References

Genetic variations among living humans present an exciting new sphere of research for investigators of modern human origins. The DNA sequences in our mitochondria and in the chromosomes in our cell nuclei hold a record of what we have gone through as a species. Deciphering that record has proven difficult, however, because many factors can influence the sequences and patterns in our DNA.

DNA resides in the cell nucleus (purple sphere in center of diagram) and in mitochondria (oblong bodies shown in green). Studies of mitochondrial DNA consistently point to a recent African origin of modern humans. (Illustration copyright Genetics Study Group, Department of Paediatrics, Tokyo Medical University, Japan.)

Scientists are struggling to sort out those various influences in order to paint a clearer picture of when, where, and how modern humans originated and what their relationship may have been to the archaic humans of the Old World, including Neanderthals.

Some types of DNA reveal genetic patterns that seemingly resulted from a recent African origin of our species, while others harbor signs of an older, multiregional origin. One type of DNA that favors a recent Out-of-Africa origin is mitochondrial DNA (mtDNA). Scientists have now succeeded in studying the entire sequence of the mtDNA molecule, in a worldwide sample of 53 people (Ingman, et al, 2000). As with previous mtDNA studies, this one found Africans to have the most diverse, and therefore the most ancient, mtDNA sequences. It also determined that African and non-African populations diverged not 100,000 years ago, as some other genetic studies have suggested, but as recently as 52,000 years ago. The results also indicated that after their split from Africans, non-Africans began expanding in population about 38,500 years ago.

Another study has focused on mtDNA variation in Europe and the Near East (Richards, et al, 2000), the former domain of Neanderthals. Researchers analyzed the mtDNA of more than 4,000 people from various European and Near Eastern populations. They found that 5 to 15 percent of the mtDNA pool of present-day Europeans could be traced back to the earliest part of the Upper Paleolithic period, some 45,000 years ago or earlier. They concluded that these earliest mtDNA sequences originated in early modern humans who colonized Europe, most likely from the Near East.

	Patterns in our mitochondrial DNA suggest that human populations grew greatly in size some time after modern humans had appeared, consistent with an expansion of modern humans out of Africa and across Eurasia. (Courtesy of Paolo Francalacci, Sezione di Antropologia, Dipartimento di Zoologia e Antropologia Biologica, Università degli Studi di Sassari, Italy.)

One of the most significant findings to come out of mtDNA studies is that non-Africans often show genetic signs of a severe reduction in population size, a "bottleneck," some time in the past, followed by a population expansion (Ingman, et al, 2000: 710-712). This bottleneck and expansion is presumed to have occurred when a branch of the early modern human population of Africa split off to form a small subpopulation, which then expanded in size as it spread out to colonize Eurasia. Some evidence from both mtDNA and nuclear DNA suggests that Africans also expanded in population in the past, either at the same general time as non-Africans (Zhivotovsky, et al, 2000), or earlier (Harpending, et al, 1993).

The latest mtDNA research suggests, however, that the expansions of Africans and non-Africans took place thousands of years after the first appearance of anatomically modern humans, which according to the fossil record occurred about 100,000 years ago. If correct, then the factor that facilitated the spread of modern humans across Africa and Eurasia would not have been some advantage inherent in the modern form itself, but was more likely the technological advances of the Upper Paleolithic, which began to appear about 45,000 years ago.

One problem with mtDNA, however, is that it has a high mutation rate. Some geneticists speculate that because of that high rate, most of the variations in mtDNA between individuals probably reflect mutations and population expansions from the past 50,000 years or so (Harris and Hey, 1999: 3323; Kaessmann, et al, 1999: 79). Information on more ancient demographic events and population relationships, perhaps crucial to understanding modern human origins, may be unreachable through mtDNA data.

This problem can be avoided, however, by studying variation in the Y chromosome, which has a much lower mutation rate than that of mtDNA. Like mtDNA, the Y chromosome is inherited through just one sex. But while mtDNA is inherited from the mother, the Y chromosome is passed from fathers to sons. Y-chromosome studies generally support the results of recent mtDNA analyses. They indicate that modern humans originated in Africa and that the human population expanded thousands of years after the first modern humans appeared. One study of Y-chromosome variation in a worldwide sample of over 1,000 men determined that Africans and non-Africans shared a common ancestor 59,000 years ago and that the non-African branch of humanity left Africa about 44,000 years ago (Underhill, et al, 2000).

Concerning the peopling of Europe, a Y-chromosome study of over 1,000 European men traced about half of their Y-chromosome lineages back to a mutation estimated to have arisen some 30,000 years ago (Semino, et al, 2000). The researchers think this mutation was associated with the spread into Europe of the Aurignacian, the early Upper Paleolithic culture often regarded as the hallmark of the first modern Europeans.

Central Asians, such as these Kyrgyz people, may have had Ice-Age ancestors who were the first modern humans to colonize Europe, if certain Y-chromosome evidence is correct. (Courtesy of Waugh's Central Asia web site. Photo © 1996 Daniel C. Waugh.)

The results suggest, however, that these Aurignacian people came not from the Near East, as is often assumed, but from Central Asia.

One problem with both the Y chromosome and mtDNA is that the evidence they provide of population expansions, used to support a recent origin and spread of modern humans out of Africa, may be an illusion created by natural selection (Hawks, et al, 2000). A low level of genetic variation can indicate that a population has recently expanded after branching off from a larger ancestral population. Natural selection, however, reduces genetic variation and can therefore mimic the effects of a population expansion. And since genetic diversity increases with time, a drop in diversity caused by natural selection can also make a genuine expansion appear to have occurred more recently than it actually did.

Selection can especially stifle variation in the Y chromosome and mtDNA, because they do not undergo recombination -- the reshuffling of DNA sequences that occurs when eggs and sperm are formed. Recombination influences the inheritance of alleles -- the variant forms, differing between individuals, that a DNA sequence takes at specific points along its length. Without recombination to break up DNA sequences, all the alleles along a sequence become "linked" -- they all get inherited together. As a result, selection for a favorable allele not only eliminates variants of that allele, but also variants of all other alleles in the sequence.

The genetic diversity of the Y chromosome is lower than that of all other chromosomes, and natural selection acting on linked alleles may partly explain why. But its diversity is also lower than that of mtDNA. This may be due to the fact that a woman often moves to the home territory of her husband when she marries (Seielstad, et al, 1998). This custom decreases diversity in men's Y chromosomes within a population and increases diversity in women's chromosomes and mtDNA. This factor, as well as natural selection, may be causing Y-chromosome studies to misinterpret demographic events related to the origin and spread of populations, and, in particular, to underestimate the age of those events.

Another type of nuclear DNA marker that scientists have used to shed light on modern human origins is called a short tandem repeat (STR) or microsatellite. A microsatellite consists of a series of two to five nucleotides that repeats itself consecutively a number of times in a DNA strand. Microsatellites are polymorphic, which means that different individuals carry different alleles of a microsatellite. These alleles differ in terms of the number of times the nucleotide sequence repeats itself. Because STRs are polymorphic, they are also called short tandem repeat polymorphisms (STRPs).

Microsatellite studies generally support a recent African origin for modern humans. Microsatellites are most diverse in Africans, and their genetic variation in non-Africans has been reported to be a subset of that in Africans (Calafell, et al, 1998). Similar results come from studies of another, longer type of tandem repeat called a minisatellite (Destro-Bisol, et al, 2000). Scientists have used microsatellite data to calculate that Africans and non-Africans split about 156,000 years ago (Goldstein, et al, 1995). Microsatellites also provide evidence of past population expansions presumably associated with the spread of early modern humans, although the evidence is inconsistent concerning exactly when the expansions occurred (Reich and Goldstein, 1998; Zhivotovsky, et al, 2000).

Certain other uncertainties also surround microsatellite analyses. Some researchers argue that the greater microsatellite diversity in Africans does not necessarily reflect a solely African origin for modern humans. It could have arisen, at least in part, from Africa having a larger population when modern humans were evolving (Hawks, et al, 2000: 15; Relethford and Jorde, 1999). Microsatellites also share with mtDNA the limitation of having a fast mutation rate, which places in doubt their ability to reflect population histories prior to the last 50,000 years (Harris and Hey, 1999: 3323; Kaessmann, et al, 1999: 79).

Geneticists have tried to increase the usefulness of microsatellites by studying them in combination with more stable markers with slower mutation rates. These combinations are called STRP haplotype systems. By analyzing STRPs and stable markers that are genetically linked, one can identify inherited haplotypes -- particular combinations of alleles of each type of marker. One can also assess how tightly linked particular alleles of each type of marker are to each other. Tight linkage, in which specific alleles are inherited together in a high percentage of members of a population, is called "linkage disequilibrium." The higher the linkage disequilibrium, the more recent the population.

Results from one study of an STRP haplotype system

	The DNA in our chromosomes presents a less consistent picture of our evolution than mtDNA, but the Y chromosome (bottom right) and several non-coding DNA sequences from other chromosomes suggest a recent African origin of modern humans. (Photo copyright Genetics Study Group, Department of Paediatrics, Tokyo Medical University, Japan.)

in more than 1,600 people indicate that non-Africans diverged from African ancestors 102,000 years ago (Tishkoff, et al, 1996). In the study, sub-Saharan Africans showed greater variability in haplotypes than non-Africans and also had less linkage disequilibrium, indicating that Africans are the ancestral population. Other studies of STRP haplotype systems report similar results.

Other nuclear DNA sequences also provide support for the view that modern humans originated and expanded relatively recently from a subset of the African population. Interestingly, all of these sequences are from regions of the genome that do not code for proteins. They either lie outside of coding genes or occur in non-coding segments within genes, known as introns.

Non-coding sequences from a portion of chromosome 16 called 16p13.3, from a region of the X chromosome labelled Xq13.3, and from intron 7 of the Duchenne muscular dystrophy (Dmd) gene, also on the X chromosome, have all provided evidence of greater genetic diversity in Africans than non-Africans and of a recent expansion of the human population (Alonso and Armour, 2001; Harpending and Rogers, 2000; Kaessmann, et al, 1999; Nachman and Crowell, 2000). But unlike the mtDNA and Y-chromosome data, which suggest the human population expanded long after modern humans had appeared, the 16p13.3 data indicate Eurasians expanded 106,000 to 143,000 years ago, at roughly the same time as the emergence of modern humans.

It is possible, however, that the genetic signs of a population expansion at these loci might have resulted from natural selection. Although non-coding regions such as these are unlikely targets of selection, they can be linked to coding regions that are targets. Selection acting on linked coding regions is especially likely to have affected the Xq13.3 locus, since it is a region of low recombination (Wall and Przeworski, 2000: 1872-1873). Although intron 7 of the Dmd gene is in a region of high recombination, the researchers studying that intron think it too may have been affected by selection (Nachman and Crowell, 2000). That is because another intron in this gene, intron 44, gives no hint of a population expansion. Since an expansion should affect all genetic loci equally, while selection would not, the researchers believe selection created the pattern at intron 7.

Other non-coding sequences of nuclear DNA also point to a past expansion of modern humans, but in conjunction with other evidence that threatens simple Out-of-Africa scenarios. Sequences from a non-coding region of chromosome 22 and a mostly non-coding region of chromosome 1, for example, suggest that Africans and non-Africans are more similar in their levels of genetic diversity than some other non-coding loci have indicated (Yu, et al, 2001; Zhao, et al, 2000). This suggests that if early modern humans leaving Africa went through a population-size bottleneck that reduced their genetic diversity, it must have been milder than has been supposed. Similar indications of a larger than expected population moving out of Africa come from

Our genes, which code for the proteins that make up our bodies, harbor DNA patterns that cast doubt on a recent African origin of modern humans. (Courtesy of the Human Genome Project Information web site of the U.S. Department of Energy Human Genome Program, Germantown, Maryland.)

a study of an STRP haplotype system (Mateu, et al, 2001) and from the analysis of intron 44 of the Dmd gene.

The chromosome 1 and 22 sequences also indicate that non-Africans shared a common ancestor long before modern humans appeared -- 634,000 years ago according to the chromosome 22 research (Zhao, et al, 2000), and an average of 757,000 to 805,000 years ago according to the chromosome 1 results (Yu, et al, 2001: table 8). These very ancient dates for non-African DNA sequences threaten the view that a single African population of modern humans exited Africa some 100,000 years ago and totally replaced all the archaic humans of Eurasia.

Coding regions of the nuclear genome present even greater problems for the Out-of-Africa theory. Not only do these regions yield evidence of an ancient, pre-modern ancestry for non-Africans, they also find no evidence of a past expansion of the human population. A study of a sequence encompassing the gene for the beta-hemoglobin (ß-globin) molecule in over 300 people from Africa, Asia, and Europe, for example, traced the common ancestor of Asians back more than 200,000 years, before the appearance of modern humans (Harding, et al, 1997). Instead of finding evidence of a population expansion, this study uncovered signs of a relatively steady population size and extensive gene flow across the Old World during the period when modern humans were evolving.

Similar findings come from a study of a portion of a gene, located on the X chromosome, that codes for part of the pyruvate dehydrogenase enzyme (Pdha1) (Harris and Hey, 1999). Genetic patterns at this locus indicate Africans and non-Africans split about 200,000 years ago. The researchers concluded that modern humans must have evolved in a geographically subdivided population and that genes associated with the transformation to a modern form probably flowed between populations.

This evidence of simultaneous evolution to a modern form among geographically separated populations would seem to favor the multiregional theory of modern human origins. But the Pdha1 researchers point out that at the time the transformation to a modern form was occurring, the African and non-African populations that gave rise to modern humans may not have been that widely separated geographically. The Pdha1 results may therefore lend support to a modified version of the Out-of-Africa theory known as the "weak Garden of Eden hypothesis" (Harpending, et al, 1993).

The strictest version of the Out-of-Africa theory, also called the "strong Garden of Eden hypothesis," holds that modern humans arose in Africa, the proverbial "Garden of Eden," 100,000 to 200,000 years ago and expanded out of the continent in a single large wave not long after that. But this fails to explain certain genetic signs that we had a small population over a long stretch of our evolution and that we did not expand in size until well after Africans and non-Africans had diverged from each other. Proponents of the weak Garden of Eden hypothesis attempt to account for these facts with the following scenario:

	Many geneticists see Africa as the sole birthplace of modern humans, but their views differ on the timing and pattern of early modern humans' movements beyond the continent. (Photo from the Safari Enthusiast web site. Courtesy of Dick Fishbeck.)

The line leading to modern humans split off from other African hominids long ago, perhaps as early as two million years ago (Harpending, et al, 1998), and stayed small in size for a long time. At some point, perhaps before modern human features had evolved, perhaps after, this lineage split into separate groups within Africa, which began to diverge genetically. These groups, some ancestral to later Africans and others to non-Africans, remained small. They also stayed close geographically, although some may have spilled beyond Africa into adjacent parts of the Near East. Then, some time after modern traits had appeared, these groups grew in size and spread out over the Old World. Time estimates for this expansion range from just over 100,000 years ago (Alonso and Armour, 2001) to after 30,000 years ago, the most recent dates coming from Y-chromosome studies (Shen, et al, 2000; Underhill, et al, 2000).

Weak-Garden-of-Eden advocates claim these genetically divergent groups could not have evolved into a modern form over all of Africa and Eurasia, as the multiregional theory would require. They note that current levels of genetic variation indicate we had an effective population size of only about 10,000 over most of our evolution. The effective size of a population is the number of reproducing adults, breeding under ideal conditions, needed to produce that population's observed amount of genetic diversity. With such a small effective size, our ancestors could not have maintained viable breeding groups over the vast reaches of the Old World inhabited by hominids over the last 1-2 million years (Harpending, et al, 1993, 1998). Most of those hominids must therefore have been genetically separate from us and off our evolutionary path.

Multiregionalists, however, dispute the view that we maintained a small effective size over most of our evolution (Hawks, et al, 2000). They claim that the human population could have grown exponentially over the course of its evolution and still left a genetic trail making it appear as if little or no growth took place until relatively recently. If true, our ancestors could have been rather numerous, which means they could have been the archaic humans who lived in relatively large numbers across Africa and Eurasia.

According to multiregionalists, our ancestry goes back two million years to the earliest Homo erectus populations of Africa. They claim that humanity arose at that time from a small population of hominids that became isolated from other hominid groups and evolved into the new species Homo erectus, which they prefer to call early Homo sapiens. To them, the DNA evidence is consistent with a genetic bottleneck occurring at that speciation event, followed by a population expansion as our Homo erectus, or early Homo sapiens, ancestors spread across Africa and Eurasia 1-2 million years ago. Multiregionalists deny that a sudden expansion took place at a later time, during the era of the first modern humans (Hawks, et al, 2000). They note that only some genetic loci point to this alleged later expansion. Other loci bear no trace of it.

An explanation may exist, however, for the contradictory signals given by different genetic loci. It has to do with the differing actions of two types of natural selection. One type, "directional selection," ensures that genes coding for favorable traits survive and are passed to the next generation, while unfavorable genes are weeded out. Directional selection acts to reduce genetic diversity. "Balancing selection," however, arises when certain factors in the environment enhance the fitness of individuals and populations that maintain genetic diversity at high levels. Diversity remains elevated even in the face of events such as population bottlenecks and expansions that would normally reduce it.

Some scientists have noted that the nuclear loci that show evidence of a population expansion are the very loci that should be least affected by directional selection, because they are almost exclusively from regions of the genome that do not code for genes (Rogers, 2001). In contrast, sequences that fail to show an expansion are primarily from coding regions, which could have been affected by balancing selection. Despite some exceptions to these patterns (Nachman and Crowell, 2000; Shen, et al, 2000), these researchers argue that the coding regions that point to a constant population size during the era when modern humans were springing up across the Old World are presenting a false picture due to balancing selection. In contrast, they trust as accurate the signs of an expansion apparent in non-coding loci.

Other geneticists take the opposite position. They believe that the evidence from some genetic loci of a population expansion during the era of the first modern humans is an illusion created by directional selection. They note that such selection can affect non-coding loci by acting on nearby coding regions linked to the non-coding segments. These researchers believe that loci indicating a more constant population size during our evolution present a truer picture of our past (Wall and Przeworski, 2000). They dispute the idea that balancing selection has affected these loci.

Until geneticists sort out the relative importance and prevalence of directional vs. balancing selection in our evolutionary past, it will be impossible to deduce which DNA sequences are telling us the truth about our origins, and which are deceiving us. And it will be difficult to know what contribution Neanderthals and other archaic humans may have made to our genetic makeup.

Back to Top