April 1, 2013 (Vol. 33, No. 7)

The discovery of the structure of DNA and the subsequent developments in genetics and genomics have had a great impact on all of the biological sciences, including human evolution.

Our ideas about human evolution 60 years ago came primarily from the fossil and archaeological records. These fields revealed that the last two million years were a dynamic period of our evolutionary history.

The human lineage two million years ago was a population with ape-sized brains limited to sub-Saharan Africa. The human lineage expanded into Eurasia around 1.85 million years ago, and our brain size increased throughout the Pleistocene. Anatomically modern humans first appeared in Africa about 200,000 years ago, with anatomically modern forms appearing outside of Africa at more recent dates.

Since the 1980s, genetics has become an increasingly important source of information about human evolution over the last two million years, with one of the more dramatic examples being the sequencing of genomes directly from fossils. What was revealed is that DNA sequences from archaic Eurasians are present in living humans confirming that the anatomically modern human populations that expanded out of Africa hybridized at low levels with the Eurasian populations that they encountered—the mostly-out-of-Africa hypothesis of modern human evolution (Templeton, 2002).

Limited Use

Although these ancient-DNA studies attract public attention, there are severe limitations to their use in making evolutionary inferences. Evolution is a population-level phenomenon, and tests of many evolutionary hypotheses require population samples. The human fossil record is sparse, and this fact alone limits the types and reliability of evolutionary inferences from human fossils.

Fortunately, there is another type of fossil amenable to modern molecular genetic techniques—DNA itself. The DNA sequences that exist in current human populations are replicated copies of DNA that existed in the past, albeit with possible mutational changes. The evolutionary history of DNA regions is most clearly written in parts of the genome that experience little to no recombination, as recombination can place together two or more DNA segments that may have had different evolutionary histories.

It is for this reason that much work on human evolution over the past quarter of a century has focused on the nonrecombining elements of the human genome: our mitochondrial genome and the Y chromosome. We now know that recombination in the nuclear human genome is often concentrated into hotspots, with areas of no to low recombination in between. Therefore, much more of the human genome is now available as a source of “fossil” DNA.

To convert the evolutionary histories of DNA regions with little or no recombination into inferences about evolution, we need to generate population-level expectations about how past events or processes could influence these DNA-level evolutionary histories. Here we rely on coalescent theory.

Coalescent theory models how the current array of genetic variation found in an homologous DNA region in a current population traces backward through time, with the number of DNA lineages coalescing (the time inverse of DNA replication) one by one until all the variation observed today traces back to a single ancestral DNA molecule.

This coalescent process is influenced by a number of factors that took place in the past—mutation, fragmentation, range expansions, gene flow patterns, etc. A partial evolutionary history of a low/no recombination DNA region of interest can be reconstructed from the DNA replication events marked by mutational change. Hence, our “fossil” DNA histories can give insight into past evolutionary processes and events.


One powerful way of extracting this information about past evolution is through multilocus nested clade analysis (MLNCA). This method converts the evolutionary history of a DNA region with little to no recombination into a series of nested branches (clades), which captures time (the deeper the branch in a nested series, the older the time), and then overlays the spatial distribution of the currently observed genetic variation upon the nested series. In this manner, we can estimate the evolutionary history of current variation through both space and time.

Whenever a statistically significant temporal/spatial association is detected, the pattern of association is interpreted in light of coalescent theory. One-hundred and fifty positive controls (cases where we know what happened to a species through some other source of information) were used to validate the coalescent inference structure, making MLNCA the most empirically validated method of phylogeographic inference.

Because the coalescent process is noisy, we do not expect all DNA regions to have identical or even congruent evolutionary histories, but we do expect past demographic events or processes to affect the coalescent process of multiple regions. Therefore, all inferences are cross-validated across loci and subject to a maximum-likelihood hypothesis test before being accepted. This method of cross-validation with explicit hypothesis testing has been shown by computer simulations to reduce false positives below the 5% nominal level commonly used in statistics (Templeton, 2010).

MLNCA does not require a prespecified model of evolution; rather, the model emerges naturally out of the cross-validated statistically significant inferences. Thus, there is no inherent bias toward any a priori model of human evolution. The cross-validated MLNCA inferences produced a model of human evolution that had some features of previous models, but unique features as well (Figure).

Figure. Significant inferences about human evolution from MLNCA. Geographical location is indicated on the x-axis, and time on the y-axis, with the bottom of the figure corresponding to two million years ago. Vertical lines indicate genetic descent over time, and diagonal lines indicate gene flow across space and time. Thick arrows indicate statistically significant population range expansions. Lines of descent are not broken because the population range expansion events were accompanied by statistically significant admixture.

Moving Out of Africa

The first detected event is an expansion out-of-Africa into Eurasia dated to 1.9 million years ago using a molecular clock—an inference confirmed by the fossil record. Unlike any other model of human evolution at the time, a second expansion of nonmodern humans out-of-Africa into Eurasia occurs about 650,000 years ago, which corresponds well to the expansion of the Acheulean tool culture out of Africa.

These Acheulean populations did not replace the Eurasian populations they encountered, but rather admixed with them. Moreover, after Acheulean expansion, there was significant, albeit limited, recurrent gene flow between Eurasia and sub-Saharan Africa.

Next, there was a third major expansion of humans out-of-Africa into Eurasia at 130,000 years ago, corresponding to the fossil record of the spread of anatomically modern humans out-of-Africa beginning 130,000 to 125,000 years ago and reaching far Eastern Asia by 110,000 years before present.

Like the earlier Acheulean expansion, this third expansion out of Africa resulted in low levels of admixture with Eurasian populations, not complete replacement. Indeed, the hypothesis of replacement (zero admixture) was rejected with a 10-17 probability level, making this the most definitive conclusion of the analysis.

After this third out-of-Africa expansion, human populations subsequently expanded into Northern Eurasia (including northern Europe), the Americas, and the Pacific. Wherever humans went, patterns of genetic interchange were soon established (Templeton, 2005).

These results were extremely controversial when first published because of their strong rejection of the out-of-Africa replacement model, which posited that anatomically modern humans, when they expanded out of Africa into Eurasia, drove all the native Eurasian populations to complete extinction with no admixture. Even a mostly out-of-Africa expansion was anathema to the replacement advocates because a low level of admixture can still have strong evolutionary consequences.

Interestingly, the out-of-Africa replacement hypothesis rose in popularity due to a genetic analysis based on mitochondrial DNA (Cann et al., 1987). The evolutionary history of human mitochondrial DNA consisted of African-only branches at the most ancient part of the history followed by a mix of African and non-African branches, with no deep divergences (that is, branches separated by many mutational changes with no intermediates present in current populations).

In the 1980s, there were three principle models for human evolution: 1) the out-of-Africa replacement model; 2) the multiregional model that proposed that humans evolved toward modernity across the globe because of genetic interchange between African and Eurasian populations, and 3) the candelabra model that posited that Africans, Europeans, and East Asians evolved to modernity as separate evolutionary lineages that diverged after the initial expansion out of Africa with little to no gene flow since then.

No Deep Divergences

The fact that the mtDNA evolutionary history had no deep divergences and coalesced to a common, ancestral mtDNA molecule about 200,000 years ago certainly falsified the candelabra model. Unfortunately, the original paper reporting this mitochondrial evolutionary history falsely equated the candelabra and multiregional models, thereby leaving only the replacement model.

However, the mitochondrial evolutionary history was also compatible with the multiregional model. Indeed, there has never been a genetic dataset or analysis that favored the replacement model over alternatives in a statistically significant fashion (Templeton, 2007). Nevertheless, the replacement model became the standard model of human evolution through the 1990s onward.

Now that ancient DNA studies offer direct confirmation of the MLNCA inference that there was admixture, this major controversy in human evolution can now be regarded as settled—at least, as settled as any scientific debate can be.

The future of human evolution studies is particularly bright due to genetics. As techniques become more refined, we can anticipate more ancient DNA studies that allow us a direct glimpse into our evolutionary past. Already, new sequencing techniques are producing much data on current populations that can be analyzed with MLNCA and newer parametric phylogeographic techniques that are complementary to the non-parametric MLNCA (Templeton, 2010).

Hence, there is much potential for further evolutionary discoveries about our shared history.

Alan R. Templeton, Ph.D. ([email protected]) is the Charles Rebstock Professor of Biology and Genetics and Professor of Evolutionary and Environmental Biology in the Department of Biology at Washington University in St. Louis.

Cann RL, Stoneking M, Wilson AC (1987). “Mitochondrial DNA and human evolution.” Nature 325, 31-36.
Templeton AR (2002). “Out of Africa again and again.” Nature 416, 45-51.
Templeton AR (2005). “Haplotype trees and modern human origins.” Yearbook of Physical Anthropology 48, 33-59.
Templeton AR (2007). “Perspective: Genetics and recent human evolution.” Evolution 61, 1507-1519.
Templeton, AR (2010). “Coherent and incoherent inference in phylogeography and human evolution.”
Proceedings of the National Academy of Sciences 107:6376-6381.

Previous articleDainippon Sumitomo Invests $16M in RIKEN Company
Next articleArQule Regains Rights to AKT Inhibitor from Daiichi Sankyo