Modern human migrations in insular Asia according to mitochondrial DNA and non-recombining Y chromosome


  • 3D-S8-02

Jean Trejaut, Mackay Memorial Hospital, #45 Min-Sheng Road, Tamsui 25115, Taiwan

Mitochondrial DNA (mtDNA) and non-recombining Y chromosome (NRY) are inherited uni-parentally from mothers to daughters or from fathers to sons respectively. Their high polymorphism (and rate of mutation) has been used to confirm the ‘Out of Africa’ hypothesis, and to time the spread of farming by Neolithic farmers (between 8000 and 5000 years ago) from the Levant to Europe and from Mainland Southeast Asia (MSEA) to Island Southeast Asia (ISEA).

Here, we first combined results obtained from sequencing of the mtDNA hyper variable regions (HVS I and II) and other relevant coding regions among 1400 individuals from ISEA, Taiwan, Fujian and MSEA (Thailand, Vietnam). Analysis of the extensive polymorphism confirmed genetic continuity, revealing very specific genetic profiles gradually changing eastward across the regions. Lineages that best represented a clade in the initial phylogeny (a branch of the genetic tree) were further analysed using complete genomic mtDNA sequencing. Finally, a most parsimonious tree was constructed. While establishing the most up to date and informative mtDNA dataset available on populations of ISEA and Taiwan, we could infer more clearly, ancient population migration routes between MSEA and ISEA. When combined with worldwide studies, the African haplogroup L3 was, as expected, the precursor to super-haplogroups M and N. We note here that M and N are the ∼65 000 years’ ancestors of all haplogroups seen in populations of the northern hemisphere, Australia and the New World.

In genetics, inferences about the genetic makeup of the past are drawn from studies of modern-day populations. According to archaeological and linguistics studies, it is now generally accepted that Western ISEA (Sumatra, Borneo, Java, Philippines and Sulawesi) have been a very active demographic platform of expansion and dispersal of the first eastern settlers in early Palaeolithic (∼55 000–45 000 years ago). The people who remained in the tropics are the Melanesian of today, and those who, very early, moved north, formed the east Asian populations. Some of them, much later (<10 000 years) after favourable climate changes, moved back to ISEA and even further to settle Polynesia in the Pacific and Madagascar in the Indian Ocean. The most favoured model, the ‘out of Taiwan’ was initially based on linguistic arguments, and suggests that the Austronesian-speaking populations of ISEA, Near Oceania, and Remote Oceania (the Polynesians) have a common origin among early Taiwanese agricultural groups who dispersed into ISEA ∼4000 years ago, reaching Near Oceania ∼3500 years ago [1,2]. In the process, these people completely replaced the culture (cultural diffusion) and the local populations of ISEA (demic diffusion).

Interestingly, our study showed that 86% of mtDNA sequences in ISEA and Taiwan were characterized by sub-haplogroups of super haplogroup M (M7b3*, M7c3c*, M7b1*, D5b3*, D6*, E*; with ‘*’ indicating all subgroups thereafter) or N (B4a1a1*, B4a2*, B4c2*, B4d1*, B5a1b*, B5a2a*, F1a1a*, F1a3*, F1a4*, F3b*). These haplogroups are not seen among the Melanesian peoples (the descendants of the palaeolithic settlers such as the Australian Aborigines and the people of Near Oceania: Papua New Guinea, Bismark archipelago), but they are predominantly seen among Austronesian-speaking groups, in Taiwan and ISEA (Philippines and Indonesia). In the Pacific and Madagascar, their mtDNA diversity was mostly restricted to mtDNA haplogroup B4a1a1 and B4a1a1a (the precursor of the Polynesian motif and the Polynesian motif).

Our analysis supported that these mtDNA haplogroups had Southeast Asian origin and were therefore the result of an ancient demic and cultural diffusion.

To obtain a more objective and balanced genetic viewpoint, we re-analysed all poorly defined mtDNA haplogroups using complete sequencing and also extended the analysis to the study of the paternal polymorphism obtained from slowly evolving Y single nucleotide polymorphism (Y-SNP) and the much faster evolving Y chromosome short tandem repeats (Y-STR).

Interestingly, in western ISEA, migration events inferred from Y chromosome analysis generally agreed with mtDNA and also suggested some possible bidirectional population movements from Taiwan to ISEA (the out of Taiwan) or from ISEA to Taiwan. On the contrary, further eastward, migration of Austronesian speakers into the pacific ended up with populations (the Polynesians, north coast New Guinea and many Bismark islands) displaying very different maternal and paternal genetic profiles: maternally Austronesian with a strongly reduced polymorphism, and paternally Melanesian).

Together, the polymorphism of mtDNA and NRY chromosome unveiled outstanding findings on the evolutionary history of East Asia, insular Asia and near Oceania. For example, if Palaeolithic people did settle Taiwan, our study shows that demic diffusion by people from MSEA was complete (total replacement). On the other hand, we found that this was not the case for the Philippines and Indonesian where approximately 8–14% of Palaeolithic genetic markers (Melanesians) could still be found nowadays in the populations of western ISEA. In agreement with most recent studies, our analysis suggested that the relationship between linguistics, archaeology, demography and genetics was much more complex than first hypothesized.

Some results of this analysis are discussed below using a conceptual approach.

Palaeolithic expansion of the Melanesians (mtDNA haplogroup P)

In 2005 and 2007, Friedlander and Hudjashov [1,2], produced trees of complete mtDNA sequences using founding super-haplogroups M and N. Initial analysis of the phylogeography of these haplogroups suggested that Aboriginal Australians were most closely related to the autochthonous populations of New Guinea/Melanesia. According to the molecular clock, it also indicated that prehistoric Australia and New Guinea were occupied initially during a unique Palaeolithic colonization event ∼50 000 years ago. The question still remained as to whether PNG and Australia were reached separately, sequentially, several times after an initial settlement event or if there were exchanges between them after an initial settlement. For this, Friedlaender and Hudjashov separately re-analysed the distribution of all subtypes of Melanesian mtDNA haplogroup M and N. Only one subtype of super haplogroup N (haplogroup P) will be described here [3,4] to show that it is during the early undifferentiated period of haplogroup P (∼51 000 years ago) that anatomically modern humans most likely moved separately to PNG and Australia. As the result of isolation, haplogroups P1 and P2 expanded in PNG while P5, P6, P7 and P9 expanded in Australia. More recently described subtypes of P, P3 and P4, were seen in PNG and Australia [5] and indicated possible gene flow between the two regions. The phylogenetic tree in Fig. 1 shows different subtypes of P3 or P4 in Australia (P3a and P4b1) and in PNG (P3b and P4a/b), suggesting gene flow. This also indicates a period of expansion of the undifferentiated P haplogroup in either Australia or PNG/Melanesia where P3 and P4 first appeared, (most possibly in PNG/Melanesia as the diversity there is highest). A later gene flow between 39 000 and 15 000 years of P3 and P4b from PNG/Melanesia to Australia must have been followed by an uninterrupted period of isolation till the present days. This allowed P3 and P4b to differentiate into P3a and P4b1 in Australia, while P3b and P4a speciated in PNG/Melanesia. Alternatively, P could have first differentiated in eastern ISEA into P3 and P4a/b and separately moved to Australia and PNG (39 000 years ago) where they differentiated locally.

Figure 1.

 Phylogeography of mtDNA haplogroup P: most Parsimonious tree of haplogroup P (P is a subtype of super haplogroup N and macro haplogroup R). Before 2009, all known branches of P (P1–P6) were seen either in Australia or in Melanesia (PNG). Here, new branches, namely P8 and P10, were found in the Philippines [4]. Sequence accession number can be obtained from Phylotree [12]. P3a and P4b1 are only seen in Australia, and P3b and P4a are only seen in PNG. Their coalescent time estimate ranges between 15 000 and 30 000 years. This indicates either an early dispersal of the unspeciated P3 or P4 from PNG or from Australia, or separate dispersal of P3 and P4 from ISEA with no later gene flow between PNG and Australia.

In summary, it appears that PNG and Australia were first reached separately (51 000 years ago). A later gene flow from PNG to Australia (<39 000 years ago) is most probable but will not be ascertained till further study is conducted among the populations of eastern ISEA.

Our group later described the presence of two new subtypes of haplogroup P in the Philippines, haplogroups P10 and P8 [3,4]. Interestingly, there is no traces of haplogroup P in the region situated between PNG and the Philippines. P may have either disappeared by drift or has not yet been sampled. In any case, when ruling out the possibility of mutation recurrence, or back migration from PNG towards western ISEA, our findings suggest that haplogroup P made its first appearance in Western ISEA, very shortly before the settlement of PNG and Australia. This idea is further reinforced by the presence of another haplogroup P in Malaysia (R21). These findings strongly indicate that Western ISEA was an active platform for the Palaeolithic expansion and dispersal of the first modern human settlers.

An unexpected high number of ‘basal lineages’ that, like haplogroup P, branch directly from super-haplogroups M or N, were also found throughout ISEA [3,4]. The presence of such unique and unshared basal lineages, all along the southern hemisphere coastlines, from the horn of Africa [6] through ISEA, and then New Guinea or Australia [1], suggested that the first colonists had a small effective population size, and progressed rapidly in their eastward migration, most likely at a rate of ∼700 m per year [7]. Moreover, the presence of this unexpectedly large number of novel basal haplogroups in West ISEA reinforced the idea that western ISEA was an active centre of expansion and of dispersal in early Palaeolithic. The high frequency of these basal haplotypes (14% in Java only), further suggest that maternal demic diffusion (total replacement), universally accepted as the result of an early Holocene wave of migration from MSEA, was not complete. The remaining 86% of sequences in ISEA were characterized by haplotypes that belong to an already well defined and much younger twig of haplogroup M, such as G1, D4, M9, M7, M13 and Z, and of haplogroup N such as B4, B5, F1 and N9a. Today, these haplogroups are all commonly seen in ISEA non-Melanesian populations, and most particularly among Austronesian speakers.

Neolithic expansion of the Austronesian speakers (mtDNA haplogroup B4)

Only one branch of super-haplogroup N (haplogroup B4 and its subgroup haplogroup B4a1a) is described here (Fig. 2). Using a large number of complete mtDNA B4 genomes, supplemented by data obtained from the analysis of control region genotyping, we show new and very convincing evidences that the most prevalent (80–90%) present days maternal lineages among Pacific islanders (haplogroup B4a1a1a) descended ∼20 000 years ago, during the last glacial maximal, from an east Asian mainland mtDNAs ancestor bearing haplogroup B4a1 [3,8,9]. The next descendant in line, haplogroup B4a1a, appeared at the beginning of the Holocene (8500 years ago), nowadays, it is only seen among Austronesian speakers in Western ISEA/Taiwan. B4a1a is the closest ancestor to the proto-Polynesian and Polynesian motifs. Since the origin of Austronesian languages is not older than 6000 years, this suggests that the ancestors of group of people speaking Austronesian languages today had already genetically differentiated from the mainland populations, and they had done this in insular Asia, before they actually spoke Austronesian. Some questions remain: were the bearers of B4a1a proto-Austronesians speakers? Is the absence of B4a1a on mainland Asia the result of drift? Till this is answered, the possibility for an origin in southeast Asia of proto-Austronesian speakers’ bearers of B4a1a remains open.

Figure 2.

 Phylogeography of mtDNA haplogroup B4 and its of sub-haplogroups in mainland East and North Asia (green), in ISEA/Taiwan (the Taiwan motif: B4a1a, yellow), in Near Oceania (the proto-Polynesian motif: B4a1a1, black) and in the Pacific (the Polynesian motif: B4a1a1a, red). The four evolutionary stages of expansion of haplogroup B4 are region-specific.

While there is strong correlation between archaeology and genetics on the path of migration taken by the bearers of all descendants of haplogroup B4a1a on their eastward movements towards the Pacific ocean, the timing of the staging of people movements determined by genetics and the timing associated to the ‘out of Taiwan’ model determined by linguistic and archaeological studies are not compatible. The ‘out of Taiwan’ proposes a dispersal beginning about 4000–5000 years ago. In contrast, the phylogeographic model (starting from B4a1a in western ISEA) reveals that its descendant haplogroup, the ‘Polynesian motif’ (B4a1a1a), has never been seen in Taiwan, and had its major expansion approximately 6650 years in the Bismarck Archipelago where it has now reached its higher frequency and diversity. Interestingly, these dating estimates are also compatible with recent Y-chromosome [10] and autosomal data [11], thus providing an excellent synthesis from all genetic systems [8].

These results describe an early B4a1a Holocene incubation period in ISEA/Taiwan and then expansion of the Polynesian motif (B4a1a1a) ∼6650 years ago in Near Oceania, which most importantly predated the development of the Lapita culture. The Lapita culture is a cultural complex including obsidian tools, finely decorated dentate-stamped pottery and shell ornaments that first appeared on the coasts of the Bismarck Archipelago ∼3500 years ago, spread into Remote Oceania ∼3000 years ago and is generally believed as the ‘land mark’ being the most strongly associated with the arrival of the first Austronesian speakers in Melanesia.

In brief, during the first stage of the Austronesian Diaspora (West ISEA to Near Oceania), language and culture progressed when the genetic profile of people in ISEA was already well established, but in the second stage (Near Oceania to the Pacific), language, culture and genetic makeup moved in concert.


No potential conflict of interests to declare.