This post will look into the "age" or timeline of our Y chromosome. There are different methods used to estimate how this chromosome changes over time by accumulating mutations (*). This leads to a mutation rate expressed in mutations per site per year. It can be calculated by comparing the Y chromosomes of two different men, or a modern human and an archaic one, or another ape and humans, and count the mutations by which they differ, if we know when their lineages split.
(*) Actually, to be correct, a mutation is a chance change in a base pair, an error in the transcription of our genetic code. A substitution is when a mutation becomes fixed within a population, one that has not been erased by genetic drift, or natural selection. For instance, a chance mutation may lead to a "C" to appear instead of a "G", but and then another chance mutation could flip it to a "T" in that case looking at the original and the final versions we'd see one substitution, and imagine one mutation, but there were 2 mutations. Another example would be a flip back, to "C" (back-mutation or reversion), where we'd see no mutation or substitution, but in fact, there was one mutation.
The formula to calculate the Time to the Most Recent Common Ancestor (TMRCA) is shown below. Where the TMRCA in years; k is the number of base pair differences between both men, 2 is a factor added because the difference corresponds to the divergence in both men, half for each one and we don't want to count them twice, L is the total sequence length sampled in which k was detected; and μ is the mutation rate.
TMRCA (years) = k (mutations) / μ (mutations/site · year) · L (sites)· 2
From which the mutation rate can be calculated by shuffling terms as shown:
μ (mutations/site · year) = k (mutations) / TMRCA (years) · L (sites)· 2
A real life example comparing Chimpanzee and Human base pair divergences found in Shen et al., (2000) is the following: assumed TMRCA: 4,900,000 years, acutal values measured (the length sampled in base pairs) L: 38,568, (the number of divergent base pairs) k: 470, calculation of μ shown below:
μ (mutations/site · year) = 470k (mutations) / 4,900,000 (years) · 38,568 (sites)· 2
μ (mutations/site · year) = 1.25 × 10−9.
Assumptions and Data
As you can see, the formula is logic, and straightforward, we take a sample of a DNA strand from a Y Chromosome with a length of L base pairs (my intro to Y Chromosome post explains the basic terms) in both subjects, and then count the differences we observe (k), knowing how long ago both individuals shared their last common ancestor (TMRCA) we calculate the mutation rate μ.
The same method can be applied to any chunk of nuclear DNA including the X chromosome and any of the non-sexual chromosomes or autosomal DNA, as well as the mtDNA. The interesting part is that they all mutate at different rates.
Y chromosome mutation rates (μ)
Over the course of the years, different studies using different methods have attempted to calculate the mutation rate of the Y chromosome in humans. The table below shows some of these studies. The values of μ are given in mutations per site per year. For non-scientists, note that expressing a number multiplied by 10-9 is another way of writing: "divided by 109" and 10 to the ninth power is 1000,000,0000. This means that 1.24 x 10-9 = 1.24 / 1000,000,000 = 0.00000000124 (very small indeed!).
- 1.240 × 10-9 Thompson et al., (2000)
- 1.500 × 10-9 Kuroki et al., (2006)
- 1.000 x 10-9 Xue et al., (2009)
- 0.617 × 10-9 Mendez et al., (2013)
- 0.820 × 10-9 Poznick et al., (2013)
- 0.530 × 10-9 Francalacci et al., (2013)
- 0.716 × 10-9 Trombetta et al., (2015)
- 0.871 × 10-9 Helgason, (2015)
- 0.760 × 10−9 Fu, et al., (2016)
The difference between the first value reported by Thompson et al, and the figure given by Francalcci et al is 2.34 times!
There are big discrepancies in these μ values
Using one or the other to estimate the age of a given specimen would result in one figure being over twice the age of the other one.
The different methods employed to calculate μ involve different assumptions and they all have their shortcomings. In the following commentary I will follow the excellent work of Wang CC, Gilbert MT, Jin L, Li H. (2014) ( Evaluating the Y chromosomal timescale in human demographic and lineage dating. Investig Genet. 2014 Sep 10;5:12. doi: 10.1186/2041-2223-5-12. PMID: 25215184; PMCID: PMC4160915).
The data used for comparing divegencies in the Y chromosome's base pairs can come from different sources: Ancient DNA, samples taken from prehistoric human remains which have been dated by using radiocarbon or other methods. Genealogical, using samples from a certain family for which the genealogy has been confirmed and dated, these usually involve short time spans and few generations. Archaeological Events, taking an estimated date for an event, such as the peopling of America or the settlement in a given region in Europe, and applied to samples from that date.
Confounding Factors
As mentioned in my post on phylogenetic tree branch lengths (a factor that shows that mutation rates are not constant), the value of μ should be considered as an estimation, and not something written in stone. Trombetta et al., (2015) point out that "variants" (mutations) appear at different rates across Y chromosome haplogroups, geographic locations, and time: "... we observed a remarkable heterogeneity in the distribution of variants, not only across different regions, but also across lineages and different times. Hg A00 stands out for showing strong associations with almost all genomic features considered. In the rest of the tree Hg's A0, A1a, A2'3 and B differ from Hg's DE, FC and R, and ancient branches differ from recent ones. The two levels are not entirely independent, as far as recent branches are enriched in lineages belonging to Hg's DE and R. It is possible that different social habits, lifestyles and environmental conditions experienced by populations harbouring different haplogroups resulted in systematic variations of the generation time and average paternal age at conception." They attribute this to older or younger age of the fathers when their children are conceived —older dads have more mutated sperm, and the effects of environment on DNA mutations.
Chimpanzees and Men
When comparing human and chimp Y chromosomes, we are not only separated by a gulf of 5 to 7 million years of separate evoluton, the evolution itself has been different in both species. The chimpanzee Y chromosome is much smaller than that of humans. it lost roughly one-third of its genes in the MSY, or male-specific Y region of their Y chromosomes, compared to men.
Kuroki, Y., Toyoda, A., Noguchi, H. et al. (2006) noted that there is a greater divergence in the sequence of human Y chromosome vs. chimpanzee Y chromosome than between the whole genomes of both species (1.78% and 1.23% respectively).
The reliable dating of the Chimpanzee-Human split is still being debated, and figures range from 4.2 to 12.5 million years ago. A factor of three!
There are also structural differences in the shape of our and the chimp's Y chromosome which complicates the alignment of segments for comparison. Finally, chimpanzees and modern humans have different pair-bonding sexual behaviors (Hughes et al, 2013). Schaller et al., (2010) note that receptive females copulate with multiple male partners creating selective pressure towards the male fertility genes in the chimp's Y chromosome. Monogamous pair bonding in humans lacks this intense selective force. This alters the rate of the mutational clock. The pair bonding in hominins can be seen in the reduced sexual dimorphism in australopithecines and the loss of sperm competition adaptations, suggesting less male-to-male strife, and the growth of cooperation as a means for reproductive success (Gavrilets, 2012).
Genealogical Methods
Pedigree-based studies look at men belonging to the same family, sharing the same paternal lineage, and whose birth dates are known, or at least, how many generations separate them. This provides a well defined dating.
Xue et al., (2009) studied 13 generations of men in haplogroup O3a. However, there are some potential sources of error in this method: Since mutation rates are random, variable, and unpredictable (the statistical term for this is highly stochastic), are we sure that 13 generations (approx. 390 years) is a long enough interval.
Another is the haplogroup itself, we have mentioned that haplogroups accrue mutations at different rates. Is haplogroup O3a a good reference for all other haplogroups?
Finally, if selection and genetic drift act mutations eliminating some of them over longer timescales, the number of mutations found in a genealogical study will be smaller than the actual one detected in longer scales.
Mutation Rates adjusted for autosomal mutation rates
This method was developed by Mendez F., et al., (2013) when they dated the extremely ancient A00 haplogroup Y chromosome, found in the Mbo people of Cameroon, Africa. The μ used by Mendez team was based on a study conducted in Iceland that calculated the autosomal mutation rate by analyzing the divergence between parents and their children. This implies several unverified assumptions: autosomal and Y chromosome mutation rates are similar (they are not), substitution rates and mutation rates are equivalent (they are not). They also used a generation time that spanned from 20 to 40 years when life expectancy for men in Cameroon in 1950 was below 40! (Source). Other authors using genealogical data, like Boattini et al., (2019) found generation lengths of 33.57 years. Hunter-gatherer groups in Equatorial Africa 150,000 years ago probably began mating at the age of 15. How can we know for sure? (See Elhaik E, Tatarinova TV, Klyosov AA, Graur D., (2013) and their critique to Mendez et al.
Older generation times lead to more mutations, and an overestimated TRMCA. A00's age is far shorter than the one put forward by Mendez et al.
Archaeological Data
In my posts, I have mentioned many ancient DNA samples taken from the remains of prehistoric, ancestral human beings and hominins (Yana River, Mal'ta and Ust'-Ishim). They provide a certain age with reliable radiocarbon-dating, and if the DNA is not degraded or contaminated, a reliable count of mutations.
Fu, et al., (2016) compared the remains of the Ust'-Ishim man with those of men alive today, and looked for "missing" mutations, those that appeared in modern men after Ust'-Ishim died. The team calculated a mutation rate of 0.76 × 10-9
Human Migration approximations
Assuming dates for certain migratory events, like the peopling of America, placed at around 15,000 years ago, the dates for certain haplogroups found exclusively in America can be set close to that event. This has been used to calculate μ for splits between Asian and Amerindian lineages, or between Amerindian groups within America, obtaining a μ of 0.820 × 10-9 (Poznick et al., (2013)).
A similar method was applied by Francalacci et al., (2013) to men native to the island of Sardinia in Europe, peopled around 7,700 years ago, and using the mutations detected in a sample of Sardinian men to calculate μ for their haplogroup (I2a1a). The value obtained was very low compared to those shown in the Table further up: 0.530 × 10-9.
Can we be certain that the haplogroup diversified in its current location in Sardinia or America? Could it have taken place earlier (European mainland, or Siberia, respectively). Do the current Sardininans belong to the group that reached the island 7.7 kya? Or did they arrive later?
Implications
The sum of these factors show that environment, generation times, sexuality (pair bonding), natural selection, and genetic drift can promote or reduce mutation rates. Studies have a 2.4-fold variability, ranging from 1.24 × 10-9 Thompson et al., (2000) to 0.53 × 10-9 Francalacci et al., (2013), and these are the mean values, the confidence intervals are even wider. Supposing a threefold difference, what some estimate as a divergence taking place between two Y chromosome haplogroups 50,000 years ago during the final Out of Africa migration, may have taken place within Africa 150,000 years ago! And the A00 split instead of taking place 250,000 years ago may reflect one that ocurred 750,000 years ago.
The possibility that our most ancient ancestors had more chimp-like behavior would have implied a faster mutation rate during that period, followed by a slower rate later on. Even pair-bonding during the patrilineal, sedentary, agricultural period of the past 8,000 years may have slowed down mutation rates in comparison to the matrilineal hunter-gatherer period that preceeded it.
The image above shows how mutation rates can influence the age of the TMRCA. For the same measured value of "m" mutations marked by the gray line, the use of different mutation rates μ influences the depth or age to the TMRCA. A quick mutation rate like the blue one (μ1) accumulates mutations quicker and takes less time to reach the detected "m" mutations, so its TMRCA is younger (T1), a slower mutation rate (red line) like μ2 takes longer to accumulate the "m" mutations and its T2 is longer. Finally, a variable mutation rate like μ3 (green line) that was faster in the distant past, and slower later on will have an even longer and older timeline (T3).
For those interested in maths, the slope of each curve marks the mutation rate dm/dt = μ(t), steeper curves mean quicker mutation rates. It an analogy of speed as the differential of space over differential of time.
I believe that the μ3, variable mutation rate is closer to reality, reflecting variable social-sexual-cultural patterns of the small and egalitarian promiscuous hunter-gatherer groups of early hominins and human evolution. Later settled societies with paternal monogamy and private property led to slower mutation rates. This of course would imply an older root for the human Y-chromosome haplogroups, in Africa, prior to our migration into Eurasia.
Fall has begun here in Buenos Aires, in the Southern Hemisphere! Lucky you who are now entering Spring!
Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2026 by Austin Whittall ©
No comments:
Post a Comment