distant human individuals as opposed to the total number of mutations that should exist in a population – this would allow him to calculate the number of generations required for those two individuals to share a matrilineal ancestor.

So I don’t think your criticism is correct.

I think the problem with his paper is that rates of mutation vary wildly depending on the region of the mtGenome from which they are measured and the size of the population. The other problem is the fact that measured mutation rates are expected to be higher than rates inferred from phylogenetic methods for reasons set out here.

I’ll use that formula I linked to to come up with different answers based on different mutation rates.

Here are some typical mutation rates for different mt regions: http://i.imgur.com/xgyEIfC.png

Taking the rate 1.6E-7 (the rate for one of the hypervariable regions – way too high), we calculate that the maximum divergence for 2 individuals should result in about 1059 differences.

Taking the rate 2.4E-8 (the last rate shown in the table), we calculate that the maximum divergence for 2 individuals should result in about 158 differences which is fairly close to what we find today.

Taking that same rate (2.4E-8) and plugging in 10,000 years (in an attempt to recreate his YEC result), I calculate that the maximum number of between two individuals should be about 8… I have no idea how he came up with 123!

    Adam Benton · 17th May 2016 at 1:42 pm

    You’re giving Dr Jeanson far too much credit. He figured out the number of mutations in each generation (a figure I think he got from previous work he published for AiG). This was then converted into mutations per year (using information on generation length from Africa). This figure for mutations per year was then multiplied for 200,000 for the evolutionary model, or ~4,000 for the creationist one. This gives the estimates of the number of mutations discussed in the paper. That’s what produces the problem I pointed out. He’s comparing the raw number of mutations to a measurement of relative number of mutations.

    You are right in that one way to circumvent this problem would be to compare the two most distantly related individuals. In that scenario the relative and raw numbers should be identical. However, he doesn’t do that. He notes that the most divergent individuals he identified were San and Taiwanese Aboriginals. But does no further examination of them beyond that.

      cynicalme · 17th May 2016 at 1:56 pm

      Adam… His number (Green bar – 123) is the maximum pair-wise mtDNA difference count which has been observed and documented between two individuals. This seems correct to me.

      What is incorrect is his mutation rate of 0.197 mutations per generation. This is way too high

        Adam Benton · 17th May 2016 at 1:59 pm

        It is correct, as far as I can tell, being lifted directly from a paper on the subject.

        cynicalme · 17th May 2016 at 2:10 pm

        Well I say it seems correct to me because it would imply a mutation rate of about 1.85E-8 which is in line with what Soares has published:

        Correcting for Purifying Selection: An Improved Human Mitochondrial Molecular Clock

        Adam Benton · 17th May 2016 at 2:56 pm

        And that’s all fine. My point was simply that trying to infer nucleotide diversity from mutation rate alone is dubious to say the least. After all, humans and flies have a similar mutation rate yet vastly different levels of nuclear diversity.

        However, it’s clear I didn’t do a very good job of explaining all this. Thanks for bringing this to my attention. I’ll be trying to rewrite this post over the next week to clear it up a bit.

        cynicalme · 18th May 2016 at 9:43 am

        Adam… perhaps I’m being dense (has been known to happen), but I can’t see any place where Jeanson deals with nucleotide diversity.

        The numbers in this graph deal with maximum genetic distance (or predicted maximum genetic distance) between individuals.

        His green bar was labelled misleadingly – he shouldn’t have used the word “diversity”. When you have a look at his description for that diagram it becomes clear that he actually means maximum genetic distance, not diversity. He writes:

        “Both predictions were compared to the maximum pairwise mtDNA difference in the published literature (green bar)”

        Adam Benton · 18th May 2016 at 11:36 am

        I was under the impression that nucleotide diversity typically refers to the average diversity per site. Of course, that means there is actually a range of diversity within the population. Jeanson seems to have picked out the highest number of that range from:

        Kim, H.L. and Schuster, S.C., 2013. Poor man’s 1000 genome project: recent human population expansion confounds the detection of disease alleles in 7,098 complete mitochondrial genomes. Front. Evol. Popul. Genet, 4(13), pp.10-3389.

        cynicalme · 18th May 2016 at 1:38 pm

        Yes that sounds like a reasonable measure of nucleotide diversity to me but as far as I can tell, Jeanson doesn’t use this concept – rather he is trying to find the maximum genetic distance between two individuals in a population which seems to be a reasonable way of determining the distance to their LCA (mitochondrial Eve)

        Adam Benton · 18th May 2016 at 1:46 pm

        Either way, I think the point still stands that this value – whether average diversity or maximum diversity – can’t be inferred from mutation rate alone.

        Clearly though there’s a lot of rewriting needed to clear up some of this points. I should be able to get around to it today.

    cynicalme · 17th May 2016 at 1:48 pm

    Note that I made the following assumptions in my calculations above:

    Time since mtEve: 200,000 years
    Time since the flood: 10,000 years
    Number of positions in the mitochondrial genome: 16,553

    Looking at Jeanson’s article, he takes the mutation rate to be 0.197 mutations per generation which corresponds to 5.95E-7 mutations per nucleotide per year. This is an absolute joke and is far in excess of mutation rates described by Soares et al. (2009). This is probably because as I alluded to in my previous comment, he is using a direct method to calculate his mutation rate (e.g. counting differences between mothers and daughters) and this is expected to be higher than a mutation rate based on phylogenetics.

      Adam Benton · 17th May 2016 at 1:53 pm

      That hints at another issue I mentioned in response to a previous comment: Jeanson is assuming that all mutations are neutral/beneficial (although not explicitly making this assumption). Deleterious mutations would eventually be purged. As you said, this would make the “net” number of mutations lower than the “gross” number. Yet his work relies on the assumption those two values are equal.

      Rather ironic to see a creationist arguing all mutations are beneficial (or at least neutral). It seems to be the opposite of the position they like to take.

        cynicalme · 17th May 2016 at 3:45 pm

        Okay… I think I’ve cracked it and worked out where Jeanson’s crazy mutation rate comes from.

        Looking again at the Soares paper, the fact that he used a direct method doesn’t even account for his ridiculous mutation rate. Here is how the overall mutation rate should drop off over time. So using a direct method he should have come up with a mutation rate of about 2.5E-8. Instead he was using a mutation rate of around 5.9E-7 which is 23 times higher!

        If you have a look at this table from Jeanson 2015b you will see that his mutation rate is based on a single study on 333 pedigrees from an island nation – Sardinia. These Sardinians had a new homoplasmic mutation popping up every 5 generations! Compare this to the fact that we only expect a new mutation to popup once every 3624 years (or every 180 generations) and that explains the difference.

        Here is the paper from which they pulled these figures. They got their figures from table 1

        There is a huge problem here:

        This table only lists mitochondrial sites that differ from the reference genotype across 49 specific positions so they didn’t look at the entire mitochondrial genome for positions where mother and child might differ. But this is exactly what Jeanson has assumed! In his table which I link to above he claims that they examined a total of 5.5 million base pairs. They didn’t – they actually examined 49 sites per individual – which comes to a total of 16,317 base pairs. So if Jeanson was being honest, he would recognise that they found 63 mutations in 16,317 base pairs which comes to a mutation rate of 0.0038 per nucleotide per generation or 62 new mutations per generation. Bwahahaha!! Forget young earth creationism. This would blow all of our estimates out the water and imply that the world was 40 years old!

        Here are the 49 sites they looked at

        Clearly there is an issue with the data from Ding et al 2015. Either Jeanson and I have failed to understand it correctly or there were sequencing errors (a fairly common occurance) and these sequencing errors explain the majority of the 63 differences they found between mother and child.

        Adam Benton · 17th May 2016 at 3:56 pm

        Oh boy that is rich. This post clearly needs a follow up.

        cynicalme · 17th May 2016 at 10:02 pm

        Correction: there were 829 variants looked at altogether across the 999 participants – not 49.

        I will need to look into this further to determine whether this includes all possible variants (did they sequence the mt genomes in full?) or whether this study just looked at a subset of the possible variants (did the study involve the use of a microarray to pick up specific variants)

        Do be too quick to publish a follow up… I may be mistaken here

        Adam Benton · 17th May 2016 at 10:22 pm

        Your take on this is rather deep and fascinating. I’ll be sure to hold off until you’re all done. Speaking of which, it would be great if you wanted to put some of this into a guest post or segment for that follow up

        cynicalme · 17th May 2016 at 10:59 pm

        Sorry Adam, it looks like the Ding study did sequence full mitochondrial genomes as opposed to running the samples through a microchip array. When I initially looked at the variant datasheet in my browser it only listed 49 variants from all individuals tested but when I later downloaded the sheet, it listed 829.

        This means that the numbers quoted in this table probably do reflect the actual number of variants found in mothers and children.

        Having said that, I have thought of another serious issue that stems from interpreting the data the way Jeanson has.

        The table tells us that there were:
        – 7273 homoplasmic variants in found in 333 children
        – 7266 homoplasmic variants in found in the 333 mothers of those children
        – Only 7238 homoplasmic variants were common to mother and child

        Jeanson took this to mean that 35 child variants were not found in the set of their mothers and that 28 mother variants were not found in the set of their children.

        But this would be incorrect!

        The first line of this table (which he draws his data from) is only considering homoplasmic variants and not heteroplasmic variants. It is quite possible that those 35 homoplasmic child variants were not found in mothers because those corresponding locations in mothers were heteroplasmic. Likewise it is quite possible that those 28 homoplasmic mother variants were not found in children because those corresponding locations in children were heteroplasmic due to a subsequent mutation that had happened in one or more mitochondria (these samples were taken from somatic lymphocytes).

        The reality is that this data does not tell us how many homoplasmic loci in mothers were homoplasmic but different in children and vice versa. So there is nothing that we can conclude from this study regarding homoplasmic differences between mother and child or mutation rates.

        Adam Benton · 18th May 2016 at 11:37 am

        Perhaps it’s for the best then that I avoided writing up an instant reply

Natural Historian · 18th May 2016 at 4:27 pm

Fascinating discussion here. I appreciate the time spent cynicalme on this. I have spent much time looking at some of the same articles but everyone reports numbers in different ways so it is so hard to find apples-to-apples comparisons of numbers. Observed mutation rates and real mutation rates are hard to communicate to the public much less even figure out in the scientific literature. I got very frustrated yesterday. There is a new paper on the genomes of the giraffe and okapi.They report overall differences of synonymous (dS) and non-synonymous (dN) differences between the genomes and that of cows but even in the supplemental information they don’t provide enough information for me to calculate a raw number of nucleotide differences in the protein coding genes which is what I wanted to do to compare to other species within “kinds.” Creationists use the difficulty for anyone by experts to do genetic comparisons to their advantage because so few can figure out why they are wrong. So, again I appreciate this discussion, it is very helpful even if it still isn’t clear just what Jeanson was thinking.

    cynicalme · 18th May 2016 at 7:59 pm

    I’ve written up a post which I will hopefully be publishing to this blog as a guest. I believe I’ve now explained what Jeanson was thinking. I’d like to get his response to this but that’s probably a pipe dream.

Ashley Haworth-roberts · 19th May 2016 at 3:22 pm

The ICR have also been pushing the original article:

    Adam Benton · 19th May 2016 at 3:25 pm

    I know, that’s the link in the final paragraph of the introduction.

Venelouis Polar · 17th June 2016 at 6:01 pm

I read this article by the student Adam Benton and I see there is no problem with Answersingenesis article by Dr. Nataniel, evolutionists make so unverifyable claims that seens stupid how evolutionists still trying to refute the way creationists base their science on the Bible, well, Creationists at least have a base, evolutionists only assumptions of their minds…

    Adam Benton · 20th June 2016 at 1:56 pm

    What you want to claim as a base, the fact remains that these are fundamental problems with the methodology. An unrooted tree simply can’t be used to investigate ancestry.

    If you are curious about this subject, there’s a follow-up guest post elsewhere on this site:

Leave your filthy monkey comments here.

This site uses Akismet to reduce spam. Learn how your comment data is processed.