Pedigree Mutation Rates vs. Genome Divergence

So the Ho et al 2005 paper used a program called BEAST which does Bayesian estimation of parameters from the data meaning it returns a posterior probability of a parameter based on the data and some reasonable model of molecular evolution (I think in this case they used some variant of an HKY model, it should be listed in the methods). The date therefore is a highest point in a posterior distribution of the parameter. The upper and lower and lower bounds are apparently associated with uncertainty around the dates of the calibration points.

Yes, the numbers they get for divergence using the time corrections are around those we see from other studies so they make sense in light of everything else.

Again the point of this is they are saying you can not extrapolate out mutation rates from short time scales like those obtained from pedigrees and use those same rates, unadjusted, to ask questions about divergence among species. This was a point I had brought up in my debate with Nathaniel Jeanson that he either didn’t understand or just ignored, or a bit of both more likely.


So @Herman_Mays, for humans, what is the scale of the difference between the estimates? And in what direction? In this case, I want to know for autosomal and sex chromosomes, not mitochondria.

1 Like

@Herman_Mays, have you seen this paper yet?

The mutation rate in human evolution and demographic inference, Aylwyn Scally,

The figure here shows mutation rate estimated in several ways:

You can see that the pedigree based methods (blue) produce results that are consistent with several other methods, not just pedigree. You can see that the variance of pedigree methods is high, but it is not that bad.

The paper also goes into depth about the (relatively small) mismatch between human-chimp divergence and the divergence computed by mutation rates and time alone. As @glipsnort alludes, some of this is resolved by taking coalescence into account. There is a discrepancy, but it is not large.


That’s where the graphs in the paper and subsequent equations come into play right.

So the substitution rates over time in millions of years is based on the BEAST programme?

It’s definitely interesting, and difference in substitution rates dont over time is large esp in the first two million years.

Yes. The long term rates are calibrated by fossil and biogeographical boundaries. Something for religious reasons Jeanson refuses to do so he just takes the pedigree rates and extrapolates them out to deal with any divergence he likes.


“ The three new date estimates were considerably younger than those estimated in previous studies, which gave ranges of 365–853 ka (Ovchinnikov et al. 2000), 550–690 ka (Krings et al. 1997), and 317–741 ka (Krings et al. 1999) for the Neandertal-human divergence; 151–352 ka (Ovchinnikov et al. 2000) for the last common ancestor of Neandertals; and 106–246 ka (Ovchinnikov et al. 2000), 120–150 ka (Krings et al. 1997), and 111–260 ka (Krings et al. 1999) for the last common ancestor of humans (fig. 5). This discrepancy arises because high, short-term rates of change were taken into account by our approach. The standard error ranges on the new estimates do not take into account uncertainty associated with equation (7). Furthermore, the revised date estimates are sensitive to μ and λ, which are difficult to estimate accurately unless abundant data are available. However, our results clearly show the effect of correcting date estimates of geologically recent divergence events using a rate decay curve estimated from sequence data and demonstrate the unsuitability of using a single short-term or long-term rate to date evolutionary events on an intermediate timescale.”

The divergence times from Ho et al 2005 are indeed a big lower but, of course, nowhere near those hoped for by YECs. These are difficult approaches that are fraught with assumptions throughout and in those cases you just have to draw upon multiple independent lines of evidence to arrive at a decent answer to the question.

In BEAST for different loci you simply have to scale these different loci by their effective population size (1 for nuDNA, 3/4 for X chromosome, 1/4 for mtDNA and Y chromosome data).


I am a little confused by this statement. While pedigree based methods are fairly close to each other, the overall variance in the values from all the methods is high in percentage terms…
If I take the equation D=RT as more or less applicable, I get the understanding that diversion timelines are inversely proportional to the rates… so a 50% higher calculation of mutation rates should lead to an approx 50% lower time of diversion.

Is this a correct understanding? If it is… why isnt it significant?
The paper also points to such big differences in predictions of diversion time.

For some events,

such as the speciation of humans and chimpanzees, a
higher rate had been increasingly difficult to reconcile
with fossil and archaeological data, and a lower value
(implying older date estimates) mostly improved concor-
dance [35,36]. However for more ancient events a longer
timescale was problematic, and to a large extent remains
so still. For example, applying a present-day human
mutation rate of 0.5 -
109 bp1 year1 to the 2.6% ge-
netic divergence between humans and orang-utans [37]
yields an divergence time of 26 Mya. Even allowing for a
large ancestral coalescent time of 5 Myr this is substan-
tially older than the dates of 12–16 Mya typically quoted
in paleoanthropological literature [38]. The difference
increases for older dates: the human-macaque divergence
[39] implies a speciation more than 40 Mya, whereas
paleoanthropological studies generally place this node
at 25–30 Mya [40].

It looks like the non-pedigree methods (PSMC, aDNA) are still measuring rates well within 2MYA and according to the Ho et al paper I was talking about, if I recall correctly, this would be around the limit where you could apply these rates. Meaning none of those rates would be constant if you are asking questions about human chimpanzee divergence. Although I didn’t see what “other” meant in that figure.


That’s to be expected from the nuclear genome, as most of it is evolving neutrally. The story is likely different for mtDNA.


Is it not also the case, in addition to the non-coding portion of the genome being different, that for every one cell division, there are multiple mitochondrial divisions?

I refer to animal mitochondria here. Plant mitochondria, for reasons unknown to me, have an extremely low mutation rate.

That’s true. But the mutation rate per replication is also higher. Different polymerase (nuclear-encoded) and none of the proofreading that nuclear genomes have.


'K, rookie question or request for clarification here…

Nuclear DNA for the zoo of eukaryotic beasties is going to vary all over the place because all these resulting phenotypes will be adapted to various ecological niches. Mitochondria, on the other hand, have a much more narrow niche, earning their keep with their powerhouse role. There is not a lot of scope for changing environment there. Over the epochs, I would expect that mitochondria would then become fairly optimized to their role by some time in the past, so that mtDNA is less inclined to neutral change than nuclear DNA, or relatively touchy with more mutations tending to severe consequences. Would this not have a strong selective effect on the degree of variance we observe? There is, of course, some variation as my 23andMe results would indicate.

The short answer is that this appears not to be the case. Mitochondrial evolution is faster than nuclear genome evolution in most animal species. The neutral rate for mitochondria is much faster than the neutral rate for nuclear genomes. While it’s true that much less of the mitochondrial genome is evolving neutrally, since most of it’s protein-coding, around a third of positions are degenerate, meaning that transitions, at least, don’t change the amino acid coded for, and most of these are 4-fold degenerat, meaning that transversions don’t either. Anything that changes amino acids is much more conserved, but over time even that can change a lot. Some amino acids are similar enough that flipping back and forth is close to neutral. Further, the most important part of a mitochondrial protein’s environment is other mitochondrial proteins with which it interacts. A change in one can make favorable a change in another, so that one fixation can trigger others. Over evolutionary time, mitochondrial proteins can change radically in sequence without significant change in their function or performance. However, a mitochondrial-encoded protein in one species may be incompatible with a nuclear-encoded protein in another species, even though function hasn’t changed.

One more thing: if you know what saturation is, mitochondrial changes become saturated much more quickly than nuclear changes, so genetic distances won’t increase in proportion to the actual number of changes.


Also it’s important to remember mtDNA has much faster substitution rates just because of the smaller effective population size alone where nuDNA is much more prone to incomplete lineage sorting especially for recent divergence times.

1 Like

Well said. The co-evolution of cytoplasmic (mitochondrial and plastid) and nuclear genomes is a super interesting topic.
For example:

Fundamental biological processes hinge on coordinated interactions between genes spanning two obligate genomes—mitochondrial and nuclear. These interactions are key to complex life, and allelic variation that accumulates and persists at the loci embroiled in such intergenomic interactions should therefore be subjected to intense selection to maintain integrity of the mitochondrial electron transport system. Here, we compile evidence that suggests that mitochondrial–nuclear (mitonuclear) allelic interactions are evolutionarily significant modulators of the expression of key health-related and life-history phenotypes, across several biological scales—within species (intra- and interpopulational) and between species. We then introduce a new frontier for the study of mitonuclear interactions—those that occur within individuals, and are fuelled by the mtDNA heteroplasmy and the existence of nuclear-encoded mitochondrial gene duplicates and isoforms. Empirical evidence supports the idea of high-resolution tissue- and environment-specific modulation of intraindividual mitonuclear interactions. Predicting the penetrance, severity and expression patterns of mtDNA-induced mitochondrial diseases remains a conundrum. We contend that a deeper understanding of the dynamics and ramifications of mitonuclear interactions, across all biological levels, will provide key insights that tangibly advance our understanding, not only of core evolutionary processes, but also of the complex genetics underlying human mitochondrial disease.

1 Like

Thank you, that post was a super clear introductory explanation.

I’m working on data now suggesting evidence of introgression of both mtDNA and nuclear encoded mtDNA transcription factors across species.

1 Like

Wow, sounds fascinating!

What species, if one may ask?