Modern human DNA analyses with special referenceto the inner dual-structure model of Yaponesian

Timothy A. Jinam, Yosuke Kawai , Naruya Saitou

Anthropological Science Vol. 129(1), 3–11, 2021



We explored the possibility of a three-wave migration model by examining three datasets of modern human DNA:
(1) whole mitochondrial (mt) DNA genomes of 1642 Yaponesians;
(2) mtDNA haplogroup frequencies of 59105 Yaponesians from 47 prefectures; and
(3) genome-wide SNP data of two Yaponesians (Ainu, Okinawa) and whole-genome sequence data of Yamato individuals, the Funadomari Jomon F23 individual, and three East Asian populations (Korean, northern Chinese, and southern Chinese).


The term ‘Yaponesia’ was proposed by Toshio Shimao in the 1960s (e.g. Shimao, 1977) to refer to the Japanese archipelago, and later Saitou (2015, 2017) defined the three geographical areas of Yaponesia:

northern Yaponesia (Sakhalin Island, Kuril Islands, and Hokkaido Island);
central Yaponesia (Honshu, Shikoku, Kyushu Islands, and surrounding smaller islands); and
southern Yaponesia (Nansei Islands including Amami and Okinawa regions).

Saitou (2015, 2017) and Saitou and Jinam (2017) proposed a three-wave migration model for Yaponesia..

(1) Paleolithic and the middle period of Jomon (40000 years before present (BP) to 4500 BP) for the first wave;
(2) late and final Jomon periods (4500 BP to 3000 BP) for the second wave;
(3) Yayoi period (3000 BP) to present day for the third wave.

The people involved in those three waves may be hypothesized as follows. Hunter-gatherers from Siberia, continental East Asia, and/or Indochina migrated to Yaponesia during the first wave.

We propose a somewhat mysterious ‘sea people’ as the second-wave migrants.
 They may have been huntergatherers who mainly relied on fishing and were distributed from the coastal area of southern China to the Shandong peninsula, the Yellow Sea, and the Korean peninsula.

Shinoda et al. (2019) analyzed ancient DNA of people who lived on Gadeok Island, located of the southeast coast of Korean peninsula about 6300 years ago, and reported that they had a much higher Jomon component than modern Koreans.

 Gadeok Islandは加徳島で韓国釜山の近くの島
These two literatures seem to support the existence of ancient hunter-gathers who mainly relied on fishing on coastal areas of continental East Asia. These ‘sea people’ were replaced by rice farmers whose population size quickly expanded between 7000 BP and 5000 BP, and these ‘sea people’ eventually migrated to Yaponesia during the late and final Jomon periods (4500 BP–3000 BP).

Rice agriculture was introduced 3000 BP (e.g. Fujio, 2015; Nasu and Momohara, 2016) by the third-wave migrants whose homeland was somewhere in continental East Asia.


This migration wave still continues today, for the highest proportion of international marriages in Japan includes Chinese and Koreans (e-Stat, 2018). We will discuss the plausibility of this three-wave migration model based on genome-wide single-nucleotide polymorphism (SNP) and mitochondrial (mt) DNA analyses.

Materials and Methods

The first dataset was a total of 1642 complete mitochondrial genomes from Yaponesians (Table 1).
These 1642 Yaponesian mtDNA genome sequences were used to estimate effective population size changes over time.

The second dataset used was mtDNA haplogroup frequency data of 59105 Yaponesians living in all 47 prefectures provided by Genesis-HealthCare, Ltd, as shown in Supplementary Table

We classified 47 prefectures into a central axis area and a peripheral area based on Saitou’s (2017) proposal as shown in Table 2 and Figure 1.

The third dataset comprised 639912 autosomal SNP data from two Japanese populations (Ainu, Ryukyuan) reported by the Japanese Archipelago Human Population Genetics Consortium (2012).

This SNP dataset was merged with whole-genome sequence data of the Funadomari Jomon F23 individual (Kanzawa-Kiriyama et al., 2019), 27 Japanese (Bergström et al., 2020), 40 Koreans (Zhang et al., 2014), 45 Han Chinese Beijing (CHB), and 45 southern Han Chinese (CHS) (Bergström et al., 2020).

After pruning the dataset for linkage disequilbrium (LD), the resulting 192898 SNP data were used for admixture analysis

Results and Discussion

Dataset 1

There were sharp increases in the population size during the last 3000 years when using either coding region only (Figure 2A) or complete mtDNA (Figure 2B).

2021-05-14 (2)
This increase in population size corresponds to the start of the Yayoi period, just after the Jomon period (16500 BP–3000 BP) (Fujio, 2015; Yamada, 2019). 

The bottleneck followed by a rapid increase of the population size was noted both by our Figure 2 and by Figure 2 of Okada et al. (2018).

They observed a clear bottleneck during the early Yayoi period followed by a sharp increase in population size. The D haplogroup is known to be distributed only among Japanese, Tibetan, and Andamanese, and is thought to have contributed to the formation of the Neolithic Jomon males in prehistoric Japan (Kanzawa-Kiriyama et al., 2019). Watanabe et al. (2019) argued that this sharp decrease was due to a shortage of food resources in the colder final Jomon period.

Dataset 2
2021-05-14 (5)

Figure 6 shows a two-dimensional PCA plot of 47 Japanese prefectures using the frequencies of 36 mtDNA haplotypes frequencies presented in Supplementary Table.

We constructed phylogenetic networks of prefectures using mtDNA haplogroup frequency data. Figure 5 shows a network of all 47 prefectures.
2021-05-14 (6)

We thus made another phylogenetic network for the 46 prefectures without Okinawa (Figure 6).

Dataset 3

If we fit this k = 3 situation to the three-wave migration model proposed by Saitou (2015, 2017), the blue, green, and orange components may correspond to first, second, and third waves, respectively.

2021-05-14 (9)


We used three datasets in this study, and each dataset was analyzed using different methods. Dataset 1 comprises mtDNA genome sequences, and the population size change in the past was inferred from these data of modern Yaponesian. These mtDNA data clearly showed a steep increase in the population size after the Yayoi period. This confirms the shift from hunting and gathering to agriculture after the Yayoi period started. Dataset 2 comprises mtDNA haplogroup frequency data of 47 prefectures, and the dichotomy between central-axis and peripheral populations supports our ‘inner dual-structure’ model. Dataset 3 (genome-wide SNP data) seems to be consistent with the three-wave migration model. We need to compare nuclear DNA data of many areas of Yaponesia to strengthen the existence of the ‘inner dual-structure’ and the three-wave migration model.


Zhang W., Meehan J., Su Z., Ng H.W., Shu M., et al. (2014)

Whole genome sequencing of 35 individuals provides insights into the genetic architecture of Korean population.