A remarkable feature of modern humans is our wanderlust, which the poet Charles Baudelaire famously referred1 to as “l'horreur du domicile”. From our evolutionary birthplace in Africa2, modern humans have migrated to nearly every habitable corner of Earth (Fig. 1), overcoming obstacles such as ice, deserts, oceans and mountains. The number, timing and routes of human dispersals out of Africa have implications for understanding our past and how that past influenced contemporary patterns of human genomic variation. Three studies on pages 207, 201 and 238 (Malaspinas et al.3, Mallick et al.4 and Pagani et al.5) describe 787 new, high-quality genomes of individuals from geographically diverse populations, providing opportunities to refine and extend current models of historical human migration.
In the past decade, the maturation of whole-genome sequencing technology has enabled data to be generated on a scale that was previously difficult to imagine. Genome-scale studies in humans, such as the 1000 Genomes Project6, which was completed last year, have contributed to a catalogue of genetic variation and genomic regions that confer the ability to adapt to diverse environments. Nonetheless, existing genetic data are often constrained by several factors, including limited breadth of population sampling and low-coverage data (in which each region of the genome is sequenced only a few times, leading to high error rates and missed variants). To address this issue, the current studies collect high-coverage sequence data for individuals from more than 270 populations across the globe. By studying the genetic diversity within and between these populations, the groups can tackle many questions about our past.
Cataloguing genetic data from indigenous populations, which are often difficult to access and are rapidly disappearing, is an important achievement. Mallick et al. and Pagani et al. made great efforts to comprehensively sample regions that are typically under-studied; these include African populations, which have considerable genetic, linguistic and cultural diversity. Similarly, Malaspinas et al. describe the first extensive survey of human genetic diversity in Australia — a poorly studied region that, together with New Guinea, contains some of the earliest archaeological and fossil evidence of modern humans outside Africa.
The high-resolution portrait of human genetic diversity afforded by these studies allows new inferences to be made about our migration out of Africa. There are currently two conflicting models for such human dispersal. The first hypothesizes a single event that occurred around 40,000–80,000 years ago. Under this scenario, all present-day non-Africans trace their ancestry to a single population. By contrast, the multiple-dispersal model7 posits that an initial migration out of Africa occurred as early as 120,000–130,000 years ago8, culminating in the peopling of southeast Asia and Australasia, possibly via a southern migration route along the coastline of the Arabian peninsula and the Indian subcontinent. This early dispersal was followed by a second migration from Africa, through the Levant, which resulted in the peopling of mainland Eurasia.
Superficially, the current studies seem to come to different conclusions about out-of-Africa dispersals. Pagani et al. found that about 2% of genomes from individuals of Papua New Guinean ancestry indicate that their ancestors separated from Africans earlier than did other Eurasians. This observation is consistent with a multiple-dispersal model in which an early expansion of modern humans from Africa led to the peopling of Australasia around 120,000 years ago. This early out-of-Africa migration would have been followed by subsequent dispersals, and would have contributed only a small amount of ancestry to present-day Papuan individuals. Cranial morphology and other genetic data also support the idea of an early expansion9.
Malaspinas et al. and Mallick et al. consider a different sequence of events, in which all contemporary non-Africans branched off from a single ancestral population. Malaspinas and colleagues provide evidence that, on leaving Africa, modern humans immediately separated, leading to two waves of dispersal. As previously proposed10, one wave led to the peopling of Australasia, whereas the other contributed to the ancestry of present-day mainland Eurasians. Mallick and co-workers propose that this early separation instead occurred between west and east Eurasians, meaning that present-day people in Australia and Papua New Guinea might be descended from the same wave as east Asians.
However, neither Mallick et al. nor Malaspinas et al. exclude the possibility of multiple out-of-Africa dispersals. Indeed, their models are consistent with earlier dispersals, as long as these early voyagers made little or no contribution to the gene pool of contemporary non-African populations (which is essentially what Pagani et al. find). Studies of ancient DNA clearly show that large-scale population turnovers have happened throughout human history: populations that once lived in Eurasia, for example, vanished without a trace, except for their bones11,12. Thus, although some differences between the proposed models are yet to be reconciled, they are not as disparate as they might seem to be.
The three studies also provide resources to better define models of genetic mixing between modern humans and their archaic hominin relatives, such as Neanderthals and Denisovans. Malaspinas and colleagues propose that the genomes of present-day Aboriginal Australians might harbour traces of an ancient liaison with an unknown hominin group. Although evidence for gene flow from an unknown hominin group is tentative, it highlights the potentially surprising things that can be learnt from a comprehensive sampling of human genomic variation.
These studies fill in some missing pieces in the puzzle of human history, but many fascinating questions remain. The continued sampling of human genomic diversity and the development of increasingly sophisticated statistical tools promise to reveal more secrets about our past. Nonetheless, it is crucial to recognize the limits of genetics. As previously pioneered13, the integration of data across traditionally distinct disciplines, such as linguistics, archaeology, anthropology and genetics, will be necessary to fully retrace the steps taken by early humans as they explored and colonized the world.