An intro to Y chromosome haplogroups

My last post mentioned the possibility of Denisovans being linked to Haplogroup P of the Y chromosome, and the possible presence of haplogroup P in America. I thought that it would be straightforward to associate Denisovans with haplogroup P. But after giving it some thought, it isn't. So I decided to recap and go back to the basics and try to find out if it is feasible to associate Denisovans with any human Y chromosome haplogroup.

Transmission and Mutations of Chromosome Y

Chromosome Y is inherited in a patrilineal manner. All men carry one chromosome X and one chromosome Y, they inherit the X from their mothers and the Y from their fathers. In human beings, carrying a pair of X and Y means you are a man. If you inherit the X from both parents, you are a woman.

Base Pairs

Like all chromosomes, Y is made up of DNA (Deoxyribonucleic acid) a molecule that is made up of two counter-spiraling helicoids (like a winding circular stairway), both strands are made up of sugar-phosphate and are the backbone onto which four different compounds (bases) attach. These are Adenine (A), Cytosine (C), Guanine (G), and Thymine (T); these are the steps of the stairway. The bases bind in a particular way, A with T and G with C.

DNA carries the instructions that the cells can read and use it as a template to build proteins.

The bases are laid down in a certain sequence along the spirals, for instance one strand could have: A T G C C T A G T... and the opposing one would have the complementary bases (a T for every A, a C for each G, and viceversa): T A C G G A T C A...

Each pair of linked bases (the rungs of the stairway) is a base pair, for instance A—T. There are 60 to 300 million base pairs in each of our 46 chromosomes, a total of around 3 billion of them in our genome.

Chromosome Y is the smallest in terms of base pairs: roughly 60 million on average.

Genes

Genes are a specific sequence of aligned base pairs in one chromosome. They are the basic unit of heredity. A gene has the codes required to produce special molecules known as RNA or specific proteins.

When cells replicate, or in the case of our sexual gametes (ovarian eggs and sperm), the chromosomes undergo a process of splitting and the DNA strands unwind and replicate. With 3 billion base pairs, copying the new strands can lead to alterations in the base pairs: mutations.

Some base pairs are lost (deletions) others are copied twice (duplications). This alters the blueprint and may have an impact on how the gene that contains these mutations functions. Mutations can be negative (deleterious), neutral, or even positive. As we will see below, mutations in chromosome Y are problematic, as they accumulate.

Hominin Evolution

Our closest primate relative is the common ancestor that we share with chimpanzees, who lived between 6 and 8 million years ago.

This distant ancestor evolved, through a series of mutations into our homo ancestors: Homo habilis and Homo erectus, and others, reaching the common ancestor of Neanderthals, Denisovans, and Homo sapiens.

The original male hominins living 3 or 4 million years ago, carried certain base pair sequences in their Y chromosomes. We can imagine a small population with a few hundred males sharing identical base pairs (this is of course an over simpification, they differed). These "men" then passed their Y chromosomes with these same sequences to their sons. Some of them probably died in their childhood and did not mate, others only had daughters, so their Y chromosomes were lost, only those who had sons passed them on to the next generation.

Mutations

Each generation went through the same process, but the sequences that were passed on, changed over time as chance and external factoes introduced random mutations in the base pairs of the DNA strands of the Y chromosome.

Below are some of the factors that cause mutations:

• Chance, random mutations.
• Age of conception, those men who reproduce later will have more male germ-cell divisions, and each division entails the risk of a failed copy in the sequence. Formation of sperm (or spermatogenesis) implies constant cellular division over a man's lifespan. Female oocytes that result in eggs are all produced at birth, in one go.
• Methylation, the addition of a methyl group (—CH₃) to the DNA strand due to epigenetic (lifestyle or external) factors such as stress, famine, or toxins (alcohol, chemicals, smoking).
• Oxidative stress. Sperm are also modified by inflammation, heat, radiation (cosmic rays) which can produce free radicals which are oxidants and degrade the DNA.
• Inadequate repair systems, although the Y chromosome has limited repair mechanisms as it is mostly non-recombining (it has no partner like the other non-sexual chromosomes and does not recombine with the X chromosome). It has limited ability to fix glitches due to its high content of repeat sequences called palindromes.

Unlike other chromosomes, mutations can't be purged in Y chromosomes, so if they are harmful, they will accumulate and lead to genetic malfunction (sterility, illness, death, stillborn boys, and miscarriages). The chromosome will not work as expected. Mutations can reverse, undoing the original variation, but it is an unusual event.

The hominins evolved, but the basic structure of their Y chromosomes was similar, only the accumulated mutations, those that had allowed viable offspring survived, the others vanished as those who carried them died.

The whole genome is subjected to mutations, the X chromosome, and the other chromosomes, and natural selection acts, promoting the survival of the mutations that provide an advantage to those carrying them. It is possible that certain Y chromosomes, even though they were fit and possibly provided survival benefits, were eclipsed by deleterious mutations in other chromosomes. This led to the loss of many Y chromosome variants that had evolved over millennia.

The image Below shows an extremely oversimplified version of a Y chromosome. The original, ancestral version is (1) it has 50 million base pairs (not shown), but one mutated, say an A for a T (shown with the red band). It survives in the following generation and after many generations during which othe over the years, and today, when we look at the global population and sample the men, we find the variants marked (2) to (8), each one carries the original "red" mutation but have added others, each identified with a different color (blue, black, orange, green, violet, and gray).

Y Chromosome markers explained. Austin Whittall ©2026

Haplogroups

Here is where modern geneticists and anthropologists use their computer software tools, algorithms, and theory to build phylogenetic trees. They choose certain base pair mutations known as SNPs as "markers" that define "haplogroups" that split populations into branches from a main trunk (the basal one). Assuming that there are no back-mutations, and that repeat mutations are extremely uncommon, they propose that each marker (a mutation at a given base pair) that is fixed in a given population arose in a sequential manner.

In the example shown above, the phylogenetic tree would be the one shown below, assuming that mutations accumulate and don't reverse:

Caveats

We could argue that (8) resulted from (4) that lost its "blue" mutation, but as mentioned further up, orthodoxy considers that back mutations are rare so they ignore them. Problems also arise when we ask which mutation came first, (2), (7), or (8) they are all just one mutation away from the ancestral root.

In the real world, this is far more complicated, especially when we sequence the Y-chromosome of Neanderthals and Denisovans, which have degraded, decayed, and are incomplete. The strands of DNA of ancient remains are full of voids, and bases that have switched, or flipped. Comparing them with modern strands is done with software that "matches" them and points out the differences.

Toomas Kivisild (2017) highlights the complexity of analyzing haplogroups in ancient Y-chromosome samples: "it can be challenging to distinguish true mutations from those induced by damage, particularly in case of C to T and G to A substitutions", contamination is another factor, and the errors caused by low quality readings caused by "coverage" (how many sites were measured in a sample for comparison with a reference genome) and "sequencing depth" (how many reads covered the sample). All of them can lead to incorrect branch lengths, tree inferences, and dating.

SNPs

And mutations can appear in markers leading to mistaken identifications, like the ones reported by A.T. Fernandes, R. Goncalves, and A. Brehm (2004), in the Azores, where "It was found that some individuals share the same haplotype but belong to different Y-chromosome haplogroup suggesting that SNP mutations may occur frequently." SNPs are Single Nucleotide Polymorphisms (a switch in one base, like an A for a T). This paper notes that "The human Y-chromosome haplogroups are characterized by several mutations according to the phylogeny and nomenclature proposed by the Y-chromosome Consortium. Haplogroups are considered to be stable due to the very low mutation rate of most binary markers (SNPs), around 10⁻⁹ per base per generation, showing evidence of recurrent mutation at only 6 of 240 SNPs." This study involved 240 unrelated men and found "three individuals that share an haplotype with a double duplication suggest[ing] that a recurrent mutation occurred in SNP M78 because the duplication event is rare and it is unlikely to occur twice. For the individuals sharing the same haplotype but belonging to different haplogroups two explanations can be possible: recurrent mutations in several SNP namely in M78 and M81 (E3b1/E3b2) and M172 (J/J*) or several STR mutations may have occurred." So much for haplogroups and the assumptions that they are based on! 6 in 240 may seem a low frequency but it is high, 2.5%.

STR, mentioned above is a Short Tandem Repea, a snip of 2 to 6 base pairs long that is repeated two or more times in a location along the DNA strand.

The Branches of the Haplogroup tree

Another factor to consider when looking at ancient and modern DNA is that a man who died 50,000 years ago shows us a picture of a lineage that stopped accumulating mutations then. During the following 50,000 years all other lineages continued adding mutations to their DNA strands at a rate of 10^-9 per base per generation (I am using the SNP haplogroup marker value given above). So assuming generations of 25 years, in 50 ky, there are 2000 generatons, and with 50 million base pairs in a Y chromosome, we can calculate 50 x 10⁶ x 2 x 10³ x 10^-9 = 100 mutations.

When we look at our last shared common ancestor with the Denisovan group who lived ~550,000 years ago, if we assume no mixing with these people since then. After 500 ky, when we met them again in Asia during the Out of Africa migration, each branch, ours, and theirs would have accumulated an average of 1000 mutations (1000 in 50 million base pairs is a very low proportion: 0.002%). With Neanderthals from who we split later, around 350 kya, and met during our first Out of Africa 150 kya, only 400 mutations would have accumulated during the 200 ky we remained apart.

Intra Homo sapiens comparisons like the ones that compare a modern Chinese or a Native American from the Amazon, with an African San, are comparing lineages that have accumulated mutations since they split, probably 60 ky (Chinese and Amerindian) ago from the African line, accumulating mutations separately since then ~120 mutations in each line. And Native Americans with a 30 ky split from Chinese would have added 60 mutations.

Branch Shortening

Finally, and this will be the subject of my next post, mutations do not accumulate at the same rate. Africans have "shorter branches" on the phylogenetic trees. A paper by Petr et al, (2020) using data from an ancient man found in Siberia, Ust’-Ishim, 45,000 years old and modern humans noticed that the number of mutations from the root of each "branche" that leads to Africans and Non-Africans differed, implying different mutation rates (or, in my opinion, incorrect dating of the root, or fork): "Importantly, we discovered that the branch-lengths in Africans are as much as 13% shorter compared to non-Africans, which is consistent with significant branch length variability discovered in previous studies and suggested to be a result of various demographic and selection processes. Notice how they attempt to explain the issue away with "various" processes.

Further reading. Though old and dated, it is short, clear, and comprehenisive. Mark A. Jobling and Chris Tyler-Smith, (2003). The human Y Chromosome an evolutionary marker comes of age, Nat Rev Genet. 2003 Aug;4(8):598-612. doi: 10.1038/nrg1124.

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2026 by Austin Whittall © 

On the direction and root of phylogenetic trees

When I see a phylogenetic tree (also known as an evolutionary tree), I always wonder why do we believe that those branches, trunk and the root which anchors it, are correct. I ask myself why is it assumed that the mutation took place in one direction and not the other. And this trivial question is fundamental because the branches open up from other branches based on the differences between the DNA as you move along them.

Below is a very simple example of what I mean. Imagine we reach a planet, and come across a species whose DNA is sequenced and reveals the folowing genes: A, B and C.

We then take a sample of individuals, and sequence their genome. The nine individuals in our sample come from different continents and the "order" of the genes is different in each individual:

AAAAA, AACAA, BAAAA, BAAAB, BBAAA, CACAA, AACAB, BAACB and BABAB.

We assume that mutations take place at random so a B can spontaneously mutate into C or A, an A into B or C and a C into A or B. So, a group of scientists after looking at the genomes assumes that AAAAA is the oldest group of that species and that the other populations are the result of mutations that modified the original genome. They build tree (1) shown below. The most distant population is the one with the BABAB genome.

The red arrow marks the "founding" population and the green arrow the "newest" group, descended from them.

But another group of scientist based on some ancient remains and other assumptions, says "No, the original population is not AAAAA, it is the people carrying the BABAB genes" (exactly the opposite to what the first group of scientists have proposed and proven in Tree (1).

The second group builds Tree 2, where as we can see (the green arrow shows the original population and the red one shows us where they place the population AAAAA. For this second group of scholars, populations AACAB and CACAA are the "most recent" populations. The tree below shows the mutated gene in red:

Two different trees built from the same genome samples. Copyright © 2018 by Austin Whittall

The scholars could then identify haplogroups where the A to C or the B to A mutation marks a haplogroup and theorize on how these haplogroups evolved one from the other... Does this sound familiar? Yes, it is how the mtDNA and the Y chromosome DNA haplogroups were created -by adopting certain mutations as key indicators for branches and defining that it took place in a certain way (in our DNA, for instance, an aadenine (A) switch for a cytosine (C) may mark a haplogroup. A for C but, we could also -as in our thoretical planet imagine that the C switched for an A and that the supposed parent genome is actually the child and not the other way round.

Thus the "new" American genomes could actually be the oldest and the African ones the youngest (like switching from tree 1 to 2 above).

This is of course an oversimplification, but we do have the DNA of Neanderthals, Denisovans, Homo sapiens from different sites around the world, and anchors from our ape relatives, the chimps. But often, when I look at the sequences (CGACGGAATACG... and so on - see this image below (from Nature where a standard human sequence -top row "Reference"- is compared to Neandertal sequences in the bottom two rows), I wonder how true and accurate are our "reconstructions". Which base mutated first, which later?...

And also See this image, which compares Denisovan, Neandertal and some apes and monkeys.

Trees are created by computer programs that use "assumptions" and theoretical considerations built into them by the scientists that programmed them. They supposedly work using statistically sound calculations, which are so complex that I doubt anyone can verify them without the help of computer software... so maybe some bias is built into them, for instance assuming that AAAAA is the "original" genome in our distant planet, or here, assuming that the DNA of an African is "older" than that of an Amerindian...

I don't believe in snake oil, but I do believe that we should look at facts with open eyes (like the two trees that can be built using those nine sequences in our distant planet. Same data different conclusions) and not be biased by a prejudice (prejudice = Pre Judgement, we use the data to prove what we believe to be true, not to prove the facts...).

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2014 by Austin Whittall © 

The recurrent 9-bp deletion in mtDNA (more)

Today's post adds some more information to yesterday's post on the 9-bp deletion in human mtDNA's Coding Region.

More Europeans with the deletion

My previous post menitoned African, Asian, Amerindian and two Portuguese carriers of the 9-bp deletion, today I will add some more European data:

1. Haplogroup I, from Italy.[1] and Haplogroup I, from Spain [2]

2. Another paper (Thomas et al., 1998) reports two individuals from Glasgow (2 in N=115), Scotland (forgive my lazyness, but this paper was published in 1998 and does not specify the haplogroup it does mention the mutations, so those interested may classify these two persons). I will quote this paper extensively because it is intereresting:

The paper found these two persons who were unrelated and noticed that their mtDNA Control Regions were quite different, which lead them to conclude that the mutation was not recent and predated the arrival of H. sapiens to Europe:

"The sharing of these rare polymorphisms suggests that the 9-bp deletion in these two individuals stems from a single mutation event. The CR sequences of these individuals di¡er by eight nucleotides, suggesting that their common ancestor was not recent and probably lived before the arrival of modern humans in Europe."
[...]
"date their most common ancestor to about 50 000 years before present [...] it is interesting to note that the estimated date is earlier than the arrival of modern humans in Europe." [3]

The paper points out that such an ancient origin means that "this polymorphism should be present elsewhere in Europe and we suggest that it would be rewarding to do additional surveys for the deletion" [3], as mentioned above, it is found in Europe but at very low frequencies which supports its ancient origin.

The paper does not contemplate the possibility of the "one-repeat" sequence (that is, the 9-bp deletion) being the archaich ancestral state while all repeat sequences (2, 3 and 4 repeats plus insertions of partial segments of the 9-bp sequence) are the mutations. It considers the basal state as the "tandem repeat" and all mutations (the deletion) or multiple repeats as derived from the tandem sequence.

They do however note that "In this study, we have not addressed the effect of different repeat numbers on general fitness or replication efficiency." [3], which as I mentioned in yesterday's post is something to be considered since it appears that the 9-bp deletion is correlated to certain diseases.

The authors also found one female individual (1 in N=115) carrying not one, but several insertions of the CCCCCTCTA motif and studied her maternal lineage and concluded that the individual was heteroplasmic.

Heteroplasmy.
A heteroplasmic individual has more than one (hetero = "various", plasmy = "kinds") type of mtDNA in his or her cells.
Our cells have hundreds of mitochondria and they mutate in a random fashion. Since they cannot repair their mtDNA, these mutations accumulate over time in our cells which contain a mix of mutant and wild-type mtDNA. A woman carrying the heteroplasmy will transmit them in variable amounts to her offspring. The outcome can be healthy children or on the contrary, a child affected by a terrible disorder.

They studied this woman's mother, aunt and grandmother and found that the frequency of the different repeat values which was conserved from grandmother to daughters (her mother and aunt), changed in one generation (from her mother to hereself).

The Four repeats increased at the expense of the three repeat mutation, which is at odds with the population frequencies we mentioned in our previous post, where the 3 repeat is more frequent than the 4 repeat. Probably in heteroplasmic individuals some other factors such as natural selection are acting on the relative frequencies of repeats.

Comments

Detecting two Southern Mediterranean individuals belonging to Hg. I carrying the mutation while others do not carry it [2] (samples from Spain and North Africa) seem to point to a common origin of this mutation in a specific group of Hg. I individuals before their arrival to the region and their later spread to Spain and Italy.

In a similar manner, the two Scots with the deletion also suggest a pre-European origin for the mutation ~ 50 kya. [3]. This is indeed interesting as it supports my hunch that the deletion is indeed archaic and found in many haplogroups.

The heteroplasmic woman is also interesting since she shows that the mutations tend to an increase in repeat numbers and not a decrease (i.e. a deletion from tandem to one-repeat), which means that it may not have originated independently in different populations, but once, long ago in Africa and survives until now as a trans-haplogroup trait, which if carefully analysed would allow us to detect the inconsistencies in our current haplo-trees!

Sources

[1] Torroni A. Petrozzi M., Santolamazza P., Sellrro D., Cruciani F. and Scozzari R., (1995). About the "Asian"-Specific 9-bp Deletion of mtDNA. Am. J. Hum. Genet. 57:507-508, 1995
[2] GenealogiaMolecular.com. (2007), Linajes mayores del genoma mitocondrial trazan antiguas expansiones humanas (in Spanish)
[3] Mark G. Thomas, Charles E. Cook, Kevin W. P. Miller, Michael J. Waring and Erika Hagelberg, (1998). Phil.Trans. R. Soc. Lond. B 955-965

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2014 by Austin Whittall © 

The recurrent 9-bp deletion in mtDNA

AAs you know, different markers (or SNP mutations) are used to identify the haplogroups into which we classify human mtDNA these markers are assumed to have arisen by chance in a purely random manner within a population, and that they are not lost as they are passed on from mother to daughter generation after generation. This allows us to trace the relationship between different populations by verifying which markers are present and which are not.

Markers are therefore mutations, and they can be found in different regions of the human mtDNA. Today's post will look at the mtDNA's "Coding Region" and in a very particular mutation which is actually a deletion (part of the sequence found in some populations is absent in those carrying the deletion), which was lost due to the random mutation: the 9-bp deletion (9-bp d).

The 9-bp d in different populations around the world

First we will define the 9-bp deletion as a "length variant in the intergenic region between cytocrome c oxidase subunit II and the mitochondrial tRNA for lysine (COII⁄tRNALys)" [3]. What this means in a physically sense is the loss (deletion) of one (1) of the two (2) tandem repeats of the sequence "CCCCCTCTA" between nps 8281 and 8289. It is known also as the polymorphic anthropological marker MIC9D.

On the 9-bp sequence
Most people carry two (2) copies of the 9-bp sequence "CCCCCTCTA". This is generally assumed by mainstream science to be the ancestral state.
State which can be modified by mutations such as insertions of one or two additional 9-bp sequences, or by repeats of additional cytosine (C) or deletions and also by the loss of one full sequence (the 9-bp deletion). [12]

It was first detected among Asians by Wrischnik et al. (1987) [4], but later it was found at high frequencies (77% to 100%) in other populations in Polynesia such as the Maori, Niuean, Samoan and Hawaiian as well as among Australian Aboriginals:

The Polynesian branch seems to be associated with the influx or people from Taiwan who migrated southwards across Indonesia and led to the peopling of Polynesia. [5]

Another region where it is found is in the Indian subcontinent and insular areas next to it, where it is carried by southern Indian populations at frequencies that range from 68% in Yandi and 44% in Irula to 24% in Nicobarese and 4% in Siddi, [3] and 21% in Koyas and 3% in Chenchus. [2]

The Asian variant is found at frequencies of 15.6% (Thailand), 40% (Taiwanese Han), 20% (Hong Kong and Shanghai), but 3.3% in Shandong and Uygur populations. [5] In China it averages 14.7% (ranging from 0% to 32% in different ethnic groups). [15]

In addition to these Asian groups, it is also found in Africa and the Americas (as I said, it is global)

Africa

Among sub-Saharan Africans it is found at a frequency of 8.8% (or 81 individuals out of n=919), who carried it in different haplotypes (41 haplotypes in total).

The African Khoisan do not carry it, and is not frequent or absent among the people of the Western and Southwestern Africa. Pygmies and other populations in Central Africa, Malawi and Southern Bantu-speakers do carry it, suggesting according to orthodoxy a Central African origin and a recent dispersal to other regions. [1]

America

In America it is found in those populations belonging to mitochondrial haplogroup B, where it serves as a marker. It is one of the four founding lineages of Native Americans, but also has the peculiarity of being the only mtDNA haplogroup to cover the Pacific: it is found on both shores of the Pacific Ocean, in North and South America, Polynesia, Indonesia, Thailand, the Philippines, China and Japan.

Global yet different at a regional level

This global range is quite unusual since it spans different haplogroups.

The African and Asian (East Asia and India) mtDNA haplotypes are distinct and separate (see Fig. 2 in [3]).

A later study by Kivisild et al., (2003) [2] covering certain Indian populations noted that their 9-bp d was different in its haplotypes to the other Asian and African deleted lineages.

To study the relationship between the different populations carrying the 9-bp d, the following tree (Fig. 4 in [5]) is very useful. It shows that Africans (both those with the deleted 9-bp and the non-deleted populations) cluster together away from the rest.

The Australian Aboriginals and Indians without the 9-bp d are also distinctly separated on the tree from Indians and Australians carrying it.

The Chinese "cluster I" is very close to the non-deleted Chinese, which suggests a recent local origin.

The Amerindians and deleted Amerindians are very far apart.

There is a distinct separation which has three branches: (a) with all of the 9-bp deleted Thai, Cambodian and Nicobarese, (2) the Taiwan aboriginal, Amerindians, PNG, Indonesians and Polynesians and (c) the Chinese "cluster II".

The tree which does not specify the Haplogroups of the different populations clearly shows thre clusters which display the proximity of Amerindians, Asians and Melanesians; of Africans and of Aboriginal Australians.

Diseases linked to the 9-bp locus deletion

The locus of this 9-bp segment seems to have some strong correlation with the individuals' health: A study by Yiqui Jin et al, (2012) [6], links it to a form of cancer (Hepatocellular carcinoma or HCC). This is a very common form of cancer worldwide. The scholars found that the presence of the 9-bp deletion increases the risk of cancer (they call the deletion "one repeat", which is technically correct since the normal form is a tandem repeat [two repeats], so a deletion menas a "one repeat"):

"carriage of 9-bp one repeat fragment was associated with a significantly increased risk of developing HCC
[...] the frequency of 9-bp one repeat [that is: deletion] in our control samples (16.0%) was similar to that of previous reports in Chinese (14.7%). However, this frequency was 22.1% in the HCC groups, indicating a significant association between the 9-bp one repeat and the incidence of HCC." [6]

The reason for this is somehow linked to microRNA (miR) -mediated regulation because the deletion lies in a binding site region for certain miRs.

The locus is also associated with several other diseases: ataxias, idiopathic disorders, polycystic ovary syndrome, dilated cardiomyopathy, etc. [7].

If there is disease risk associated to the presence of one repeat, it seems logical to assume that natural selection would act to eliminate this mutation whenever it appears. However this does not appear to be the case, it is found in old (> 50 ky old haplogroups) with a global range.

Why? The deletion must provide its carriers with a positive side effect. But I have not managed to find any paper mentioning it.

The deletion and its origin

We have seen above that it appears all across the World (though at very low frequencies in Europe) and is considered by orthodox science as a marker that appeared independently in different human populations, yet despite this mutability it is neverheless used as a marker to define mtDNA haplotypes!:

The 9-bp d as a haplotype marker. From [9].

It appears in several haplotypes as markers in some, as a mutation in others:

• B ^[9]
• B6 ^[9]
• T2b15 ^[9]
• T2f ^[9]
• L0a2 ^[10]
• W ^[11]
• H ^[15]

the fact that it has such a high recurrence, meaning that it is found in different populations and in different mtDNA haplogroups around the world is very odd for a marker, since we would expect it to be distinct and appear only once and remain immutable.

Furthermore, the 9-bp d appears in the Coding Region and "Coding-region SNPs, [are] used as reliable markers to define Hgs because they are considered stable evolutionary events" [8]. But as I pointed out above, it seems that they "are nevertheless not entirely stable" [8] and a paper (Behar et al., 2007 [8]) shows that this instability leads to "inconsistent" haplogroup typing in 3.2% of the samples:

The 9-bp deletion is the major contributor to these inconsistencies, with an overall frequency of 1.2%. Which is quite logical since as we have seen above it appears in at least 7 haplotypes, and if the sequencing does not cover many markers an incorrect haplo-typing is surely to occur. [8]

Origin

According to Sodyall, Vigilant, Hill, Stonekin and Jenkins, 1996, [1] it arose independently in different populations around the world: the "locations and frequencies of variant sites" in Africans and Asians are not the same (different haplotypes), which suggests that ir appeared independently in both regions, and perhaps more than once in Africa [1] and twice in Asia [12].

Behar et al., 2007, also interpret this 9-bp deletion as a Homoplasy (which combines the Greek words for "same" -homo- and "form, shape" -plasis), in other words the chance duplication of a feature with a similar shape, structure or function but with an independent origin and no shared ancestry.

Homoplasy is found in the shape of the fins of fish, dolphins and Ichthyosaurs or wings in bats, birds, pterodactyls; which arose independently in separate and separate lineages of different kinds of creatures (reptiles, fish, mammals and birds) by chance, yet the shape is similar because of their similar function.

The independent origin of the 9-bp d is supported by the fact that the "most likely mode of origin of the deletion, namely, slipped mispairing during replication, can occur (and recur) in every population..." [12].

However Kivisild et al., (2003) [2], apart from suggesting an independent origin of the deletion also consider the possibility of "an ancient common root" [2]. And this is indeed interesting, so we will look into it below.

Ancient relict state

Behar et al., [8] found that out of a total amount of mutations of 592, there were 343 transitions, 199 tranversions, 35 insertions and 15 deletions. So the most rare mutation is a deletion (2,5% frequency). Yet for mainstream science, the deletion is the most obvious cause for the presence of one instead of two "CCCCCTCTA" sequences. But insertions are over twice as frequent as deletions. And in fact the tandem repeat is more common than the one-repeat (or deletion). Furhtermore, as we will see below, triplicate, cuadruplicate and even intermediate (expansion and contraction of the tandem repeat by loss or addition of individual bases) are present. What does not exist is a total deletion of the segment implying that the trend is towards growth not contraction.

Its low frequencies may be due to its association with illness, but since it has not been selected to extinction, it must provide some benefit to those carrying it.

The fact that it is found repeatedly in several haplotpes and especially in the ancestral African L0a2 line is an indication of its ancient origin. The most parsimonious explanation is that it appeared in the African populations (in fact their placement on the tree above, away from most groups clearly shows a distinct origin) and was lost in some of the downstream haplogroups due to a "9-bp insertion" that erased the "one repeat" and created the now most common tandem repeat, found in most populations. It is likely that a small number of individuals conserved the archaic deleted state. Which, as we see above makes its carrier prone to certain diseases and therefore maintained a low frequency of this variant.

It moved with humans into Asia, generating the Indian - Australian cluster and then (this resembles the Y chromosome C hg. expansion!) into South East Asia, Eastern Asia, Siberia and America.

It grew among mtDNA haplogroup B in such a way as to become its marker. This must imply some yet unknown benefit in carrying the 9-bp deletion.

Mutations

The Sakha from Siberia have a high frequency of mutation associated to the 9-bp locus: 1.8% deletions, 1.2% triplications and 0.26% four repeats all others have the usual "tandem repeat". These mutations were found among different mtDNA haplogroups, where B, T and W were the most common for deletions. [11]

Interestingly "No neighboring populations have been reported to carry a non-haplogroup B deletion [...] suggesting that shared ancestry or admixture or both are unlikely explanations for the presence of these polymorphisms in the Sakha. The identification of high levels of variation may be a function of the large sample size [N = 779] and the in-depth analysis of all derived polymorphisms" [11].

The comment on sample size is a very positive one, in my opinion most studies are based on very small samples and their sequencing is not done thoroughly, so many differences go unnoticed.

The authors conclude that these mutations are due to "multiple recent events" [11], but in my opinion the lack of admixture or shared ancestry means that these are archaic or relict mutations that have been carried by the Sakha and their ancestors for many generations.

The "three repeat" is very infrequent globally, besides the Sakha with 1.2% frequency, the triplicate insertion has been found among Chuckchi (Siberia), Nepalese Tharu (in 1 out of N=127) [12][13], a Portuguese Brazilian belonging to hg. H (1 in N=245), and Portuguese in Europe (2 in N=96) [14]; it is also very rare among Chinese at 0.25% (3 in N=1218) [15].

The normal tandem (two repeat) is also found mutated in some Indian populations. It appears as "contracted" among the Caste populations (Brahmin and Jalari) at an average frequency of 1% and as an "expanded" two repeat mutation among tribal populations (Khonda Dora and Santal) with a 3% frequency. [3] The expansion is due to a T to C transition followed by a replication of the C component. The contractions have "length variant intermediate between 1 - 2 repeats in length". [3] They have also been detected in Borneo, the Philippines and Samoa. [3]

Selective pressure is surely acting on this ancient mutation maintaining it despite the current prevalence of the tandem sequence.

Closing comments

This is a trait found in the most ancient lineages in Africa and also among other populations at low frequency. Yet one of Amerindians founding mtDNA haplogroups is identified with this specific mutation and it is found at very high frequencies among native Americans.

It causes some propensity to certain diseases yet despite this fact it has not disappeared, this implies that it has some yet to be detected positive aspect for those carrying it.

Deletions are infrequent yet this one is global. It surely signals its ancient origin. Furthermore, its presence in seven different haplotypes indicates in my opinion that the basic axioms of the mtDNA clock and neutral mutations used to set it are not to be taken seriously.

Its dispersion across the globe mimics that of Y chromosome hg. C, which I have suggested may be an ancient (Homo erectus) lineage and much older than mainstream science suggests. Could the 9-bp deletion be an archaic hominin trait?

Clearly some more research is needed here. I will keep my eyes open for additional data.

Sources

[1] H Soodyall, L Vigilant, AV Hill, M Stoneking, T Jenkins, (1996). mtDNA control-region sequence variation suggests multiple independent origins of an “Asian-specific” 9-bp deletion in sub-Saharan Africans. Am J Hum Genet, 58 (1996), pp. 595–608
[2] Kivisild T, Rootsi S, Metspalu M, Mastana S, Kaldma K, et al., (2003). The genetic heritage of the earliest settlers persists both in Indian tribal and caste populations. Am J Hum Genet 72: 313–332. doi: 10.1086/346068
[3] Watkins W. S. et al., (1999). Multiple origins for the mtDNA 9-bp Deletion in Populations of South India. American Journal of Physical Anthropology 109:147-158
[4] Wrischnik LA, Higuchi RG, Stoneking M, Erlich HA, Arnheim N, Wilson AC., (1987). Length mutations in human mitochondrial DNA; direct sequencing of enzymatically amplified DNA. Nucleic Acids Res 15:529–542
[5] Yong-Gang Yao, W.S. Watkins and Ya-Ping Zhang, (2000). Evolutionary History of the mtDNA 9-bp deletion in Chinese populations and its relevance to the peopling of east and southeast Asia. Hum Geet (2000) 107:504-512 doi: 10.1007/s004390000403
[6] Jin Y., et al., (2012). The mitochondrial DNA 9-bp deletion polymorphism is a risk factor for hepatocellular carcinoma in the Chinese population. Genet Test Mol Biomarkers. 2012 May;16(5):330-4. doi: 10.1089/gtmb.2011.0208. Epub 2012 Jan 27
[7] Borgione E., et al., (2013). The 9-bp deletion in region V of mtDNA: a risk factor of hearing loss and encephalomyopathy in Caucasian populations?. Neurol Sci (2013) 3$:1223-1226 doi: 10.1007/s10072-013-1297-9.
[8] Behar DM, Rosset S, Blue-Smith J, Balanovsky O, Tzur S, et al., (2007). The Genographic Project Public Participation Mitochondrial DNA Database PLoS Genet 3(6): e104. doi:10.1371/journal.pgen.0030104
[9] www.phylotree.org
[10] www.dnahaplogroups.org
[11] L. Tarskaia, R. R. Gray, B. Burkley and C. J. Mulligan, (2006), Genetic variation at the mitochondrial DNA 9-bp repeat locus in the Sakha of Siberia. Human Biology 05/2006; 78(2):179-98. doi: 10.1353/hub.2006.0038
[12] Redd A. J., et al., (1995). Evolutionary History of the COII⁄tRNALys Intergenic 9 base pair deletion in human mitochondrial DNAs from the Pacific. Mol. Biol. Evol. 12(4):604-615.
[13] Passarino G1, Semino O, Modiano G, Santachiara-Benerecetti AS, (1993). COII/tRNA(Lys) intergenic 9-bp deletion and other mtDNA markers clearly reveal that the Tharus (southern Nepal) have Oriental affinities. Am J Hum Genet. 1993 Sep;53(3):609-18.
[14] Alves-Silva J., Guimaraes P., Rocha J., Pena S. and Prado V., (1998). Identification in Portugal and Brazil of a mtDNA Lineage containing a 9-bp triplication of the intergenic COII⁄tRNALys Region Hum Hered 1999,49:56-58
[15] Yao YG1, Watkins WS, Zhang YP, (2000). Evolutionary history of the mtDNA 9-bp deletion in Chinese populations and its relevance to the peopling of east and southeast Asia. Hum Genet. 2000 Nov;107(5):504-12.

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2014 by Austin Whittall © 

Biases in Genetic Models that are generally overlooked

Before continuing with the Y chromosome C haplogroup and its peculiar global range, I want to dedicate today's post to "Models" in general (and those used in genetics in particular), and how they shape the way we believe things work.

As an engineer I have a clear notion that a model is a simplification of the real world, we make assumptions and find that models allow us to make predictions that are reasonably similar to the real world. This allows us to build a bridge that will not collapse and yet use a minimum amount of steel, or shoot a cannon ball from A and hit a target in B. The models can be more complex (relativity is taken into account to make your GPS work correctly and quantum mechanics in all things electronic), but all of them are just that: a "model", an approximation to reality, not reality itself, merely a simplification of the real world.

Models and Genetics a critical overview

Let's review some causes of errors and the hidden biases in genetic models and methods:

Populations

Some studies use strange populations such as "MXL, Mexican Ancestry from Los Angeles USA" or "ACB, African Carribbeans in Barbados", [10] which are far less significant than a native aboriginal individual in his or her homeland: such as an "Alakaluf from Magallanes, Chile" for instance.

What conclusions can be reached on the peopling of America by looking at the genome of a person living in LA, of Mexican ancestry? Mexico has several native groups, later overalaid by Spanish conquistadors (Spanish are themselves a mixture of many ethnic groups... aboriginal Iberians, Basques, Celts, Romans Carthaginians, Arab invaders, Goths, etc.), the African slaves they brought over to work in their plantations and a touch of other European and Southern and Mesoamerican groups. In other words the Mexicans are very admixed population. MXL are actually irrelevant or former slaves of African origin living in Barbados!

Then we have studies that ignore the New World completely. Samplings cover populations in Africa and Eurasia (sometimes including Australia and PNG), seldom the Americas. This of course limits the usefulness of these studies and maybe conceals interesting findings that will remain ignored until American samples are contrasted with Old World ones.

The Sampling within the populations

Once the populations have been identified (with all the caveats mentioned above), its ancestry is studied by means of a sample of individuals that in theory (but not in practice) is drawn from it in a random manner.

In other words a sample of size "n" is taken from a population with a size "N". In most populations "N" is several orders of magnitude lager than "n" (imagine a sample of 200 Italians from Tuscany out of the 61 million people living in Italy). This implies that we may have a sampling bias, and leave out some unique or even critical genetical sequences that appear at low frequencies in a given population.

In other cases the sample is not a random one (Ascertainment bias); for instance, in the case of small tribes: the sample "n" is small because "N" is also small and perhaps encompasses a clan or family group. Thus diversity is low and, as we will see below, calculations based on these samples will be affected by this sampling bias.

Ascertainment bias also arises when researchers take samples from databases in which some populations are missing, or some regions are under represented. Or from samples obtained from volunteers (i.e. a University class) which are definitively non-random samples.

The Typing of Haplogroups in those samples

Once we have sample of our population, then we sequence the Y chromosomes (or in the case of mtDNA, the mitochondrial DNA) for known markers (remember this: known, we will get back to it later) and type the individuals based on these markers. We don't read the whole sequence of 60 million base pairs in a Y chromosome and compare them all. Instead we choose certain ones which we believe are markers and base our analysis on these markers. This introduces another ascertainment bias: the choice of markers is not random.

Look at it this way, we have a million books printed in different languages, we take a sample of fifty from the pile of books and compare them based on a few of the words printed in them: If we find the word "Chapter" we will place them in the English group, "Capítulo" in the Spanish one, "Haupstück" in the German one. All the other words are ignored. We may even find a book printed in Armenian that by chance carries the word "Chapter" and we will place it in the "English" group and ignore that an "Armenian" group even exists. We may have not sampled a Chinese book and also ignore its existence.

Yes, the analogy is faulty (well it is a model after all), the markers we look for in genetics are placed in specific locations within a chromosome. We compare the markers in specific positions with each other to define similarity. In the books example the analogy falls through because the word Chapter can appear on any page and not on a specific page. Yet, what I wanted to point out is that even though there are millions of base pairs in a chromosome, we only use a handful to define haplogroups.

Let's look at this in depth:

Y chromosome SNPs, the Haplogroups

First of all, the Y chromosomes are sequenced and the Single Nucleotide Polymorphism (SNP) that identifies a given haplogroup (hg) is identified. An example of an SNP is shown below:

• AAGCCTA - ancestral
• AAGCTTA - derived

The fifth nucleotide, the "C" in the ancestral variant, mutated to a "T" in the derived variant. This SNP ocurrs in a specific location. These mutations are assumed to be apparently rare, and once they happen, they remain in the DNA and are inherited by all descendants of the mutated individual. This allows us to build trees based on which SNPs are present in the populations.

The SNP that identifies the whole Hg C is marker RPS4Y711. The different haplotypes also have their own markers: M8 (for C1), M38 (for C2), M217 (for C3), M93 (for C3a), and so on.

But, as I mentioned in a previous post, it is possible that these SNP mutations revert (in that post I mention the Motala remains sequenced for a Q haplotype which lacked a key marker but had all the others), so this may also introduce additional errors.

Haplotypes and Paragroups

So, in our sample which already may have a sampling bias, we will find that most individuals will belong to a given haplogroup (i.e. Y chromosome hg C) and to a given haplotype (i.e. if they tested positive for marker M38 we will be confident that they belong to the C2 haplotype).

But... Some cases will not test positive for the known markers (i.e. M8 -excluding C1, M217 -excluding C3, M347 -excluding C4 or M356 -excluding C5), so we will assume that they form a paragroup that includes all C-other lineages, and we will identify it as C* (with an asterisk).

So these pargroups clump together a large variety of haplotypes for which we have not yet identified specific "unique" markers which would set them apart as new haplotypes.

Hidden diversity

This is not a trivial matter. For instance paragroup C* is found all across the Eastern edge of the Old World, from Australia, across Indonesia, China, Japan and India, to Bering at frequencies that range from 0.3% to 10% of the population. (some are even higher: C3* among Mongols reaches a frequency of 18% [6]).

To assume that they all belong to the same group is erroneous, the paragroup masks subclades (haplotypes) which have not yet been discovered: Maybe C* in Australia is a not yet identified C8 hg, while C* in Central Asia is a yet undiscovered C9 hg for example. In other words, there is hidden diversity out there waiting to be discovered.

Bias in the choice of the markers

The SNPs are not chosen in a random manner, they are defined by geneticists to type haplogroups. This means that the real distribution of polymorphisms in a population may differ from those shown in a study. The reason is simple: The genotyping arrays (or chips)used to identify the markers contain biased sets of pre-ascertained SNPs. These SNPs tend to be older than the majority of the SNPs in a given population, and are found in many populations besides the one being studied. These hand-picked SNPs act as sieves, classifying the samples and causing alterations such as "[shifts in] allele frequency distributions ... towards intermediate frequency alleles", furthermore, "estimates of linkage disequilibrium are modified" [11].

In other words, this increases the frequency of the most commonly polymorphic loci and eliminates other markers (loci that are less polymorphic in the screening panel). This ascertainment bias in the SNP arrays strongly skews the estimates of genetic diversity by ignoring those that are not included in them.

Variety within the Haplotypes

Once the haplogroup and haplotype have been identified, we can take a look at the Microsatellites to check out even further diversity. Microsatellites are repeats (2 to 6 nucleotides long) that repeat "n" times ( n= 5 to 100). An example: the sequence "AT" repeated 25 times in a row (this is expressed as follows: (AT)₂₅).

These microsatellites are found across species and mutate quicker than single point mutations (SNPs) and for this reason they are used as markers to define the subclades within haplotypes. An example would be a mutation in (AT)₂₅ to (AT)₂₄ or to (AT)₂₆.

In the case of our Y chromosome, we will use special microsatellites known as Short Tandem Repeats (STR). These are named as "DYS-Number" (i.e. DYS393) which indicates a position for the STR. The STR, is a series of repeats of a dinucleotide (two nucleotides).

Below is a real example, for haplogroup C, for four individuals, two from Colombia, one Korean and a Kalkh:

A real set of STRs for four individuals. Copyright © 2014 by Austin Whittall

We can see in the image above that some DYS markers differ in the quantity of repeats (they are shaded pale blue and yellow).

The Evolutionary sequence

Now come the questions: Does the Korean derive from the Colombians or is it the other way round? What about the Mongolian? Note the differences at DYS392 and DYS393 between Koreans and Colombians (in yellow) and the difference between Colombians and Mongolians at the other DYSs (pale blue).

Without an A priori theory you can't answer that question just by looking at the repeats. Some additional assumptions are necessary, and computer programs are used to build the phylogenetic trees that link these individuals.

We could assume that humans came from Asia and peopled America, so the Americans are more recent than Asians. And as Koreans are Asians, they predate the Colombians so they accumulated mutations, passing from 11 to 12 in DYS392 and 13 to 15 in DYS393. So far so good, but then we have a Khalkh, from Mongolia, who are also Asians, but have less accumulated mutations than the Colombians or the Koreans. This way of comparing STRs is faulty.

Building a Phylogenetic Tree

The individuals are placed on phylogenetic trees using other assumptions that consider the differences between individuals, and are based on the different STRs, but it uses a different reasoning process to the one used above.

Distorting elements

Nevertheless we must remember that each DYS may have its own mutation rate, so if there are several DYS that differ, then they all have to be considered to calculate the "distance" between individuals.

An additional complication is that the "real" evolutionary history of any given set of individuals may differ from the "inferred" evolutionary history. As can be seen in the following image where only three (3) mutations out of twelve (12) are detected during analysis. The other nine (9) are ignored and remain undetected. This affects the estimations on divergence since the mutations are underestimated and the split time between the ancestor and its descendants is underestimated:

How mutations are Underestimated. Adapted from Fig. 2 in [4]

So "corrections" are introduced such as the Jukes-Cantor Model [4] which fiddle with the equations used in the models to make them fit better to reality.

So, how are the differences calculated?

Computer algorithms add even more assumptions

Algorithms are used. They are run on computers and compare the individuals in a pair-wise manner. There are several algorithms, each with their pros and cons. The two basic classes are:

• Distance-Based Methods - Neighbor Joining (NJ), which initially form an unresolved star-like tree and compare the branch length sum in a pairwise manner. It then groups as "close" those that have the minimum sum. This pair is linked in a branch and then the process begins again, iterating until all individuals have been grouped.
• Maximum Parsimony method, uses certain features (substitutions in the sequences) to work out a most likely evolutionary relationship among individuals. It builds the tree using the least substitutions chain from the common ancestor to the individuals being located on the tree. So it scores each possible option and minimiizes the mutation number to buid the tree.

The trees are then rooted by comparing them with some outgroup species (ie. chimpanzees are used for human and hominin comparisons). Of course this requires the assumption that molecular clocks are valid and that the divergence date with the outgroup species is well known (more on this below - see clock ticking out of time).

An example: Batwing

Batwing [8] (acronym which stands for Bayesian Analysis of Trees with Internal Node Generation), is a widely used computer program for analysis of genetic data. It has some implicit assumptions that I list below which are not mentioned in the papers that use it, but which impact on the outcome of the program's analysis. By the way, the authors of the program clearly point out that "Natural populations are unlikely to satisfy BATWING's modelling assumptions" [8].

• The data is a random sampling from the population (we have seen above it is not usually the case)
• The population is panmitic (not frequent in human groups)
• Splitting between populations is instantaneous (actually it takes plenty of time)
• There is no subsequent migration events between populations (there is always posterior admixture due to migrations between populations that have split)

Batwing uses different mutation models, but the default setting is the Stepwise Mutation Model or SSM, which we analyse in detail below.

Comment, the TMRCA (time to most recent common ancestor), Ť, is calculated under the Simple SSM model using the expression: Ť = Δ ⁄ μ.

Where Δ is the average squared difference in the number of repeats between all sampled Y chromosome and the founder haplotype, averaged over STR loci, and μ is the Simple SSM mean mutation rate per generation averaged over loci. But, if, as we will see below SSM is not very reliable, then how can clade age estimates be reliable?.

The Stepwise Mutation Model

This Stepwise Mutation Model (SSM) [2] was proposed in 1973 by Ohta and Kimura and has been widely adopted as the model for microsatellite evolution:

Microsatellites are believed to evolve neutrally: natural selection does not influence the number of repeats so, the SMM premise is: "In one generation the repeat number can only increase or decrease by one, and the probability is equal".

But this assumption is not exactly so for several reasons:

• Actually, the probability of mutation is larger for longer microsatellites [1][3].
• A long set of repeats (n larger than 20) may cause physical instability in the microsatellites and hamper its further growth, actually leading to their contraction. [3][2]
• Some microsatellites are interrupted and have lower mutation rates.
• The repeat unit also influences mutation rates: dinucleotides mutate slower than tetranucleotides.
• The motif of the dinucleotide (i.e. TG vs. TA) also plays a role: certain motifs are much longer than others.
• Variable mutation rates (those that change the repeat by more than 1) are not uncommon and happen about 15 to 22% of the time [3], in other words, the model ignores a big chunk of mutations.

Add to this that insertions or deletions next to the microsatellites also influence their lenghts. [3]

Even the "neutrality" of satellites is questionable since some repeats take place in promoter regions and may influence protein building [3] and thus be subject to natural selection. Some microsatellite repeats have been linked to certain diseases (myotonic dystrophy, Huntingtons' disease, etc.) making their neutrality doubtful too.

There are also "point mutations" that interrupt a repeat; an example: (AT)₁₈ may suffer a chance point mutation where "A" mutates to "G" in position 10, causing the new sequence to be: (AT)₉ GT (AT)₈. Transormation which may go undetected in sequence analysis, altering the mutation rate estimates.

Panmitic populations

Last but not least, the SMS model assumes that the individuals come from a random sample form a single panmitic population of constant size "N", and this is not the case. [5] Application to expanding populations or those with mixing due to migration may provide different results.

A panmictic population allows random mating without any restrictions of any kind (due to age, genes, behavior, social, environment, etc.), which is seldom the case in human populations, past or present.

Migrations

When an ancestral population splits into two groups, they are subjected to two opposing processes (see image below):

How genetic drift and migration affect allele frequencies. Copyright © 2014 by Austin Whittall

• Genetic Drift. It arises because N_e (number of effective breeders) which contribute their genes to the next generation is smaller than the total population, so they pass on their genes only and since this is a random sampling process, the frequency of these genes will differ from that of the previous generation. The smaller N_e, the larger the drift. This effect accumulates with each successive generation and separates the diverging subpopulations as time passes.

• Migration. Exchanges between the populations as they diverge will limit the drift, keeping them similar. The proportion of migrants "m" if larger will have a higer impact on stability.

Comment on drift and lack of migration: The "Beringian Standstill" was invented to justify the strangely unique American haplogroups, completely absent in the purported Asian homeland of the Native Americans.
The Standstill theory first suggested by Bonatto & Salzano (1997) and perfected by Tamm et al., 2007, is based on a one-in-a-million "Founding effect" that isolates a group of "founding fathers" in Beringia, cut off from their Asian relatives and from the vast empty Americas by ice sheets for about 15,000 years. During this long period of time they mutated their Asian mtDNA and NRY haplogroups into new ones and then in a quick wave covered America swiftly so as not to allow new diversity to arise.
Furthermore, their Asian relatives all died off, leaving no trace on the Asian side of Beringia.
Yes, I know it sounds improbable, yet even though the odds are against this kind of event, several papers apart from Tamm et al, support the theory.

But let's get back to the Stepwise Mutation Model: it is quite weak, to put it mildly.

Summary: SMM is unreliable

All these phenomena make SMM a very rough approximation to reality, yet it is used as if it was 100% reliable!

Just as an example of this lack of reliability is the quote below (Nebel a., et al., 2001) [7]:

"the behaviour of DYS388 appears to be inconsistent with the SMM, as was shown in two populations of Middle Eastern origin. Additionally, another widely used microsatellite, DYS392, has recently been demonstrated to deviate from the SMM" [7]

The Clock that ticks out of time

I have posted on the useless genetic clocks in the past, so I will not bore you, just highlight my previous objections to clocks:

Divergence from Chimps. Scientists devise clocks to calculate mutation rates. To do so we estimate the divergence dates of the human line from the chimpanzee line. But the date of this event is uncertain, and has been increasing since the 1970s from an estimated 5 Mya to 6 - 7 Mya, and in June 2014, to 13 Mya [9], this recent change should surely impact on the dating of human origins!

Assumptions are also made regarding the duration of a generation (what can we know about how long a generation was 50 kya? did females mature earlier or later? what about males? was it 27 years or 35?). Population sizes and their trends (expansion, migration, admixture with other groups as well as bottlenecks and founder effects) also should also be factored in.

When those clocks are calibrated against real mutation rates measured in (again a discrete sample) familes over the last few hundred years strong discrepancies arise: these family (pedigree) calculated mutation rates usuall differ from the former ones (evolutionary). But these differences remain unexplained in the papers. They merely show both figures but avoid explaining the causes (i.e. the mutational clock does not tick at a regular pace).

The mutation rates are also calibrated against the estimated dates for the peopling of certain regions based on the information provided by archaeology. i.e. 40 kya for Australia or 17-20 kya for America. However just by looking at the published error margins we can see the uncertainty involved in these calculations.

Diversity is taken as an indicator of antiquity so if Region X has a large variety of haplotypes while Region Z has fewer, population in Z is assumed to be younger. But, actually what happens is that if the people in "Z" are a subset of population from Region "X", the fact that they are a subset means that they will have less diversity than the original group. This does not mean that they are more recent, it means that they left certain genes behind. Add to this the pressure of natural Selection (and chance i.e. genetic drift) and the genes of certain individuals within the subset at "Z" will get lost too. So if we measure "X" against "Z" by their diversity we would incorrectly judge "Z" to be more recent, when they are really just as ancient as "X".

Frequency, Migrations and antiquity

Often the current distributions of haplogroups occur at differing frequencies in certain territories. What does this mean? That the less frequent "A" hg is a recent arrival of a small group carrying it, entering the territory of the prevailing "B" hg.? Did "A" exist in the same population as "B", but in very low frequencies, and those have been maintained or even decreased?

Or is "A" an ancient colonizer that suffered attrition over thousands of years and has been gradually losing ground to better equipped newcomers with hg. "B"? Maybe "A" and "B" were found in equal proportions in the original colonizers but "B" grew due to genetic drift or natural selection...

Questions like those are seldom asked or answered in mainstream papers. It is clear that the choice of the correct answer requires an in depth analysis which is not found in the academic literature (I have read tens of papers and these matters are not even addressed).

The low diversity among Amerindians is always invariably attributed to a founder effect or a bottleneck during the peopling of America event. The massive death of millions of Natives (virtually a genocide) during the process of discovery and conquest of the New World between 1492 and 1560 is ignored. Disease and war acted selectively wiping out tribes without leaving a trace of them, but this issue is simply ignored and the "lack of diversity" is assumed to be due to the original peopling event some 15 kya.

Dogma

The unidirectional migratory route from Africa to the World has some inconsistencies which can only be explained by back-migrations. These into Africa migrations are reluctantly accepted by orthodoxy but, fortunately, are gradually altering the OoA picture with a more parsimonious explanation. Sometimes I get the feeling that OoA is supported because it is politically correct and assuages the guilt complex of the Western world for the tragic crimes of Slavery and colonialism perpetrated against Africa.

Two issues requiring a serious review are: The East Siberian void of putative ancestors to the Amerindians, which remains unexplained and The "Beringian standstill" justification for Amerindian uniqueness, which also requires a critical analysis due to its improbability.

Closing comments

What I have tried to express in today's post is that there are many assumptions underlying the "facts" expressed by mainstream geneticists regarding human diversity and evolution.

Models are simple representations of reality and not reality itself. They should be taken as such and not as truth written in stone.

Algorithms and simulations run on models are only as reliable as the models they are based on. And we have seen the flaws in some of these models and programs. Flaws that introduce errors in their output yet are not explicitly mentioned in the papers that basethemselves on them.

The complexities in the statistical assumptions mentioned in papers (those pages or paragraphs, full of equations that you skip when reading a paper) mask some very evident biases that skew the results and produce patterns that do not correctly reflect reality, which is richer and much more varied than what these papers show us.

Sources

[1] Esra Ruzgar and Kayhan Erciyes, Phylogenetic Tree Construction for Y-DNA, Haplogroups.
[2]Amke Caliebe et al., (2010). A Markov chain description of the stepwise mutation model : Local and global behaviour of the allele process. Journal of Theoretical Biology 266(2010)336–342
[3] Peter Calabrese and Raazesh Sainudiin, (2004) Models of Microsatellite Evolution
[4] Yan Li, Phycs498BIO Assignment 2, How to Build a Phylogenetic Tree
[5] Valdes, Ana M. Slatkin M. and Freimer N., (1993). Allele Frequencies at Microsatellite Loci: The Stepwise Mutarion Model Revisited. Genetics 133: 737-749 March 1993
[6] Boris Malyarchuk, et al., (2010). Phylogeography of the Y-chromosome haplogroup C in northern Eurasia. Annals of Human Genetics (2010) 00,1–8 doi: 10.1111/j.1469-1809.2010.00601.x
[7] Nebel A., et al.,(2001). Haplogroup-specific deviation from the stepwise mutation model at the microsatellite loci DYS388 and DYS392. Eur J Hum Genet. 2001 Jan;9(1):22-6
[8] Ian Wilson, David Balding and Mike Weale, (2003), Batwing User Guide. See pt. 1.2.
[9] Oliver Venn et al., (2014). Strong male bias drives germline mutation in chimpanzees. Science 13 June 2014: Vol. 344 no. 6189 pp. 1272-1275 DOI: 10.1126/science.344.6189.1272
[10] www.1000genomes.org.
[11]Lachance J, Tishkoff SA. et al., (2013). SNP ascertainment bias in population genetic analyses: why it is important, and how to correct it. Bioessays. 2013 Sep;35(9):780-6. doi: 10.1002/bies.201300014. Epub 2013 Jul 9.
<(p>

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2014 by Austin Whittall © 

A shared Y chromosome lineage Neanderthals and Modern Humans

I am following up on some (there were quite a few) of the subjects that appeared while writing my most recent posts on the human Y chromosome.

Among them, are my very serious doubts regarding the markers used in classifying Y-chromosome haplogroups, the ages of the different sub-clades or haplotypes within them, and the different dates for common ancestors based on the estimated mutation rates for the human Y chromosome. As expressed in my previous posts, the methods, assumptions and convoluted statistical calculations seem far fetched and very uncertain. The fact that some confidence intervals are too large in my opinion invalidate the logic behind them, but then, I am a mere engineer not a geneticist.

The Y chromosomes of the Neanderthal

It is in this context that I read an interesting paper (Musaddeque Ahmed and Ping Lian, 2013) [1], from which I quote the following regarding Neanderthal Y chromosomes ... (bold font is mine):

"...the Y-chromosome of Neanderthals or paternal inheritance has yet to be examined. Comparisons of the Y chromosome sequence of Neanderthals with currently established Y-haplogroups for modern humans should provide some insights into the admixture hypothesis.
With respect to the recent finding of admixture of Neanderthals with non-African populations, the Neanderthal Y chromosome should not match the Y haplogroups A or B, as these haplogroups are the oldest of the clades and almost restricted to Africans and their descendants. Since haplogroup E is found in Africa, the Middle East, Southern Europe, and Asia, the Neanderthal Y chromosome may match this haplogroup, but it should not match the haplogroups E1b1a*, E2b1, or B2a1a, as they are specifically treated as Bantu expansion markers, while Neanderthals interbred only with non-Africans." [1]

I was very surprised by this paragraph, as I expected something similar to what we see in mtDNA, where the ancestral population of both humans and Neanderthals, split into two groups: one that would later evolve into modern humans, and another that would evolve into Neanderthals. Each would carry, in the beginning, the "original" archaic Y chromosome of the ancestral population, which, as each lineage accumulated mutations due to chance and positive selection would begin to grow in different directions, forming two distinct branches, which then in turn would continue branching as more mutations appeared.

So, the oldest human Y-chromosome haplogroups "A" and "B" found exclusively in Africa, were born from the branch leading to modern humans (the "root" of mankind), while, on a separate branch, (Neanderthals), the Y chromosome haplogroups of Neanderthals appeared.

However, towards the end of this post we will see that this reasoning may not be correct.

Nevertheless, mainstream science considers it to be the correct interpretation. According to Krause et al. (2007), [2] since the oldest common ancestor with human Y chromosomes dates back to 90 kya, and as Neanderthals and humans split long before that date, they "expect[ed] Neandertal Y chromosomes to fall outside the variation of modern human Y chromosomes unless there was male gene flow from modern humans into Neandertals"; in other words: Neanderthals would have their own peculiar Y chromosome haplogroups unless H. sapiens men mated with Neanderthal women and passed on their Y chromosomes to their human-neanderthal mixed male sons. But then these would not be pure Neanderthals but admixed ones, with 50% human genome in them and their Neander mom's mtDNA.

The following image shows what I expressed above, two branches: red for Neanderthals and black for us, humans, both splitting from our common ancestor:

Hypothetical Y chromosome phylogenetic tree for humans and Neanderthal. Copyright © 2014 by Austin Whittall

Actually, we have not yet found Neanderthal Y chromosomes among modern human genomes and this is indeed curious since we have admixed with them. It could be due to several reasons:

(a) They are present in such low frequencies that they have not yet been sampled. Imagine a 0.001% frequency, that is, only 1 in 100,000 men would carry it (35,000 in the whole world, a very small number indeed).
(b) They are very similar to modern haplogroups (as suggested by Musaddeque and Ping, 2013. [1]) and we see them but have not yet realized that they are Neanderthal.... (More on this below, because it seems to be a very interesting possibility).

While studying the Neanderthal genome, Krause et al., (2007) [2] found that two of their Sidrón Neanderthal specimens (Sidrón 1253 and Sidrón 1351c) were males. They then tested them at the five positions that mark the main nodes of Human branches (i.e. the splitting points of our Y chromosome haplogroups) in both Eurasia and Africa. They noticed that "all 15 Y chromosomal products for the five assayed positions show the ancestral allele" [2]. In other words, the ancestral or chimpanzee-like state was found. They did not find any specifically Neanderthal-like state. This is odd, considering that 7 million years separate chimps from Neanders, why would they both share the markers in their Y chromosomes and humans, a mere 300 ky away don't. It may be an error in their sequencing.

Admitting the paucity in their sampling and that these Y-chromosomes may appear in extremely low frequencies among modern human populations, the authors conclude:"These [Neanderthal] Y chromosome results must derive, then, either from Y chromosomes that fall outside the variation of modern humans or from the very rare African lineages not covered by the assay..." [2], in other words, the Neanderthal Y chromosomes may be similar to those of modern humans that were not screened for.

I must point however, that according to Figure S1, the "markers" chosen by Krause et al., screened only certain branches of the Y chromosome tree: (i) and (ii) the split that leads to all haplogroups downstream from the A clade. (B, C,.... R2). (iii), The one leading to B. (iv) The one leading to P, (and therefore Q), R1, R2. And (iv) the one leading to A2.

So the Neanderthal Y chromosomes could belong to Haplogroups A1 and A3, which were not screened for. Or, an option that is evident, but they did not consider it at all: Neanderthal Y chromosomes are identical to those of Modern Humans and cannot be told apart from them.

Actually, isn't it surprising that not one Neanderthal Y (or X with its mtDNA) chromosome survived until our days? We have a considerable amount of autosomal genetic content introgressed from our relatives, the Neanderthals, but their mithocondria and their sex determining chromosomes did not survive. Why?

An extremely old Y chromosome lineage

Mendez et al., (2013) [3] , discovered a Y chromosome haplogroup (named A00) which shares its most recent common ancestor with the rest of mankind 338 kya, of course (as usual!) the confidence interval is 237 to 581 kya. So this is a very old and wide time frame which goes back well beyond the oldest known fossils of anatomically modern human (AMH) beings.

This haplogroup was found in the genes of an African American and later identified in the chromosomes of Mbo individuals from Cameroon, at very low frequencies: 0.19% [CI = 0.11%–0.35%].

They would predate the oldest modern human fossils, which are only 195 ky old. So the Mbo lineage should have split from humanity at least 40 ky before our species even appeared.

The paper also looked at the rest of our Y chromosome lineages, working up the other branches of the tree, finding that the African A0 hg diverged from the other branches 202 kya (CI = 125 - 382 kya), which is much older than previous estimates of 142 kya.

This in turn moves the split between African and non-African branches (hg C to T) backwards in time, making them older: 63 kya vs. the previous 39 kya date.

Orthodoxy strikes back
Of course, such an ancient root for modern humans is inconsistent with mainstream genetic beliefs so a rebuttal appeared (Eran Elhaik et al., 2014)[4], pointing out that it: "contradicts all previous estimates in the literature and is over 100,000 years older than the earliest fossils of anatomically modern humans.".
E. Elhaik et al believe that Mendez et al.'s paper overestimated mutation rates resulting in an artificially older age. Their analyisis "indicates that the A00 lineage was derived from all the other lineages 208300 (95% CI= 163900 - 260200) years ago." [4] Which as can be seen, falls suitably close to the oldest dated AMH fossils.

New May 2018: Recent findings have pushed back both the earliest H. sapiens remains' datings in Africa and in Asia. This means that an older origin for AMH is indeed possible.

But let's get back to the very ancient A00 and the "unorthodox" point of view:

What are the implications of this finding? There are several possible scenarios:

1. Archaic admixture. These Mbo people inherited their A00 haplogroup from an achaic human population. These ancient African humans carried the A00 lineage in them, mated with AMH (within the last 195 ky) and passed on their Y chromosomes to their male offspring, their A00 passed from father to son in these AMH, which survived until now (Mbo's are their youngest descent). In the meantime, the archaic group became extinct.

Mendez et al. support this view: "Interestingly, the Mbo live less than 800 km away from a Nigerian site known as Iwo Eleru, where human skeletal remains with both archaic and modern features were found and dated to ~13 kya." [3]

2. Deeper time line. Maybe the depth of modern humans into the past is greater than the current (spotty) fossil record shows (roughly 200 kya).

The Confidence intervals given by Mendez et al., span from 237,000 to 581,000 years ago, this is well into the dark past of mankind, a time frame which encompasses many hominids in Eurasia and Africa: H. erectus (from 1.8 Mya until roughly 300 kya); Denisovans (until 30 kya), Homo heidelbergensis (600 to 400 kya), Neanderthals (600 to 25 kya) and H. rhodesiensis (300 to 125 kya).

Could this Y chromosome lineage belong to one of these hominids? Before answering let's just consider one additional piece of information.

Chuan-Chao Wang et al., [5] noticed that Y-chromosome STR haplotypes from different haplogroups converge; they looked for "possible haplotype similarities among haplogroups [and found] similar haplotypes between haplogroups B and I2, C1 and E1b1b1, C2 and E1b1a1, H1 and J, L and O3a2c1, O1a and N, O3a1c and O3a2b, and M1 and O3a2 ..." [5] .

In other words identical Short Tandem Repeat (STR) markers crop up among the different Y chromosome haplogroups indicating that the same mutations arise time and time again, unlike the markers of haplogroups which are much more sporadic and unique.

Making sense of this muddle

Above, I asked why didn't Neanderthal Y chromosomes survive among their descent (albeit admixed with modern humans)? Probably there were too few of them and too many of us, and their Y chromosomes got watered down, in successive admixtures, diluted until their frequencies are so low that they have not yet been detected.

Another option is that "Haldane's rule'' kicked in; the rule declares that if hybrids of one sex only are sterile, the afflicted sex is much more likely to be the male (XY) than the female (XX). Of course, why would hominds so close to each other bear sterile offspring? It is not like horses and donkeys, this is humans and Neanderthal, and we have their genes in us, so we did interbreed and the offspring survived (but Haldane's rule may mean that the girls survived but the boys were sterile).

A paper (Sankararaman et al., 2014) [6] supports the "Haldane's rule" notion: "...interbreeding of Neanderthals and modern humans introduced alleles onto the modern human genetic background that were not tolerated, which probably resulted in part from their contributing to male hybrid sterility".

The Neanderthal-deficient regions in modern humans are found in genes that "are specifically expressed in the testes, and in the female sex X chromosome. "This suggests that some Neanderthal-modern human hybrids had reduced fertility and in some cases were sterile. An unexpected finding is that regions with reduced Neanderthal ancestry are enriched in genes, implying selection to remove genetic material derived from Neanderthals. Genes that are more highly expressed in testes than in any other tissue are especially reduced in Neanderthal ancestry, and there is an approximately fivefold reduction of Neanderthal ancestry on the X chromosome." [6]

But, this does not mean a total extinction of their sex chromosomes; not all of their offspring were sterile after all, we still carry their genes mixed with ours, why shouldn't their Y chromosomes survive too?

In my opinion, we actuall carry their Y chromosomes in us (in the men of course), but long before we admixed with them in Eurasia some 60 kya. The mutation rates that are used to date our Y lineages are wrong, allow me to explain why:

We look at populations (say Amerindians) and jot down their haplogroup markers, and we assume that they mutated when they reached America and then, we guess the date they entered America (say 15 kya). With this we calibrate our clock. We again look at the humans closest to Africa and jot down their haplogroups markers, and once again guess the date these people's ancestors left Africa (say 70 kya), and again calibrate our clock. We take another look at the oldest fossils of AMH in Africa (195 kya) and jot down the most divergent African haplogroups' markers, we recalibrate our clock again. But, as you can see, there are many assumptions in all of this (the dates and, above all, the assumption that these haplogroups are specifically human and mutated recently < 200 ky!).

But, What if they are not specific to us, but archaic? What if we carry slowly mutating Y chromosomes. The mutations found in certain haplogroups are valid, but they reflect ancient migrations. Maybe even the Out Of Africa (OOA) migration of H. erectus or, within the time range given by Mendez et al., the Y chromosomes of H. heidelbergensis or Neanderthals?

This would explain why there are no Neanderthal specific branches to be found (the red ones in the image above). We all have the same tree our and their lineages coincide.

So what the date should we consider for the makers at the "non-African" CT groups (the OOA split)? Not the date modern humans left Africa. Instead it may be the date Neanderthals left Africa.

The split in Altai, with a Western route for our "R" hg and East for the "Q" hg may indicate divergence among the Neanderthals that lived there while AMH began to move out of Africa.

The archaic humans of China and even Lake Mungo people in Australia may be the branches of the South East Asian and Austronesian Y chromosome haplogroups, instead of modern humans that reached those regions much later.

The Q hg in America may not reflect a recent peoping of America at all, but an ancient one by Neanderthals.

The pan African E hg, may reflect the recent dispersal of AMH in the continent as well as in the Middle East, Southern Europe, and Asia after their OOA movement.

The dispersal of Q hg across the Arctic regions of America and Eurasia or the presence of the very old C hg in Asia and America may reflect the ancient migrations of primitive humans and not the recent (<20 kya) dispersal of modern humans.

See This post for an update. May 2018 and this Sept. 2020 update

Sources

[1] Musaddeque Ahmed and Ping Lian, (2013). Study of Modern Human Evolution via Comparative Analysis with the Neanderthal Genome. Genomics Inform. Dec 2013; 11(4): 230–238. Published online Dec 31, 2013. doi: 10.5808/GI.2013.11.4.230
[2] Johannes Krause, et al., (2007). The Derived FOXP2 Variant of Modern Humans Was Shared with Neandertals. Current Biology 17, 1–5, November 6, 2007 ª2007 Elsevier Ltd All rights reserved DOI 10.1016/j.cub.2007.10.008
[3] Fernando L. Mendez et. al., (2013). An African American Paternal Lineage Adds an Extremely Ancient Root to the Human Y Chromosome Phylogenetic Tree. The American Journal of Human Genetics 92, 454–459, March 7, 2013. http://dx.doi.org/10.1016/j.ajhg.2013.02.002.
[4] Eran Elhaik, T. Tatarinova, A. Klyosov and D. Graur, (2014). The 'extremely ancient' chromosome that isn't: a forensic bioinformatic investigation of Albert Perry's X-degenerate portion of the Y chromosome. European Journal of Human Genetics (2014) 1-6. doi:19.1038/ejhg.2013.303
[5] Chuan-Chao Wang et al., (2013). Convergence of Y chromosome STR haplotypes from different SNP haplogroups compromises accuracy of haplogroup prediction. pre-print Arxiv server on 21 October 2013.
[6] Sriram Sankararaman et al., (2014). The genomic landscape of Neanderthal ancestry in present-day humans. Nature 507, 354–357 (20 March 2014) doi:10.1038/nature12961

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2018 by Austin Whittall ©