Casey Luskin calculates human/chimp sequence divergence at 14.9%

Conversation
,
1

As most of you probably know, there have been a series of posts on ENV in the last week that purport to show that the newest and best estimates of the sequence divergence between humans and chimps is now as high as 14.9% based on the complete sequencing of several ape genomes.

The authors of the paper in question didnt revise the current estimate of ~ 1.2% so Casey did a deep dive into the arcane supplemental data to obtain that higher number.

https://www.nature.com/articles/s41586-025-08816-3

As far as I can tell Casey took the total nucleotides for various insertions, deletions and inversions as nucleotide differences. It seems to me this approach is wrong for the question we’re considering and my very quick and sloppy calculation shows the divergence to be ~1.3% at the most. But I stress this is quick and sloppy so others need to check!!
Heres how I approach this question. Lets say I wanted to make some money by copying a current best selling novel and selling it cheap. Lets say the book is 300 pages and on the first page I introduce an extra space. I try to claim that after the space the 2 works no longer line up, but of course if you just remove the gap they line up perfectly. So then was I do is add 700 pages with just the letter ‘a’. ( or a period or a blank space). Now I can say my work is 1000 pages and only 300 pages line up, so the similarity is 30%. That may be low enough to avoid plagiarism charges. But of course the question is not how many letters and characters lines up but how many changes it took to convert one to the other - adding 700 filler pages is still just one change.

It seems to me that Casey has counted all the nucleotides in insertions and deletions as individual changes.

First I’ll say it was never the absolute sequence difference between humans and chimps that was considered evidence for common ancestry, it was the relative difference between different species. Calculating differences in this way would drastically increase the sequence divergence between individual humans as well.
What we’re interested in here is how many changes occurred in each lineage leading to human and chimps from a common ancestor. A single nucleotide change counts as one change but so does the insertion of 1000 nucs.
So what I did is this. The h/c sequence difference due to indels etc. is about 13.7%. Thats 411 million nucs. Lines are about 6 kbp, sines are ~ 300. If we take an average for the indels as 2500 that increases the divergence by a small fraction of a percent. If we count the average size at 200 bp that only increases the divergence by about .068%. I haven’t considered the regions that show rapid change such as the MHC loci and others and I could have easily missed something. When I started my hunch was the new # would be around 1.7%- 2%

3 Likes
2

ENV loves to rant about this, but they never look at any other divergence between species. As far as I know, and I think this must be true; Humans and Pans are most closely relate by ANY reasonable measure, and by some of the unreasonable ones too.

4 Likes
3

I would only propose a diving metaphor if the metaphorical pool had no water.

3 Likes
4

Yes, using Casey’s * method you still get the same phylogenetic tree for the apes they sequenced.
I’ve been thinking of what counterarguement he could use. The best one I can think of is to claim that in this case the entire length of the insertions or deletions is functional so you have to count every nucleotide. Thats based in part on what the authors of the original paper discuss in terms for possible functions for those sequences but thats optimistic speculation and gets into the whole junk DNA debate

  • he and I are not actually on a first name basis
1 Like
5

Yes, that’s exactly what he did, and he’s hardly the first. Creationists have been doing this ever since comparisons of long sequences first appeared. As a matter of fact, the first was probably Roy Britten, who estimated 5% difference based on a stretch of a few million bases, around 4% of which was indels in one or the other species. I suppose it’s a useful measure if you want to estimate the bond strength/melting point between hybridized DNA strands, but I can’t think of anything else it would mean.

(Was it this paper??: Springer M.S., Davidson E.H., Britten R.J. Calculation of sequence divergence from the thermal stability of DNA heteroduplexes. J. Mol. Evol. 1992; 34:379-382.)

On the other hand, the chimp genome paper found 35 million point mutations and 5 million indels and other non-point mutations separating the two species. So 40 million differences total, not including the unsequenced, mostly highly repetitive bits. I don’t see a way to turn that into a percent difference, though.

The Chimpanzee Sequencing and Analysis Consortium. Initial sequence of the chimpanzee genome and comparison with the human genome. Nature 2005; 437:69-87.

1 Like
6

I’ve written up a post about this which may be of interest. Human and chimpanzee genome similarity reconsidered?
Some key points:

  • it’s incorrect to count all non-aligning regions as differences (some just can’t align due to highly repetitive sequence)
  • copy number variants or repeats aren’t differences in the same way as genuinely new sequences are. Sometimes copy number variants are functionally important but they should be distinguished from other variants as they aren’t new sequences.
  • there are large differences within the human species and within the chimp species which aren’t considered. Within-group differences are relevant to claims about between-group differences.

There are some other points I’ve realised since writing this, that I may put in a follow up post - including that a lot of the “gap” differences are presumably due to the chimp genome being ~8% larger than the human genome (which doesn’t really fit the implied narrative that there’s lots of extra interesting stuff in the human genome).

4 Likes
7

So Luskin took one way to measure the difference, and said it gave a much bigger difference than previously advertised, without saying that the previous result was based on a different way to measure it.

He might as well have complained that a table has a bigger length if measured in centimetres than it does if measured in inches.

3 Likes
8

The obsession with inflating the degree of dissimilarity between humans and our closest relatives continues to elude me. We’re much less genetically similar to birds yet still related, even less to squid yet still related, and somewhere down in the range of 0-30% genetically similar to bacteria and archaea, and still related to those too. The absolute number is completely irrelevant to whether we are related. It’s the nested hierarchy that shows we are related.

And phylogenies are not even based on degree similarity anyway, though phylogenetic placement will typically correlate with similarity to a significant degree.

4 Likes
9

Reading the EN article, it seems like Luskin is primarily pulling this information from the SNP vs. gap divergence section of the Supplementary Material published with the original paper. And specifically Supplementary Figure III.11 and Supplementary Figure III.12.

In looking at that section, the authors of the paper define both the SNV divergence and gap divergance as follows:

SNV divergence is defined as the fraction of positions in the target haplotype where the two haplotypes are in different nucleotide states. Gap divergence is defined as the fraction of positions in the target haplotype that are not aligned to the other haplotype, which could be due to biological processes (e.g., gene loss/gain and insertions/deletions), missing data, or technical problems (e.g., alignment failure due to SVs, repetitive elements, etc.).

They then depict density plots based on 1Mb segments. They mention reporting both mean and medians, which they plot on the chart. It looks like the numerical values reported on the charts are the medians.

Luskin’s 14.9% seems to be based on just adding together the reported human vs chimpanzee (hg002 vs. PanTro3) median values from the SNP and gap divergence charts respectively.

While Luskin points out the various comparisons with other species, he also completely fails to report on the within-species comparisons.

For example, in the text the authors of the paper mention the autosome gap divergence is between humans and gorillas, but then immediately follow that sentence by reporting the within-species gap divergence for gorillas:

However, autosome gap divergence showed different trends, as it was highest for human-gorilla (17.9-27.3%). The within-species autosome gap divergence was also highest for gorilla (13.8%), which could be due to the assembly (larger number of contigs), SVs, or mobile elements.

Looking at Supplementary Figure III.12, while the human-to-chimpanzee autosome gap divergence is reported at ~13.3%, the within-species for humans is reported at ~3.5%. The within-species gap divergence for chimpanzees (PanTro3) is 8.1%. And as mentioned above, the within-species value for gorillas is ~13.8%, higher than the human-to-chimpanzee comparison.

Luskin tries to spin this as some sort of breakthrough revelation that is being deliberately buried within the paper (his conspiratorial spin in the EN article is nothing short of inspired). But when you put things into the context of the within-species comparisons, the species-to-species values he is looking at don’t seem all that remarkable.

The only thing that struck me as interesting was the gap divergence values for gorilla comparisons, since those seem quite high. But as they mention in the paper, this could be due to the assembly of the gorilla genome itself.

10

Bingo. And let’s face it: a Juris Doctorate prepared him to do what many lawyers are known for: impassioned spin leading to bizarre conclusions regardless of the real truth. [Yeah. I have many lawyers in my extended family. No disrespect intended, @Puck_Mendelssohn.]

When I first saw the title of this thread, I mentally re-arranged it:

Chimp calculates human/CaseyLuskin sequence divergence at 14.9%.

2 Likes
11

You know, I have long argued that law schools have got the advocacy contests (moot court, writing-and-argument classes, etc.) badly wrong. They tend to take cases that are arguable on both sides, and then see how the kids do at arguing them, debate-competition style. I believe this is a complete mischaracterization of most litigation.

Most litigation is one side that’s right, and one that’s wrong. The “right” side has the task of tying down the whole case in such a way that it can’t be easily assaulted with bullshit. And the “wrong” side has the task of figuring out what the best bullshit is, and figuring out what part of the “right” side’s job has been done badly enough that it can be turned into an apparent dispute.

I lost a lot of motions where I was in the right, in my early days, because I didn’t understand that you not only need your case to make better sense than the other side’s, but that you need to caulk all the seams so that the blasted thing is watertight.

When you’re in the wrong, there really is an art to making it seem as though you are right or, if not that, potentially right. I can’t say that I regard Casey as a high practitioner of that art, because so much of what he does is so ham-handed and downright awful. But if the audience is the slack-jawed yokelvolk who are the DI’s darlings and contributors, perhaps he’s doing a splendid job.

3 Likes
12

“I suspect that this radical finding has implications — not just for science, but also for human exceptionalism, for the reliability of heavily marketed talking points, and more — that people will be discussing for a long time.” (emphasis added)

Rumraket Mikkel R. makes a good point. It seems to me, looking at this from a layman’s perspective, and ignoring Luskin’s typical hyperbole, that the essential point of chimp-human comparisons is not so much the exact percentage of genetic overlap or difference, but demonstrating that chimps (and also other apes) share a common ancestor with humans. But, and I appeal to the experts here, biology has known this for decades, right? And with due respect to our theist friends, of, primarily, the evangelical flavor, why is this common ancestry so offensive? Does it really do violence to “human exceptionalism?” If you think about it, God, using natural selection to create humans, is infinitely more clever than just popping us into existence one Saturday afternoon 6000 years ago (give or take)…

4 Likes
13

Which can be achieved just by making it seem like the other side is wrong.

Luskin uses this tactic by repeatedly bringing up the 1% human-chimp difference and claiming this new finding overrides that. Of course he does this while completely ignoring that % differences are entirely contextual based on what they are measuring. But I imagine he’s not expecting the typical EN reader to know this.

2 Likes