Mitochondrial Eve

Mitochondrial descendants of Eve

Mitochondrial Eve (mt-mrca) is the name given by researchers to the woman who is defined as the matrilineal most recent common ancestor (MRCA) for all currently living humans. Passed down from mother to offspring, all mitochondrial DNA (mtDNA) in every living person is derived from hers. Mitochondrial Eve is the female counterpart of Y-chromosomal Adam, the patrilineal most recent common ancestor, although they lived at different times.

Mitochondrial Eve is believed to have lived between 150,000 to 250,000 years BP, probably in East Africa, in the region of Tanzania and areas to the immediate south and west.^[1] She lived during a period of time when Homo sapiens were developing as a species separate from other hominid species. She lived in a population of between perhaps 4000 to 5000 females capable of producing offspring at any given time. Mitochondrial Eve would have been roughly contemporary with humans whose fossils have been found in Ethiopia near the Omo River and at Hertho. Mitochondrial Eve lived significantly earlier than the out of Africa migration which might have occurred some 60,000 to 95,000 years ago.^[2]

Matrilineal descent

For more information on the principles involved, see matrilineal descent and genetic genealogy (matrilineal).

To find the mitochondrial Eve of all living humans, one can start by tracing a line from every individual to his/her mother, then continue those lines from each of those mothers to their mothers and so on, effectively tracing a family tree backward in time based purely on mitochondrial lineages. Going back through time these mitochondrial lineages will converge on one maternal ancestor.

This female, nick-named 'Eve', was of interest because she bore a mitochondrial genome (mitogenome) which was the template for all later human mitogenomes. Each human mitogenome bears mutations that have replaced some nucleotides on the Eve's mitogenome with other nucleotides (or in some cases added or removed nucleotides). These are called single nucleotide polymorphisms (SNPs). Mitogenomes are passed from mother to offspring, with very few exceptions—only rarely will the father's mitochondria persist long enough or in sufficient quantity to segregate to the germ cells. Consequently, recombination between paternal and maternal mitogenomes is not observed as with autosomal and X-chromosomal DNA. All complex animals can also trace their ancestry back to a mitochondrial MRCA. Chimpanzees and humans share a mitochondrial MRCA; further back in time, humans and gorillas share an earlier mitochondrial MRCA.

Mitochondria within the cell have identical DNA sequence. On rare occasion, about once every 4,000 years a stable mutation occurs in a female that is passed to a female offspring and thus can be passed to subsequent generations.^{[3][note 1]} At the time of the 'Eve' a mutation occurred that created either the proto-L0 or proto-L1 lineage, and began the two basal branches of human mitochondrial DNA, L0 and L1. However, this it can be inferred that Mitochondrial Eve had at least two daughters who survived to have their own children.^[4] Because multiple mutations occurred on both lineages after mtDNA 'Eve' this new lineage cannot be determined.

Estimating time to most recent common ancestor

Estimated Coalescence time of human mitochondria MRCA (T_MRCA) is a measure of genetic distance to predicted sequence of the human mitochondria most recent common ancestor. T_MRCA is a product of coalescence analysis. There are two other major aspects: estimation of population size (and temporal structure) and estimating the place of coalescence, P_MRCA. Of these coalescence analyses products, T_MRCA is most important because population size estimates are heavily influenced by the T_MRCA estimates. Two recent studies of human mitogenomes have produced current best estimates of the T_MRCA of 194,300 and 192,000 years before present.^[2][5] These estimates follow more than 30 years of research and the estimates continue to evolve. The process is described below.

Coalescent ancestry of mtDNA

Humans are sexually reproducing organisms composed of two dimorphic sexes. Individuals within mammalian species cannot create exact duplicates of themselves. Instead, each individual passes ~1/2 of their genetic makeup to offspring with their mate contributing the other half. Through offspring production, individuals increase their genetic representation in the next generation, increasing the probability that more of their genes will be passed.

Mitochondria are overwhelmingly inherited from the female parent. Since female or male offspring are produced randomly at a ~1:1 ratios, when a mother passes a new mitochondrial mutation to her offspring there is random risk that the new mutation will be lost in the first generation. Alternatively the mutation may be passed to one or more female offspring and survive. If a mutation survives in a population long enough it may fix in that population (see figure:Most likely time in generations to fixation) as part of a forward looking process.^{[note 2][6]}

Most likely time in generations to fixation (N = 10)

Estimating TMRCAs, however, requires the interpretation of a past process. The accumulation of SNPs creates new lineages and thus creates genetic diversity, whereas lineages are pruned by genetic drift. As scientist measure diversity they can estimate when lineages might have formed. The mutations that occurred along Eve's descendant lineages prevented her mitogenome from fixing; however the SNPs that accumulated on each lineage allow scientist to estimate the time in which she lived. Population genetic diversity is measured by pairwise genetic comparison or parsimony analysis. Felsenstein (2002) showed that parsimony analysis was preferential to pairwise distance analysis because it arranges the sequences into a tree, and thus it is possible to measure the distance from any mtDNA sequence to a predicted sequence of the MRCA. With a neutral assumption, the observation of a MRCA sequence that joins the two deepest branches at a single point infers a MRCA time (TMRCA). In addition to determining diversity mutation rate needs to be determined, and this is frequently referenced with external calibration.

Calibrating the single nucleotide polymorphism rate

There are two general methods of determining single nucleotide polymorphism (SNP) rates. The most precise method involves establishing a rate based on observing mutations in pedigrees (observing mutation rates between a very large number of parent-child pairs). The most practical and most frequently used method implements a phylogenetics approach; observing SNP rates between two deep branches in a population or between two species, such as between humans and chimpanzees. This second approach requires, generally, some form of paleontological anchor.

In a recent review of mitochondrial clocking, Endicott et al. (2009) remarked that there is conspicuous disparity between rates estimated based on pedigrees and phylogenetics. Likewise, Soares et al. (2009) noted that the mutation rate is markedly higher than the observed SNP rate. One explanation for the disparity is that many deleterious mutations appear to be stable over a few generations but will be lost over large numbers of generations. Soares et al. (2009) suggest that these mutations can persist for 100s of thousands of years. Pedigree based rates would be more accurate for clocking recent migrations, and phylogenetic rates would be more accurate for estimate the time branch between two species when another similarly age branchpoint is already known. For estimating rates intermediate between short and long distance estimates Endicott et al. (2009) recommend using local or 'soft' calibration of the SNP rate (intra population based calibration) such as archaeologically defined colonization of different world regions.

Studies that rely on the chimpanzee-human last common ancestor (CHLCA) are also proportionally affected by variance of the predicted T_CHLCA from its actual age. Studies that rely on a pedigree based rate would not be capable of directly estimating events greater tha 50,000 accurately because substantial numbers of mutations would be lost due to purifying selection, and studies that rely on archaeological evidence may be troubled by classification issues and great paucity in the paleontological record. Therefore each method has one or more weaknesses.

Estimating based on AMH archaeology

Anatomical modern humans (AMH) spread out of africa and over very large intervals of geography and time left artifacts along the northern coast of the Southwest, South, Southast and East Asia Cann, Stoneking & Wilson (1987) did not rely on a predicted T_CHLCA to estimate SNP rates. Instead they used evidence of colonization in Southeast Asia and Oceania to estimate mutation rates. Endicott et al. (2009) have reevaluated the predicted migrations globally and compared those to the actual evidence. They postulate that the molecular clock based on chimp-human comparisons is not reliable, particularly in predicting recent migrations, such as founding migrations into Europe, Australia, and the Americans. They have offered the recommendation that mitochondrial clocks for local branches that predict migrations should be calibrated based on local evidence of human occupation, and less reliant on the CHLCA. Endicott et al. (2009) tend to be most critical of the CHLCA when used to estimate local haplogroup evolution with less critique of the timing of mitochondrial Eve. It should be noted however that Cann et al. (1987) estimated the TMRCA of humans to be approximately 210 ky and the most recent estimates Soares et al. 2009 (using 7 million year chimpanzee human mtDNA MRCA) differ by only 9%, which is relatively close considering the wide confidence range for both estimates and calls for more ancient T_CHLCA.

Using RFLP technology

This method has been used in a few but important early studies when sequence based methods were out of reach. Nucleotide sites in the mitogenome are variably functional. The optimum SNP rate selected depends largely on the sites selected for study. Based on this variation, there have been three general approaches that have been used with mitochondrial DNA to calculated the TMRCA: Restriction fragment length polymorphism (RFLP) analysis, hypervariable region (HVR; part of the D-loop that approximates the mitochondrial origin of replication, 1122 base-pairs) and coding sequence analysis. RFLP analysis has no direct assessment of rates, instead it internally calibrated the polymorphism rate based on archaeological evidence for 'founding' migrations to different continents.

Using this method, Cann, Stoneking & Wilson (1987) estimated the sequence divergence rate between humans of different geographic origin at 2 to 4% per million years. Within Africa the highest sequence divergence, approximately 1.5% for D-loop, 0.25% for tDNA, 0.30% for rDNA and 0.4% for coding. Later studies would predict African sequence divergence in the D-loop to be between 2 and 5% which is higher than the 1.5% observed for the D-loop (see below). Within the non-D-loop one expects divergence of 0.1 to 0.3%, 0.08 to 0.25%, 0.16 to 0.45% for rDNA, tDNA and coding region, respectively. The RFLP values fall within the best estimate of the expected range as of 2009. The lower values for D-loop are the expected consequence of saturation as described in Soares et al. (2009). The anchoring method used by Cann et al. (2009), is based on c.1987 understanding of archaeology, which places humans in East Asia by 40 kya. However it is currently known that anatomically modern humans reached Southwestern China well before 42,000 years ago (ka). In addition, the dates of the Mungo Lake remains have been reestimated to between 42 and 63 Ka consistent with other recent evidence for earlier occupation. There is evidence of human occupation in India from 76 Ka, and the arguably anatomically modern human remains at Jebel Qafzeh have been reestimated to 93 ka. The underestimate of D-loop divergence rate resulted in an overestimate of the TMRCA while the underestimate of the age of human migration from Africa resulted in an underestimate, such that the errors largely balanced each other.

Methods using time to CHLCA estimates

Calibration methods

Because chimps and humans share a matrilineal ancestor, establishing the geological age of that last ancestor allows the estimation of the mutation rate. However, fossils of the exact last common ancestor would be an extremely rare find. The chimp-human last common ancestor (CHLCA) is frequently cited as an anchor for mt-T_MRCA determination because chimpanzees are the species most genetically similar to humans. However, there are no known fossils that represent that CHLCA, in fact there are no proto-chimpanzee fossils or proto-gorilla fossils that have been clearly identified.^{[note 3]}

Some researchers tried to estimate the age of the CHLCA (T_CHLCA) using biopolymer structures which differ slight between closely related animals. Among these researchers, Allan C. Wilson and Vincent Sarich were pioneers in the development of the molecular clock for humans. Working on protein sequences they eventually determining that apes where closer to human than some paleontologist perceived based on the fossil record.^{[note 4]} Later Vincent Sarich concluded that the T_CHLCA was no greater than 8 million years in age, with a favored range between 4 and 6 million years before present.

This paradigmatic age has stuck with molecular anthropology until the late 1990's, when others began questioning the certainty of the assumption. Currently, the estimation of the T_CHLCA is less certain, and there is genetic as well as paleontological support for increasing T_CHLCA. A 13 million year T_CHLCA is one proposed age.^[8][9]

Despite having two known weaknesses,^[10] the preferred method for determining the SNP rate is the phylogenetic based method calibrated by inter-species measurements. One weak point, the estimated T_CHLCA is independent of the problems described above, and proportional variance of T_CHLCA causes proportional errors of the T_MRCA. The non-clocklike accumulation of SNPs, would tend to make more recent branches look older than they actually are. These two sources may balance each other or amplify each other depending on the direction of the T_CHLCA error. There are two major reasons why this method is widely employed. First the pedigree based rates are inappropriate for estimates for very long periods of time. Second, while the archaeology anchored rates represent the intermediate range, archaeological evidence for human colonization often occurs well after colonization. For example, colonization of Eurasia from west to east is believed to have occurred along the Indian Ocean. However, the oldest archaeological sites that also demonstrate anatomically modern humans (AMH) are in China and Australia, greater that 42,000 years in age. However the oldest Indian site with AMH remains is from 34,000 years, and another site with AMH compatible archaeology is in excess of 76,000 years in age.^[10] Therefore application of the anchor is a subjective interpretation of when humans were first present.

A simple measure the sequence divergence between humans and chimps by observing the SNPs. Given that the mitogenome is about 16553 base pairs in length (each base-pair which can be aligned with known references is called a site).^[11] The formula is:

The '2' in the denominator is derived from the 2 lineages, human and chimpanzee, that split from the CHLCA. Ideally it represents the accumulation of mutations on both lineages but in different positions (SNPs). As long as the number of SNP observed approximates the number of mutations this formula works well. However, at rapidly evolving sites mutations are obscured by saturation affects. Sorting positions within the mitogenome by rate and compensating for saturation are alternative approaches. ^[12]

Because the T_CHLCA is subject to change with more paleontological information, the equation described above allows the comparison of TMRCA from different studies.

Early, HVR, sequence based methods

To overcome the affects of saturation, HVR analysis relied on the transversional distance between humans and chimpanzees.^[13] A transition to transversion ratio was applied to this distance to estimate sequence divergence in the HVR between chimpanzees and humans, and divided by an assumed T_CHLCA of 4 to 6 million years.^[14] Based on 26.4 substitutions between chimpanzee and human and 15:1 ratio, the estimated 396 transitions over 610 base-pairs demonstrated sequence divergence of 69.2% (rate * T_CHLCA of 0.369), producing divergence rates of roughly 11.5% to 17.3% per million years.

HVR is exceptionally prone to saturation, leading to the underestimation of the SNP rate when comparing very distantly related lineages

Vigilant et al. (1991) also estimated the sequence divergence rate for the sites in the rapidly evolving HVR I and HVR II regions. As noted in the table above, the rate of evolution is so high that site saturation occurs in direct chimpanzee and human comparisons. Consequently this study used transversions, which evolve at a slower rate than the more common transition polymorphisms. Comparing chimp and human mitogenomes, they noted 26.4 transversions within the HVR regions, however they made no correction for saturation. As more HVR sequence was obtained following this study, it was noted that the dinucleotide site CRS:16181-16182 experienced numerous transversions in parsimony analysis, many of these were considered to be sequencing errors. However the sequencing of Feldhofer I Neanderthal revealed that there was also a transversion between humans and Neanderthals at this site.^[15] In addition, Soares et al. (2009) noted three sites in which recurrent transversions had occurred in human lineages, two of which are in HVR I, 16265 (12 occurrences) and 16318(8 occurrences).^{[note 5]} Therefore, 26.4 transversions was an underestimate of the likely number of transversion events. The year 1991 study also used a transition-to-transversion ratio from the study of old world monkeys of 15:1.^{[citation needed]} However, examination of chimp and gorilla HVR reveals a rate that is lower, and the examination of humans places the rate at 34:1.^[3] Therefore this study underestimated that level of sequence divergence between chimpanzee and human. The estimated sequence divergence 0.738/site (includes transversions) is significantly lower than the ~2.5 per site suggested by Soares et al. (2009). These two errors would result in an overestimate of the human mitochondrial TMRCA. However, they failed to detect the basal L0 lineage in the analysis and also failed to detect recurrent transitions in many lineages, which also underestimate the TMRCA. Also, Vigilant et al (1991) used a a more recent CHLCA anchor of 4 to 6 million years.

Coding region sequence based methods

Partial coding region sequence originally supplemented HVR studies because complete coding region sequence was uncommon. There were suspicions that the HVR studies had missed major branches based on some earlier RFLP and coding region studies. Ingman et al. (2000) was the first study to compare genomic sequences for coalescence analysis. Coding region sequence discriminated M and N haplogroups and L0 and L1 macrohaplogroups. Because the genomic DNA sequencing resolved the two deepest branches it improved some aspects estimating TMRCA over HVR sequence alone. Excluding the D-loop and using a 5-million-year T_CHLCA, Ingman et al. (2000) estimated the mutation rate to be 1.70 × 10⁻⁸ per site per year (rate * T_CHLCA = 0.085, 15,435 sites).

However, coding region DNA has come under question because coding sequences are either under purifying selection to maintain structure and function, or under regional selection to evolve new capacities.^[16][17] The problem with mutations in the coding region has been described as such: mutations occurring in the coding region that are not lethal to the mitochondria can persist but are negatively selective to the host; over a few generations these will persist, but over thousands of generations these slowly are pruned from the population, leaving SNPs.^[3] However, over thousands of generations regionally selective mutations may not be discriminated from these transient coding region mutations. The problem with rare mutations in the human mitogenomes is significant enough to prompt a half-dozen recent studies on the matter.

Ingman et al. (2000) estimated the non-D loop region evolution 1.7 × 10⁻⁸ per year per site based on 53 non-identical genomic sequence overrepresenting Africa in a global sample. Despite this over-representation, the resolution of the L0 subbranches was lacking and one other deep L1 branches has been found. Despite these limitations that sampling was adequate for the hallmark study. Today, L0 is restricted to African populations, whereas L1 is the ancestral haplogroup of all non-Africans, as well as most Africans. Mitochondrial Eve's sequence can be approximated by comparing a sequence from L0 with a sequence from L1. By reconciling the mutations in L0 and L1. The mtDNA sequences of contemporary human populations will generally differ from Mitochondrial Eve's sequence by about 50 mutations.^[2][18] Mutation rates were not classified according to site (other than excluding the HVR reigons). The T_CHLCA used in the year 2000 study of 5 Ma was also lower than values used in the most recent studies.

Inter-comparing rates and studies

Molecular clocking of mitochondrial DNA has been criticized because of its inconsistent molecular clock.^[19][20][21] A retrospective analysis of any pioneering process will reveal inadequacies. With mitochondrial the inadequacies are the argument from ignorance of rate variation and overconfidence concerning the T_CHLCA of 5 Ma. Lack of historical perspective might explain the second issue, the problem of rate variation is something that could only be resolved by the massive study of mitochondria that followed. The number of HVR sequences that have accumulated from 1987 to 2000 increased by magnitudes. Soares et al. (2009) used 2196 mitogenomic sequences and uncovered 10,683 substitution events within these sequences. Eleven of 16560 sites in the mitogenome produced greater than 11% of all the substitutions with statistically significant rate variation within the 11 sites.^{[note 6]} They argue that there is a neutral-site mutation rate which is a magnitude slower than rate observed for the fastest site, CRS 16519. Consequently, purifying selection aside, the rate of mutation itself varies between sites, with a few sites much more likely to undergo new mutations relative to others.^[22] Soares et al. (2009) noted two spans of DNA, CRS 2651-2700 and 3028-3082, that had no SNPs within the 2196 mitogenomic sequences.

The estimated time to mitochondrial Eve

Early studies

Allan Wilson and his colleagues began examining the mitochondrial molecular clock in the late 1970s and they found some regions of mitochondrial DNA evolve rapidly. Given sequencing technology of the time this was useful because many discrepant SNP could be detected over a short sequence of DNA.^[23][24] In 1980, W.M. Brown, looking at the relative variation between human and other species, recognizes there was a constriction in the human population 180,000 years ago.^[25] A year later Brown and Wilson were looking at RFLP fragments and determined the human population expanded more recently than other ape populations and noted that humans had the mtDNA diversity that was comparable to isolated subspecies of other apes.^[26] The study described above by Cann, Stoneking & Wilson (1987) estimated the time in which mitochondrial Eve lived (human mitochondrial TMRCA) at 215 +/- 75 kya (142,500 and 285,000 years ago).

Sequence based studies

This was followed by Linda Vigilant's approach applifying the hypervariable region within the mitochondrial D-loop from the single hairs of southern African hunter-gatherers (!kung-San - a click speaking tribe of Namibia and neighboring Botswana).^[27] At the time, it was believed that sequencing this region was advantageous because the larger density of mutations and because it was believed hypervariable region neutrality caused rapid SNP rate. Two years later, Vigilant et al. (1991) used the same technique and 4 to 6 million year T_CHLCA range to produced human mtDNA TMRCA between 166,000 and 249,000 years. As described above the approach had a number of problems, indicating the need for a much larger TMRCA confidence interval in the study.

Advances in sequencing made it possible to sequence large numbers of genomic mitochondrial DNA (mitogenome). In 2000, Ingman et al. (2000) analyzing the non-HVR region of mitogenomes estimated mitogenomic TMRCA of 171,500 ± 50,000 years. This estimate was lower than previous studies, as this group continued to use a recent T_CHLCA of 5 million years. However, the study did resolved the deepest branching of mitochondrial population in humans.

Despite some agreement with this date in some anthropological circles, there was concern that this date was too recent. A growing body of evidence from the Levant (Skhul and Qafzeh), India, China and Australia (Mungo Lake- LM3) that humans had migrated from Africa well before 52 kya.^[28] Higher-set T_CHLCA places the upper limit of confidence above the age of the earliest non-Neanderthal hominids at Skhul. However, Tattersall and Schwartz (2008) recognize that some examples of late archaic homo sapiens in the Levant may be better placed in other (non-Homo [sapiens] sapiens) taxa as the Levant may not have been an early site of human occupation out of Africa.^[29] Rightmire (2009) associates archaic humans from Jebel Irhoud (Morroco 160,000) in a Mousterian tool context with the early Skhul fossils and if this dating is correct (real date not less than the estimate) then it distances both Jebel Irhoud and the oldest Skhul fossils from the geographic limits of the constrict population. Because of the sample size this study failed to see evidence of selection or population size growth; however, coalescence theory predicts that under neutral models, current population size in Africa is far too great to explain coalescence as recent as 171,500 years ago without some selection.^[30]

The region in Africa where Tishkoff found the greatest level of mitochondrial diversity (green) and the region Behar et al. postulated the most ancient division in the human population began to occur (light brown)

Gonder et al. (2007) undertook mitogenomic sequencing in areas of Africa were previous studies indicated deep diversity. This new study found new lineages of African mtDNA and more importantly narrowed the region within Africa in which humans ancestors likely arose. This new study indicated that the TMRCA likely occurred between 160,000 to 226,000 years ago (but dates between 130,000 and 280,000 cannot be ruled out, see TMRCA table).^[2] This study was followed by Soares et al. (2009) which estimated the TMRCA at 192,000 years by singling out sites that were not as subject to purifying selection in the mitogenome.

Estimated times of major mtDNA branchpoints

The deepest branching lineage within the human mitochondrial population is the L0/L1 branches uncovered by Ingman et al. (2000). Beyond this, the L1 subbranches had largely been described by in the study of HVR regions in the decade previous to that study. The L0 subbranches have undergone intense study in the since 2000. Behar et al. (2008) examined the Khoisan population adding many more sequences. They determined that Khoisan mitogenomes other than the L0d and L0k appear to be the result of recent admixture. Consequently they estimated that Khoisans separated from the core interbreeding population after both the L0d and L0k clades had formed, about 144,000 years ago +/-11,000 years. In as much as their evidence suggested very low geneflow between non-Khoisan females and Khoisan females for 10,000s of years it was no longer possible for a population constriction to fix mtDNA in both groups, the period of constrained population size had effectively ended. Behar et al. (2008) derived their TMRCA from a 6.5 Ma T_CHLCA which had a mutation frequency of 1 per 5182 years. This is about 80% the rate proposed by Soares but Soares used at larger T_CHLCA therefore using Soares et al. methodology the branching of L0k would be about 125 ka.

Cann, Stoneking & Wilson (1987) used paleoanthropological evidence for human settlement in New Guinea, Australia and the New World allowing them to estimate the sequence divergence rate was 2 to 4% per million years(at the RFLP level). An ancestor "c" contained no known African ancestors and they suggest this ancestor lived between 90,000 and 180,000 years ago. Ingman et al. (2000) presented with an 'exodus' time from Africa in non-Africans of 52,000 years +/- 27,500 years (Assuming T_CHLCA = 5 million years).^[11]

Endicott et al. (2009) have recently reviewed the evidence for mutation rate variation and consider that the level of rate variation in humans, between lineages, is considerable. They have cast considerable critique on the use of global molecular clocks, but have particularly criticized the use of general molecular clocking on the timing of regional migrations. Therefore while considering that the TMRCA for mitochondrial Eve has tended to float around an estimated age of 200 Ka, more caution should be applied when considering the precise timing of migrations based on the MRCA of haplogroups, such as haplogroup M and N.

Coalescent structure

Population size estimates are one of the most important products of TMRCA determination. Based on the TMRCA and branching structure of the parsimony tree over time and geographic space these secondary products can be estimated:

• The average population size throughout the from the TMRCA to present.
• The complex structure of the population can be estimated.
• Place of the MRCA (PMRCA).

If population size is sufficiently large we might argue that Z was inclusive of all the similar morpho-metrically defined populations (i.e. fossil cousins) that existed at the time, but if Z is sufficiently small, one begin to look at the concept of speciation, the formation of a new species. In other words, the more that the boundaries of population size can be limited the more one is able to draw inferences about the population (its range, its location, its constituents).

Summarizing, the rate at which a new allele will eventually will displace all deeper branching clades and the time that it takes is a probabilistic function of the population structure and ploidy based on forward looking statistics. Above, the complicities of calculating the TMRCA are discussed and the result of the process ls a broad range. Since populations size estimates are dependent on that range it will also occupy a large range.

where Tg = generation time.

As with the prediction of the TMRCA, variance of population size is a concern. If we assume a flat population structure then population size equals half the TMRCA in generations. Having a range of TMRCA then one need only determine the intergenerational distance (generation time) and the flatness of the structure. However depending on the branch (bushy versus stretched out). If the structure is fully stretched out with branches following a slowly increasing number over time, then one can predict the variance of population size similar to the method to establish variance about TMRCA estimates.

Based on the early mtDNA studies and other early indirect studies Takahata estimated that the N_{e, female} was between 3,500 and 4,600 individuals.^[31]

Estimation of generation time

Early studies using mtDNA used generation time of 20 years based on numbers of studies of Neolithic peoples. However, it appears that population growth and higher densities of population after the Mesolithic has favored lower generation times. Studies from Europe's Mesolithic/Neolithic transition indicated that Mesolithic hunters typically were healthier and lived longer than their Neolithic counterparts. In addition, demographic studies of African hunter-gatherers revealed that peoples like the !kung have relatively long generation times, and as this group represents a major early branch within the human population then we must factor their reproductive behavior into estimation of generation times.

Estimates of generation time now lie between 20 and 25 years, there appears to be a preference for 22 or 23 years.

Estimating Eve's population size

Estimates assuming a flat population structure

If the population structure is assumed to be flat then the range of median estimated values has to be between 2200 and 9000 individuals, simply based on uncertainty about the mutation rate and generation times. The distribution of population sizes like TMRCA is an exponential function, and thus variance is also an exponential function, and with no other information one would create the crossproduct between the variance and TMRCAs to determine the appropriate population size distribution.

Resolving more complex structure

Retrograde look at bottlenecks. Diagram shows what genetic coalescence detects (grey area) in retrospective examination of the population, the alleles of more ancient TMRCA (circled in white) are not visible in the extant population, because they have been excluded prior to population expansion

Population structure is heavily influenced by branching in several ways. An informative example is haplogroup M, after evolving in or entering Eurasia, the population expanded multifold.^[32] Thus the number of M bearing lineages propagated, if this were not the case we would see several mutations between M and its 2 basal sublineages, but there are a dozen or so basal M lineages spread between Africa, South and East Asia. Another model that explains rapid growth of lineages would follow positive selection. Atkinson, Gray & Drummond (2009) argue that there were significant increases in the population size in mulitple lineages (L0, L1, L2 and especially L3) prior to 60 kya. Therefore many lineages could not be eliminated beyond 100,000 years ago with the population sizes that have existed in Africa for the last 100,000 years, and these lineages did not replace other lineages but grew in number as evidence for human expansion outward from Southeastern Africa. While these authors did not extrapolate backwards of 150,000 years (anchor = Ht Q in Papua New Guinea at 45 kya; with generation time of 20 years) and places the sum of lineage specific population sizes above 10,000 individuals before 100,000 years ago. They also demostrate a 95% confidence range before 150,000 years ago is between 1000 and 10,000 effective females. Because of the coalsence of lineages backward through time, population size cannot be extrapolated fertility backwards.

Expansions are not required to prevent fixation events, isolation of groups can inhibit fixation. Within the Khoisan speakers of S. Africa, maternal lineages appear to have undergone 100,000 years of isolation.^[18] Consequently, given that the World's population is magnitudes larger than the estimated size and that an inflection of size had to have occurred at some point, a good choice of points in which fixation became decreasingly possible lies between a temporal depth of 1/3 to 1/2 the retrospective time to the TMRCA. This means that the flatness of the population structure maybe restricted to Africa and at depths greater than 1/3 TMRCA. This effectively limits the maximum 'flat' shaped population during that period to less than 6000 effectively reproducing females (census of 18000 females).

Solitary female Eve as a misconception

Therefore increased knowledge of the branching structure better affords the estimation of population size. It should be noted that there is a model by which population could grow from very small numbers, a virtual 'Eve' if the population continues to expand at rate tolerant of preservation of the deepest branch. Using mtDNA alone cannot discriminate the 'early flat' model and a virtual 'Eve' model, however studies of other loci also point to a population size of about 10,000 effectively reproducing individuals, and extreme recent pinches in the population would result in abundant recent fixation of autosomal and X-linked loci. In contrast 90% of all X-link loci studies so far underwent fixation 100,000s of years before mtDNA TMRCA suggesting that population did not collapse to a single mating pair, and establishes a lower limit of the population size approximately 10² females, even for brief periods.^[33] The concept of a solitary Eve within the human population is a misconception.^{[34][35][note 7]}

Selection on mitochondrial DNA sequence

One of the most common measures implying structural changes are measures of selection. Selection in this context is not necessarily the same as positive selection in a competitive context. Fortuitous colonization of new territories by a small number of individuals can make unique alleles carried by those individuals appear selective to other individuals. If we apply this to humans, by creating a given in the argument that there was a place in Africa that was core to mtDNA of all humans and that there was at the geographic center of that core unique set of alleles. As the population expanded from the boundaries after the MRCA sequence had excluded all deeper branches then all the branches that expanded on the fringe would exhibit the appearance of positive selection relative to the allele that was unique to the central population. There are two measures of departure from Neutrality, the Tajima's D statistic and the D* and F* statistics of Fu and Li. As more mitogenomic sequence information has been gathered on Africa's hunter gatherers it appears there has been selection acting on Africa's basal lineages and that selection is not uniform.

Implications

The background for the interpretation of the mtDNA TMRCA are the numerous studies showing human-like fossils and bone remnants in Africa and Eurasia. These studies progress in concert with modern molecular genetics, with sequencing of ancient bone remnants becoming more common-place, particularly for Neandertals. Prior to the sequencing of the first Neandertal mtDNA, the TMRCA and PMRCA for modern humans set ancient humans apart from Neandertals. The current sequence of a number of Neanderthal mtDNA has reaffirmed that in the genus homo's recent past the female lineages of Neandertals and Humans existed in non-crossbreeding populations. Thus the implication of recent African origins 20 years ago set off one of the biggest battles in paleoanthropology, setting major players, such as Milford Wolpoff against molecular paleontologist such as AC Wilson.

Mitochondrial 'Eve' appeared in the literature in 1987 supporting an 'Recent out of Africa' model that was a minority view at the time. The Multiregional evolution hypothesis (MREH) was the most popular at the time and continued to challenge the OoA model for more than a decade.

Despite the compatibility of TMRCAs/PMRCAs from Neandertals and Humans with a multiple species model, other genetic studies ambiguously supported this theory. Y chromosomal studies produced a TMRCA, at less than 50,000 years, this estimate occurred after humans had expanded. Studies of other loci presented with a wide variety of fixation times, from 0 to 2 million years. Others studies did not adequately sample in Africa and presented ambiguous PMRCA or PMRCA on other continents.^[36] Francisco Ayala tried to discredit the coalescent population size by arguing that the HLA-DRB1 locus had a TMRCA of 10s of millions of years.^[37]

Variance of neutral fixation times

TMRCAs of loci, Y chromosome, and mitogenomes compared to their probability distributions if one assumes that population expanded 75kya from a long-standing population of 11,000 effective individuals

One early oversight of many early studies is that the fixation of alleles (the object of coalescent theory study) is not a discrete mathematical function, it is a probabilistic function, and it is highly dependent on the ploidy being studied.

Comparison to the X-linked and Autosomal TMRCAs

Takahata (1999) was the first molecular anthropologist to point out that conclusions drawn from single locus studies suffer from the large randomness of the fixation process. Schaffner (2004) has cleared up this issue by demonstrating the 3 sets of fixation ranges, haploid, X-linked and diploid where TMCRAs for different loci are expected to fall. Takahata (1993) estimated the effective human population size at 11,000 individuals, and Schaffner working on an improved set of X-linked markers from low recombination regions of the X-chromosome identified an effective size of approximately 12,000 individuals.^[38][39] PDHA1 falls on the edge of fixation times for X-linked chromosome. For autosomes, the MX1 locus and the HLA loci appear to preserve past diversity in the human population. With few exceptions, however, X-linked and autosomes appear to coalesce under a common population size.

Comparison to the Y chromosomal TMRCA

Just as mitochondria are inherited matrilineally, Y-chromosomes are inherited patrilineally.^[40] Y chromosomal TMRCA, the time of the Y-chromosomal Adam, lie in the 42 to 110ky range, which is a little less than half the TMRCA of mtDNA. Importantly, the genetic evidence suggests that the most recent patriarch of all humanity is much more recent than the most recent matriarch, suggesting that 'Adam' and 'Eve' were not alive at the same time. While 'Eve' is believed to have lived more than 140,000 years ago, 'Adam' appears to have lived less than 110,000 years ago.^[34] According to Wilder et al. (2004), the lower TMRCA of Y is due to an effective population size of males 1/2 that of females over most of human evolution.^[41]

Even with a reduced effective population size there are problems with this explanation . Recently, with more mitogenomic sequences from Africa, evidence has grown for an early population size expansion. This expansion probably started prior to 100,000 years ago and greatly increasing after 100,000 years ago(see: Population size oscillation). The effective size of the human population should have well exceeded 10⁴ individuals between 80,000 to 120,000 years ago. Given this expansion, implicit male populations sizes would have improbably coalesced to Y-Adam within that time frame. However, the greatest age for Y TMRCA is more recent than the evidence for expansion. In addition, despite evidence of a bottleneck, the human mtDNA TMRCA range remains consistent with population sizes estimates from X-linked and autosomal loci. However, Y-chromosomes TMRCA is not consistent with mtDNA or either of these sets (see figure:TMRCAs of loci).

This inconsistency maybe explained by some form of Y chromosome selection (cultural, or genetic). A Y-chromosomal lineage might have swept the male population.^[42] However, if true the place of greatest Y chromosomal diversity could be anywhere that humans inhabited Africa. However, Y diversity is greatest in Southern Africa, close to the earliest female population split predicted by Behar et al. (2009) suggesting the earliest branch in Y should be between 125,000 and 150,000 Ka in age. This suggests a SNP rate inaccuracy in the Y-chromosomal and/or mtDNA molecular clock. A recent study of X-chromosome suggests that different rates of male sperm production between humans and chimps has altered the molecular clock in sex chromosomes.^[43] This shift in the molecular clock would not affect the mtDNA SNP rate and would affect the Y-chromosomal rate more than X-linked and autosomes, since these Y-chromosomal lineages spend the most time in male testes.

Population size oscillation

The term bottleneck has been used to describe the population structure that created mtDNA Eve. The appearance of a bottleneck was a consequence of the appearance of a 'big bang' of HVR branching about the time humans first left Africa. From that point back to the TMRCA was less than 100,000 years and the population size estimate was below 5000 effective females. Looking backwards in time this is what might be called a retrograde bottleneck, however it is an artifact of coalescence process, since the coalescence of mitogenomes on the sequence of the MRCA (the event which initiated with mtDNA Eve and extended to the extant population) conceals the population size from all points earlier than that mutation (see figure Retrograde look at bottlenecks). Therefore the population size could have been of equal size going back 100,000s of years, to the time in which Neanderthals' ancestors and Modern humans' ancestors were part of a single population.

Evidence against a population bottleneck

The work done on Neanderthal sequencing (Green 2007) has identified little evidence of Neanderthal contribution to humans, moreover it describes an effective size of the population when humans and Neanderthals split was about 3000 individuals. Taken in the light of Schaffner's and Takahata's effective populations sizes, 3000 < N_{e, female} < 6000 and 2000 < N_{e, male} < 4000 does not appear to represent a magnitude shift downward from the average size. Taking a null hypothesis, prior to and after the mtDNA MRCA population sizes appear to reflect long-term small population structure up until 70,000~150,000 years ago, not a brief constricting bottleneck, but a long period of constrained size followed by an expansion.

Evidence for a population bottleneck

Confidence intervals of population size do not require an alternative, population bottleneck, hypothesis. However, a bottleneck may have existed. If the population size were at 12,000 individuals as suggested by X-chromosomal studies, the N_e for mtDNA and Y in particular, is below the expected median TMRCAs (See image Above and on the left). Y chromosome and mtDNA may be more representative of population structure immediately prior to expansion. However, meshing mtDNA TMRCA and Y TMRCA is problematic. If these two loci could be treated together, they would likely fall significantly below the X-linked and autosome-derived size estimates for any given T_CHLCA.

Most probable number of effective females based on TMRCA, showing the best estimate, and how Takahata's and Shaffners estimates compare (after conversion of Ne to Ne females

Atkinson, Gray & Drummond (2009) show that prior to 150,000 years ago the population could have been as low as 1000 effective females (~1500 total, 4500 census) with a lower population size between 150,000 to 200,000 years ago. Whereas X-chromosome and autosomes warrant larger population size minima, 1000s of females, these loci of larger ploidy are capable of sensing population structure of much longer periods. Such periods may include recent and ancient population structures and size oscillations. Most population structure models for Africa have assumed much of the growth occurred very recently, however Atkinson et al. (2009) shows that by 100,000 years ago the minimum female population size exceed the estimated population size for females. The flat population/recent growth model is troubled in considering an ancient population core ine Tanzania (Gonder. et al. (2007) early East African/Khoisan split (Behar et al. 2008), and spread of L2 in parts of Africa where L0 and L1 are found in low abundance. Simply, the evidence of lineage growth appears to correlate with growth in geographic regions in which humans live. Retrospectively, this suggests that population size was growing as new lineages appears to expand territory. Comparing these observations with populations sizes suggested by X-chromosome (~7000 females) one might expect a low stand of the human population size of 1/3 to 1/2 this size between 150,000 to 250,000 years ago. This indicates that earlier periods had a reciprocal, or larger size (>7000 females) between 200,000 and 500,000 years ago.

Other authors such as Endicott et al. (2009) think that bottlenecks in the human prehistory were such a common feature that they intefere with TMRCA determinations, and implies the possible effect of the OIS-6 on population size reduction with a TMRCA around the time of late pliestocene climate optimum, approximately 120,000 years ago.

Geographic constraints of ancient humans

The principal philosophical battle between strict Out of Africa and Multiregional Evolution hypotheses revolves around the placement of humans at times during human evolution. The singly interpretable result of mtDNA coalescence clearly favored a more recent African origins.

The strict Out of Africa model places humans in sub-Saharan African or even more strictly, 1,000 to 20,000 interbreeding individuals. Consistent with the geographic distribution of phylogenetic reconstructions these individuals lived in a defined domain in sub-Saharan Africa for 1000s of generations between 100,000 and 250,000 years ago. Because all human females are considered to be within this time-space domain since other hominids (Neanderthals, Flores hobbit, Java man, Peking man, a set that also might include Jebel Irhoud) continue to evolve elsewhere. Which therefore indicates some fertility barrier existed between humans and these other types of humans, warranting the delineation of multiple human species. As part of this model specific events happened, population remains in a contained region, and at some point expands, expands out of Africa, expands to most parts of Africa, and since other hominids are no longer apparent, probably took part in a global-'competitive' displacement.

In contrast, the strict multiregional hypothesis argues that humans were spread broadly across Africa, Eurasia, parts of Oceania. Humans interbreed overtime across long ranges, but modern humans primarily represent evolution in-situ. As part of this model the ice age ends allowing people to travel, agriculture increase population size, but people have largely evolved in place. Massive migrations and displacement would not explain genetic makeup.

If we assume that there were humans in NE China, central China, southern China, Sumatra, Flores Island, Europe, the Levant, North Africa and Subsaharan Africa, and assuming there were no other humans other (e.g. Narmada India) then effective sizes per known regions was about 10³ effective females (The more recent studies would suggest 100s of individuals per region between 150,000 and 200,000 years ago). In an Out of Tanzania model, 3000 individuals would be spread at densities of one per 300 square kilometers, at the size of sub-Saharan Africa under the OoA model this would mean that the density falls to one per 1000s of square kilometers and under the MREH model it would mean one per 10000s square kilometers. Based on relic stone age populations from around the world the recent African origin hypotheses better fit the estimated population sizes. As a consequence of increase information in support of OoA and the near impossibility of a strict MREH model many have repositioned their stance favoring a mostly Out of Africa hypothesis, in which modern humans recently migrated from Africa, but on rare occasion intermixed with regional humans (humans paraphyletic to Homo sapiens sapiens).

Mitochondrial MRCA and the MRCA of all humans

Mitochondrial Eve is the most recent common matrilineal ancestor, not the MRCA. Since the mtDNA are inherited maternally and recombination is either rare or absent, it is relatively easy to track the ancestry of the lineages back to a MRCA, however this MRCA is only valid when discussing mitochondrial DNA. Ironically mtDNA are not human, they are organelles that live within our cells, so it is better to say these are human-mitochondrial Most Recent Common Ancestor. Despite the recent fixation of the mtDNA genome in humans, other genes have evolved that were broadly selective in the human population, these genes have swept through the human population, two such genes have been identified on the X-chromosome. Other studies have indicated the overwhelming majority of humans have a recent common ancestor within the last 5000 years (albeit between any two individuals it may not be the same ancestor)^[44], however the genetic relationship between well diverged individuals may not reflect the theoretical relationship, as geographic and cultural barriers may slow gene migration. Gene migration is not fluid in humans, as genes are passed in units called chromosomes, which undergo limited number of recombination on each unit per generation, therefore a common ancestor genealogically may not indicate the passage of DNA from that ancestor to the two divergent individuals. Whereas since mtDNA does not undergo this dilution via recombination, we can argue that the majority of mtDNA sequence (that which has not undergone mutation) from mtDNA ~16000 nts came from a single individual >150,000 years ago. A more recent common ancestor for all Males is the much larger Y chromosome (however it codes for very few genes).

In popular science

Cover of the January 11, 1988 edition of Newsweek

Newsweek Magazine reported on Mitochondrial Eve based on the Cann et al. study in January 1988, under a heading of "Scientists Explore a Controversial Theory About Man's Origins". The edition sold a record number of copies.^[45]

Bryan Sykes has written a popular science book entitled The Seven Daughters of Eve (2001, ISBN 0-393-02018-5) that presents the theory of human mitochondrial genetics to a general audience.

In River Out of Eden, Richard Dawkins discusses human ancestry in the context of a river of genes and shows that Mitochondrial Eve is one of the many common ancestors we can trace back to via different gene pathways.

The Discovery Channel produced a documentary entitled The Real Eve (or Where We Came From in the United Kingdom), based on the book Out of Eden by Stephen Oppenheimer.

In popular culture

• The main villainess of the Parasite Eve series is named Mitochondria Eve.
• In the "reimagined" Battlestar Galactica television series, Hera Agathon (the Cylon-human hybrid) is the Mitochondrial Eve as revealed in the series finale.^[46]
• Mitochondrial Eve is a science fiction story by Greg Egan that revolves around the concept of Mitochondrial Eve and Y-chromosomal Adam.^[47]

See also

Notes

Footnotes

References

External links

Listen to this article (info/dl)

This audio file was created from a revision dated 2005-04-22, and does not reflect subsequent edits to the article. (Audio help)

More spoken articles

• Krishna Kunchithapadam, "What if anything is a Mitochondrial Eve?" a simple explanation
• Mitochondrial Eve and Y-chromosomal Adam Diagrams