Re-Examining the "Out of Africa" Theory and the Origin of Europeoids (Caucasoids) in Light of DNA Genealogy
Seven thousand five hundred fifty-six (7556) haplotypes of 46 subclades in 17 major haplogroups were considered in terms of their base (ancestral) haplotypes and timespans to their common ancestors, for the purposes of designing of time-balanced haplogroup tree. It was found that African haplogroup A (originated 132,000 ± 12,000 years before present) is very remote time-wise from all other haplogroups, which have a separate common ancestor, named β-haplogroup, and originated 64,000 ± 6000 ybp. It includes a family of Europeoid (Caucasoid) haplogroups from F through T that originated 58,000 ± 5000 ybp. A downstream common ancestor for haplogroup A and β-haplogroup, coined the α-haplogroup emerged 160,000 ± 12,000 ybp. A territorial origin of haplogroups α- and β-remains unknown; however, the most likely origin for each of them is a vast triangle stretched from Central Europe in the west through the Russian Plain to the east and to Levant to the south. Haplogroup B is descended from β-haplogroup (and not from haplogroup A, from which it is very distant, and separated by as much as 123,000 years of “lat- eral” mutational evolution) likely migrated to Africa after 46,000 ybp. The finding that the Europeoid haplogroups did not descend from “African” haplogroups A or B is supported by the fact that bearers of the Europeoid haplogroups, as well as all non-African haplogroups do not carry either SNPs M91, P97, M31, P82, M23, M114, P262, M32, M59, P289, P291, P102, M13, M171, M118 (haplogroup A and its subclades SNPs) or M60, M181, P90 (haplogroup B), as it was shown recently in “Walk through Y” FTDNA Project (the reference is incorporated therein) on several hundred people from various haplogroups.
A. A. KLYOSOV, I. L. ROZHANSKII
Copyright © 2012 SciRes.84
This study concerns the origin of anatomically modern hu-mans, which presumably belong to Y chromosomal haplogroupsA through T according to the classification developed in humangenetics and DNA phylogeny of man. This paper 1) sets forth atimeframe for the origin of Europeoids (Caucasoids); 2) identi-fies their position among all haplogroups (tribes) known todayon the haplogroup tree; and 3) offers evidence to re-examinethe validity of the "Out of Africa" concept.
The principal difference of our approach from those knownin human genetics is that our methodology is based on the iden-tification of branches of haplotypes in each haplogroup and its subclade (each branch is descended from its only common ancestor), and, in each case, is calculated a timespan from a com-mon ancestor of the branch by verifying that the branch is in-deed derived from one common ancestor and by using the crite-ria described in (Klyosov, 2009a; Rozhanskii & Klyosov, 2011;Rozhanskii, 2011). As a result, we obtained a chronology of allavailable branches in each haplogroup and in their total en-tirety—from A to T (in the current classification). In other words, for each haplotype we successfully identified its place inthe whole multi-haplogroup system of mankind. It is reasonableto assume that haplotypes of the whole of mankind form a con-tinuous system, albeit locally interrupted by "population bot-tlenecks" which essentially disrupt the initially continuous fab-ric of haplotypes. This fabric can be reconstructed based on itsfragments and in the same manner as the kinetics of chemicalreactions can be reconstructed based on relatively few experi-mental points. This analogy is rather close since mutations inhaplotypes obey the same laws of chemical kinetics, this wasdiscussed in the first paper of this series (Rozhanskii & Klyosov,2011).
Thanks largely in part to geneticists, the "Out of Africa"concept was popularized during the last two decades, yet it was never directly proven; however, for many specialists its appealwas undeniably convincing. The concept was based primarilyon the premise that Africa possesses the highest variability, or variance, of the human DNA and its segments. Set apart, it isnot a strong argument because a mix of different DNA lineagesalso results in a high variability and, as we show below, it islargely what occurs in Africa. Moreover, a genomic gap exists between some Africans and non-Africans, which has also beeninterpreted as an argument that the latter descended from Afri-cans. A more plausible interpretation might have been that bothcurrent Africans and non-Africans descended separately from amore ancient common ancestor, thus forming a proverbial fork.A region where this downstream common ancestor arose wouldnot necessarily be in Africa. In fact, it was never proven that helived in Africa.Research into this question has served as the basis for and thesubject of our work. We have found that a great diversity of Y chromosomal haplotypes in Africa is a result of the mixing of several very distant lineages, some of them not necessarilyAfrican, and that Europeiods (at least) do not contain "African"SNPs (those of haplogroups A or B). These important findings put a proverbial dent in the "Out of Africa" theory.
Results and Discussion
The 22 marker haplotypes, which are the "slowest" in termsof their mutation rate constant, described in (Klyosov, 2011a,2011b; Rozhanskii & Klyosov, 2011) were mainly used in this study. They are best suited for chronological calculations downto 100,000 years and deeper in time. This is because one muta-tion in these haplotypes occurs on average once in 4250 years,while in 67 marker haplotypes, for example, one mutation oc-curs—on average—once in 208 years (ibid). However, the 22marker haplotypes include a part of the last panel of the 67marker haplotypes and hence, 67 marker haplotypes wereneeded for the study.
Extended haplotypes of haplogroup A collected in variousdatabases (YSearch, FTDNA Projects), split into at least four different and distinctive DNA lineages, each with its base hap-lotype. This is essentially represented with a series of 32 marker haplotypes ( Figure 1) and 37 marker haplotypes ( Figure 2).In a simple, uncomplicated case, the base haplotype is equi-valent to the ancestral haplotype in the lineage. This is obvious in recent lineages, in which most haplotypes still represent the non-mutated ancestral haplotype. For more ancient lineages the
|A tree of 31 of 37 marker haplotypes of haplogroup A with somesubclades. Haplotypes were taken from YSearch FTDNA’s "African Project"(http://www.familytreedna.com/publicwebsite.aspx?vgroup=African.DNAProject§ion=yresults).|
12 11 11 - 9 11 - 10 - 10 9 12 12 7 10 8 null 13 11 16 1014 9 11 11
12 10 11 - 7 13 - 8 - 10 8 15 17 6 10 9 12 13 11 16 8 1311 11 12
13 11 12 - 10 11 - 16 - 10 9 14 14 8 8 8 9 12 11 12 8 1212 11 11
12 13 10 - 10 11 - 10 - 11 8 15 15 8 9 8 null 10 9 14 8 128 11 12
These base haplotypes have been assigned to subclades A3b2,A1a, A(P97+, SRY10831.1-), and A(M23-, M32-, P108-,SRY10831.1-), respectively. It was calculated that the common ancestors of the branches lived 5500, 5000, 600 years before present (ybp), and the last one is an individual haplotype. It is clear that these four haplotypes are tremendously distant from each other, and this is taking into account the extremely low mutation rates of the markers. The deduced base haplotype for
haplogroup A from the available data is as follows:12 11 11 - 9 11 - 10 - 10 8 14 15 7 10 8 12 13 11 16 8 139 11 12
Since alleles in the four haplotypes above vary significantly,the permutation method was applied (Klyosov, 2009a) for cal-culation of a timespan to their common ancestor. The averagesquared sum of all mutations in the four haplotypes aboveequals to 984, and the number of conditional generations to acommon ancestor is 984/22/2/16/.00027 = 5177. That is,132,000 years to a common ancestor of all the available haplo-types of haplogroup A. In the formula above, 22 is the number of markers in each haplotype; sixteen (16) is the square of thenumber of haplotypes in the dataset, 2 is introduced because thenumber of mutations was counted twice (all permutations),and .00027 is the mutation rate constant per marker in the 22marker haplotypes (Klyosov, 2011a, 2011b). A correction for
back mutations is not required in the permutation method(Klyosov, 2009a).Pairwise calculations of the dataset above give, as it should be, slightly lower values of timespans to a common ancestor.For example, base haplotypes of the A1a and A3b2 subclades (see Figure 2) differ by 25 mutations in all 22 markers, which places their common ancestor at 4167
8576 conditionalgenerations, that is 112,000 ybp. Two haplotypes with null-mutation differ by 27 mutations, which places their commonancestor at 4500
9922 conditional generations, that is127,000 years to a common ancestor (see Materials and Meth-ods for the principles of calculations). This gives an additionalsupport of the obtained "age" of haplogroup A as 132,000 ±20,000 years.
A similar approach was applied to haplogroup B, and thefollowing 22 marker base haplotype was obtained for a com-mon ancestor who lived 46,000 ybp (Klyosov, 2011b):11 12 11 - 11 11 - 10 - 11 8 16 16 8 10 8 12 10 11 15 8 1211 12 11It differs from the base 22 marker haplotype of haplogroup A by 18 mutations, which gives 18/.006 = 3000 generations. Witha correction factor (for back mutations) of 1.633, the resultconstitutes 123,000 years between common ancestors of hap-logroup A and B. Because they lived 132 and 46 thousandyears before present, respectively, their common ancestor livedapproximately 150,000 years before present (see Materials andMethods).The ISOGG (International Society of Genetic Genealogy)annual review in 2010 and earlier stated "
The BR haplogroupsplit off from haplogroup A 55,000 years before present ( bp)
.It probably appeared in North East Africa". Since the ISOGG-2012 stated, "The A haplogroup is thought to have been defined about 60,000 years bp
"http://www.isogg.org/tree/ISOGG_YDNATreeTrunk.html then haplogroups A and B should be separated by only several thousand years, or by 2 - 3 mutations in their slow 22 marker base haplotypes. It is not so, and between their base haplotypes there are as many as 18 mutations in the 22 markers, which translates to 123 thousand years (see above).This finding indicates that haplogroup B did not descend from haplogroup A. Rather, they both descended from a com-mon ancestor who lived ~150,000 ybp, and he was not necessarily living in Africa. Since he belonged to a haplogroup up-stream from haplogroups A and B, his haplogroup can be named "alpha-haplogroup". It is a matter of taste and belief to call it "Adam" or not.
Haplogroups C through T
The same methodology was applied to 7415 of 67 marker haplotypes of all known haplogroups and their subclades, re-duced to the slow 22 marker haplotypes, taken from databasesYSearch and SMGF and a multitude of FTDNA Projects (see Appendix). The base haplotypes of principal haplogroups andsome of their subclades, including those of haplogroups A andB, are listed below, and chronology of their appearance is shown in Figure 3
Pairwise calculations of mutation distances between the base haplotype of haplogroup A and each of the base haplotypes of other haplogroups place a common ancestor of the α-haplogroupat 160,000 ± 12,000 years before present. For example, 18 mu-tations between A and B base haplotypes, as it was describedabove, result in 150,000 ybp for their common ancestor. Twenty-one (21) mutations with haplogroup DE base haplotype give167,000 ybp for their common ancestor. Twenty-three (23)mutations with haplogroup H base haplotype result in 171,000ybp for their common ancestor with haplogroup A. For hap-logroup I (21 mutations) it is 161,000 ybp. For haplogroup Q(22 mutations) it is 166,000 ybp. For haplogroup R (21 muta-tions) it is 160,000 ybp. The distance in 19 mutations betweenthe base haplotypes of haplogroup A and β -haplogroup places α-haplogroup at 165,000 ybp. Clearly, the base haplotype of haplogroup A and its subclades is very remote from all other haplogroups.Similar pairwise calculations with the base haplotype of haplogroup B as well with all other haplogroups (besides A) place a common ancestor of beta-haplogroup to 64,000 ± 6000years before present (see Figure 3). This haplogroup is close toor identical with the BT haplogroup according to the currentclassification. Figure 3 shows a topology of the current haplogroup tree. The α-haplogroup which is the ancestral one with respect to both African (left branch) and non-African haplogroups (right branch) arose around 160,000 ybp, and 132,000 ybp gave riseto haplogroup A. Another, quite different branch, had formed afork, then apparently went through a population bottleneck around 70 - 60 thousand ybp (perhaps the Toba event), andgave rise to β -haplogroup, ancestral to non-African haplogroups, 64,000 ± 6000 ybp. Apparently, haplogroup B was initially not of an African origin. It could have migrated to Africa and mixed there with a local population. A common ancestor of the present-day bearers of haplogroup B lived 46,000 ybp. A similar story had occurred with a group of bearers of haplogroup R1b1around 4000 ybp, who ventured to the center of African continent during their westward migration along the African Mediterranean shore, and became a Negroid population having an unusual (for Africa) haplogroup (Cruciani et al., 2010; Klyosov,2012).The Mongoloid and Austronesian haplogroup C split ~36,000 ybp and gradually populated regions of Central Asia, Australia and Oceania. Haplogroup DE split to D and E around 42,000 ybp, and currently populates vast territory from North Africa to the west to Korea and Japan to the east.The family of haplogroup from F through T is largely theEuropeoid (Caucasoid) family. Most of bearers of these hap-logroups remained Europeoids; however, some populations have acquired racial features of the prevailing races in a given region, recently or in the long past.Based on the calculations given in this study, we know thatthe far most bearers of haplogroup A live in Africa, and they lived there probably all or most of those 132,000 years since haplogroup A arose. It cannot be excluded, of course, that hap-logroup A might have been appeared elsewhere and then mi-grated to Africa. However, there is no reason to believe (andfewer reasons to insist) that the Europeoid family originated in Africa.
Lack of the African SNP (Haplogroup A) in Non-Africans
A critical datapoint has emerged that disproves the "Out of Africa" concept; specifically, recent data shows that non-African people have neither M91, P97, M31, P82, M23, M114,P262, M32, M59, P289, P291, P102, M13, M171, M118 (haplogroup A and its subclades SNPs), nor M60, M181, P90 (hap-logroup B SNPs) in their Y-chromosomes.In fact, according to the data obtained from the "Walk Through the Y" (chromosome) international project conducted by Family Tree DNA (Texas and Arizona) [see Appendix] notone non-African participant out of more than 400 individuals inthe Project tested positive to any of thirteen "African" sub-clades of haplogroup A, SNPs for which indicated above. If totake, for example, bearers of R1a haplogroup, they each have the ladder of SNPs from M42 and M139 (haplogroup BT, butnot haplogroup B, which, as it was described above, split andmigrated to Africa around 46,000 ybp or earlier), through M168and M294 (haplogroup CT), P143 (haplogroup CF), M89 andP158 (haplogroup F), L15 and L16 (haplogroup IJK), M9(haplogroup K), М 74, L138, P69, P230, P243, P244, P280,P284, P286 (haplogroup P), М 207, P224, P227, P229, P232,P280, P285 (haplogroup R), P231, P241, P242, Р 245, Р 294haplogroup R1), L145 and L146 (haplogroup R1), L120 and
L122 (haplogroup R1a1), L168 (haplogroup R1a1a). In other words, all of the SNP have been identified, which should be found according to the phylogeny of R1a, but SNPs of the all examined subclades of haplogroup A were completely absent.The same pattern was observed with all other bearers of non-African haplogroups. The bearers of haplogroup A were exclusively positive to M91 SNP, characteristic of that haplogroup.
List of SNPs identified in haplogroup R1a1 and subclades of haplogroupA. Ancestral and derived alleles are shown. For blank spaces data arenot available. Cont. in Table 2.
SNP R1a1 A1 A1a A1b A2 A3b2V168 (alpha)G
AA G AV171 (alpha)C
GG C G GV221 (alpha)G
TT G G TP108 (alpha)C
TT C C C T T
There are, however, four distinct SNPs which present in both Africans and Europeans of haplogroup R1a1, taken the latter asan example. They seem to be the most ancient SNPs, which aredefined the alpha-haplogroup (see Figure 3 ). Tables 1 and 2 illustrate this statement.The ancestral alleles of the above four SNPs should corre-spond to the alpha haplogroup. All four are mutated in hap-logroup R1a1, and the WTY data show. All four are still ances-tral in the A1 subclade. All other subclades of haplogroup Ashow various combinations of the SNPs which do not matchthose in haplogroup R1a1 (see also Table 2 ).
logroup R1a1, it maintains their ancestral state. Table 2 shows SNPs of five subclades of “African” hap-logroup A. None of those SNPs have been observed in hap-These data, based on the SNPs (Single Nucleotide Polymorphism), along with the data based on the STRs (Short Tandem Repeats)
List of alleles of “African” SNPs identified in haplogroup R1a1 andsubclades of haplogroup A. Cont. from Table 1 .
SNP R1a1 A1 A1a A1b A2 A3b2M31 (A1a)G
C G C GV50 (A2)T
C T T CM32 (A3)T
C T T G T CP289 (A3b)C
G CM13 (A3b2)G
C G C
described in this study, are compatible with each other and undeniably indicate that non-African people, bearersof haplogroups from C to T, did not descend from the “African”haplogroups A or B. Their origin is likely not in Africa. Ahigher variance of the DNA in Africa, which was a cornerstoneof the “Out of Africa” theory, is explained by Figure 3 , in which haplogroup A has been evolving (mutation-wise) for 132,000 years, while the non-European haplogroups are muchyounger. Hence, there is a lower variability in the latter. Thesame is related to language variability, which has also beenused as an argument of the African origin of non-Africans. We believe that those arguments upon which the “Out of Africa”theory was based were, in fact, conjectural, incomplete and notactually data-driven. Therefore, we are left holding the question of the origin of Homo sapiens.Based on palaeoarchaeological evidence, the region, where anatomically modern humans have likely originated, is com- prised of a vast territory from Central Europe in the west to the Russian Plain in the east to Levant in the south. Each of these regions is renowned for discoveries of the oldest skeletal re-mains of modern humans dating back to 42,000 - 44,000 ybp. To date, none of these sub-regions has clear and unequivocal advances in this regard.
Materials and Methods
7556 haplotypes, predominantly 67 marker ones, have been collected in databases FTNDA, YSearch and SMGF (Sorenson Database), and reduced to the slow 22 marker haplotype panels:
DYS426, DYS388, DYS392, DYS455, DYS454, DYS438,DYS531, DYS578, DYS395S1a, DYS395S1b, DYS590, DYS641,DYS472, DYS425, DYS594, DYS436, DYS490, DYS450,DYS617, DYS568, DYS640, DYS492.
The methodology of haplotype datasets analysis was de-scribed in (Klyosov, 2009a, 2009b; Rozhanskii & Klyosov,2011). The most important research component involved dissecting the dataset to branches of haplotypes, each branch descended from one common ancestor. This was examined andverified by the logarithmic method (no mutation counting) cou- pled with the linear method (based on mutation counting), asdescribed in (Klyosov, 2009a; Rozhanskii, 2011). The mutationrate constant for the 22 marker haplotypes equals to .0060 mu-tation/haplotype/conditional generation of 25 years, or .00027mutation/marker/generation (Klyosov, 2011a; Rozhanskii &Klyosov, 2011). Haplotype trees were composed using softwarePHYLIP, Phylogeny Inference Package program (see Klyosov,2009a, 2009b and references therein). Corrections for back mutations were introduced as described in (Klyosov, 2009a;Rozhanskii & Klyosov, 2011). Margins of error were calculatedas described in (Klyosov, 2009a). Permutation method of TMRCA (time to the most recent common ancestor) calculationwas described in (Klyosov, 2009a).Base haplotypes in the dataset were determined by minimi-zation of mutations; by definition, the base haplotype is onewhich has the minimum collective number of mutations in thedataset. The base haplotype is the ancestral haplotype or theclosest approximation to the latter.
Example: Calculation of a timespan to a common ancestor for the base haplotype for haplogroup A (132,000 ybp)
12 11 11 - 9 11 - 10 - 10 8 14 15 7 10 8 12 13 11 16 8 139 11 12 ( A
)and that for the β -haplogroup (64,000 ybp)
11 12 11 - 11 11 - 10 - 11 8 15 16 8 10 8 12 10 12 12 8 1211 11 12 ( β -haplogroup )
19 mutations between the two base haplotypes results in19/.006 = 3167 conditional generations (25 years each) withouta correction for back mutations. The number of mutations per marker is 19/22 = .8636. By employing formula for back mutations (Klyosov, 2009a), we find that a correction for back mutation is
11exp.86362 = 1.686
Therefore, the “lateral” time difference between two base haplotypes is 3167 × 1.686 = 5340 conditional generations of 25 year, that is 133,500 years. A common ancestor of the both base haplotypes, A and β-haplotype, lived (133,500 + 64,000 +132,000)/2 = 164,750 ybp. Assignments of haplotypes to haplogroups and subclades were based on their SNP classification, as provided in the data- bases. In some instances it was additionally supported by calculating their position of the phylogenic trees from their respective STR data.
The authors are indebted to Dr. Alexander Zolotarev, a par-ticipant of the WTY project, for providing data of the Project,and to Ms. Laurie Sutherland for valuable help with the preparation of the manuscript.
Cruciani, F., Trombetta, B., Sellitto, D., Massaia, A., Destro-Bisol, G.,Watson, E., et al. (2010). Human Y chromosome haplogroup R-V88:A paternal genetic record of early mid Holocene trans-Saharan con-nections and the spread of Chadic languages.
European Journal of Human Genetics, 18,800-807.
Cruciani, F., Trombetta, B., Massaia, A., Destro-Bisol, G., Sellitto, D.,& Scozzari, R. (2011). A revised root for the human Y chromosomal phylogenetic tree: The origin of patrilineal diversity in Africa.
The American Journal of Human Genetics, 88, 1-5.doi:10.1016/j.ajhg.2011.05.002
Klyosov, A. A. (2009a). DNA Genealogy, mutation rates, and some Copyright © 2012 SciRes. 85
A. A. KLYOSOV, I. L. ROZHANSKII
historical evidences written in Y-chromosome. I. Basic principles and the method.
Journal of Genetic Genealogy, 5,186-216.
Klyosov, A. A. (2009b). DNA Genealogy, mutation rates, and some historical evidences written in Y-chromosome. II. Walking the map.
Journal of Genetic Genealogy, 5,217-256.
Klyosov, A. A. (2011a). The slowest 22 marker haplotype panel (out of the 67 marker panel) and their mutation rate constants employed for calculations timespans to the most ancient common ancestors.
Proceedings of the Russian Academy of DNA Genealogy, 4,1240-1257.
Klyosov, A. A. (2011b). DNA genealogy of major haplogroups of Ychromosome (Part 1).
Proceedings of the Russian Academy of DNAGenealogy, 4,1258-1283.
Klyosov, A. A. (2012). Ancient history of the Arbins, bearers of haplogroup R1b, from Central Asia to Europe, 16,000 to 1,500 years before present.
Advances in Anthropology,in press.
Rozhanskii, I. (2010). Evaluation оf the сonvergence оf sets in STR phylogeny and analysis оf the haplogroup R1a1 tree.
Proceedings of the Russian Academy of DNA Genealogy, 3, 1316-1324.
Rozhanskii, I. L., & Klyosov, A. A. (2011). Mutation rate constants inDNA genealogy (Y chromosome). Advances in Anthropology, 1,16-34.doi:10.4236/aa.2011.12005
Simms, T. M., Martinez, E., Herrera, K. J., Wright, M. R., Perez, O. A.,Hernandez, M. et al. (2011). Paternal lineages signal distinct geneticcontributions from British Loyalists and continental Africans among different Bahamian islands.
American Journal of Physical Anthropology, 146, 594-608.doi:10.1002/ajpa.21616
Reference data were selected according to SNP assignmentfrom YSearch database: (http://www.ysearch.org) and public projects of FTDNA
section=yresults Walk through the Y (International Project)
Scientific Research Publishing .