Domestication of azuki bean and soybean in Japan: From the insight of archeological and molecular evidence

Domestication of azuki bean and soybean has enabled them to acquire non-dormant seeds, non-shattering pods, and larger seed size. Seed remains of the Jomon period recently discovered at archeological sites in the Central Highlands of Japan (6,000–4,000 BP) suggest that the use of azuki bean and soybean and their increase in seed size began earlier in Japan than in China and Korea; molecular phylogenetic studies indicate that azuki bean and soybean originated in Japan. Recent identification of domestication genes indicate that the domestication traits of azuki bean and soybean were established by different mechanisms. Analyses of domestication related genes using DNA extracted from the seed remains would reveal further details about their domestication processes.

Crop plants have been established through the domestication of wild plants. Wild plants have been bred into their current crop phenotypes over thousands of years through repeated cultivation and selection by humans. Abbo et al. (2014) defined domestication as follows: "In biological terms, domestication refers to the major genetically based phenotypes characterizing plants selected by humans. In cultural terms, domestication is an episode based on a decision and follow-up action by which humans have chosen certain species and selected particular stocks for growth. Thus, domestication involves obtaining desirable plants with distinct phenotypes by taking educated and conscious decisions". The "phenotypic traits characterizing plants selected by humans" in this definition are generally referred to as domestication traits. In legumes, the primary domestication traits are seed dormancy, pod shattering, and seed size (Abbo et al. 2014, Fuller and Allaby 2009, Smýkal et al. 2014.

Seed dormancy
While wild legume seeds exhibit seed dormancy, domesticated legumes produce non-dormant seeds. Dormant seeds refer to the phenotype in which the germination of Communicated by Toyoaki Anai Received September 8, 2022. Accepted January 5, 2023. First Published Online in J-STAGE on April 27, 2023. *Corresponding author (e-mail: takahashi0126@affrc.go.jp) mature seeds is suppressed. For example, in temperate zones with cold winter, summer annuals produce seeds in autumn, but the seeds enter dormancy and do not germinate until spring. However, winter annuals produce seeds in the spring, which remain dormant during summer before germinating in the fall. Dormancy break occurs owing to the loss of seed coat water impermeability or due to environmental signals such as low or high temperatures (Baskin and Baskin 2004). The seeds of several wild legumes exhibit seed coat water impermeability when they reach maturity; this physical dormancy is the result of controlling the absorption of water into the seed. As time passes, a specialized water gap tissue opens up to allow the absorption of water into the seed (Baskin and Baskin 2004) (Fig. 1). Seed dormancy is an important trait that allows the germination of seeds at the right time in their natural environments, but it is an inconvenient trait in crops cultivated and harvested by farmers. However, it is believed that the early farmers did not knowingly select non-dormancy (Purugganan 2019). For example, if a non-dormant mutant occurs in a population through cultivation, it would germinate and be harvested preferentially each year. Therefore, if a population that includes a few non-dormant mutant individuals is cultivated a dozen or so times, it would naturally be dominated by the descendants of non-dormant mutants regardless of whether the cultivator is aware (Ladizinsky 1987). Ladizinsky (1987) proposed a theory that legume crop domestication differed significantly from cereal crop domestication. He argued that lentils had already acquired non-dormancy before they were cultivated, when their seeds et al. 2002). Additionally, there are many examples of conscious selection of larger seeds owing to factors such as appearance and mouthfeel.

Archeology
Wild ancestors of domesticated azuki bean are found in Japan, southern China, Korea, northern Laos, northern Myanmar, Bhutan, Nepal, and the Indian Himalayas; wild ancestors of domesticated soybean are found in Japan, China, Korea, Taiwan, and the Russian Far East (Lu 2005, Ohashi 2001). Wild azuki bean and wild soybean are characteristic of dormant seeds, shattering pods, and small seeds (Figs. 1, 2).
While the previous leading theory states that both azuki bean and soybean were domesticated in China, recent findings suggest that wild soybean and wild azuki bean have been domesticated and cultivated in Japan for over 10,000 years. Remains of legume seeds excavated from archeological sites occur as carbonized seeds or seed impressions on pottery. Carbonized seeds refer to the carbonized remains of seeds that have been heated, for example, in a house hearth (Fig. 3A, 3B). Seed impressions on pottery were created when the clay used to make the pot contained seeds; the seeds burned away when the pots were fired, leaving impressions on the outer or inner surfaces of the pots. Casts of seed impressions on pottery are made using a silicone compound to study their morphology and surface structure (Fig. 3C, 3D, Ushino and Tagawa 1991). Pods and seeds of wild and domesticated azuki bean and soybean. Upon maturation, (A) wild azuki bean and (B) wild soybean pod valves twist into a helical shape, causing the seeds to scatter. (C) Domesticated azuki bean with the acquired non-shattering pods with pod valves that do not twist. (D) The non-shattering trait of domesticated soybean is inefficient; the helical tension of pod valves remains strong when they dry, making them more likely to shatter than the pods of domesticated azuki bean.
While seeds shrink when they are burned, the impressions sometimes indicate enlarged seeds owing to the absorption of water during pot making or dried before firing (Nakayama 2009, Nasu et al. 2015b, Obata 2011. Therefore, the seed impressions on pottery do not precisely indicate seed size. For example, in a firing experiment investigating wild soybean seed impressions on pottery, there was no difference between the lengths of the original seeds and those of their impressions, whereas in the case of wild azuki bean seeds, the length of the seed impressions was 112% of the original seeds (Nasu et al. 2015b). In contrast, with domesticated soybean, the length of the seed impressions was 133% of the original seeds, and with domesticated azuki bean, the length of the seed impressions was 107% of the original seeds. These results indicated that the changes in the size of seed impressions depend on the species (azuki bean or soybean) as well as on whether the seeds exhibit dormancy (wild or domesticated). Additionally, because of the possibility that immature seeds and seeds that have already absorbed water have been embedded in the pots, the figures in this section illustrating the lengths of carbonized seeds and seed impressions on pottery were created using uncorrected data (Fig. 4). Presently, the only domestication trait that can be inferred from carbonized seeds and seed impressions on pottery is the seed size. However, once the genes responsible for non-dormant seeds and nonshattering pods are identified, it would be possible to infer such traits from the carbonized seeds.  Table 2). (C) Replicas of azuki bean impressions on pottery created by injecting silicone into the impression (Sakenomiba site, 5180 cal. BP, site 45 in Fig. 5 and Table 1). (D) Replicas of soybean impressions on pottery created by injecting silicone into the impression (Sakenomiba site, 5040 cal. BP, site 45 in Fig. 5 and Table 2).
Data on the names and ages of the archeological sites where azuki bean and soybean seed sizes (lengths) were recorded are shown in Tables 1 and 2, respectively. Based on this data, the relationships between the ages of the archeological sites (median of estimated range) and mean sizes (lengths) of the excavated seeds are shown in Fig. 4, the names and locations of sites in Japan where azuki bean and soybean seeds were excavated are shown in Fig. 5, the sites in Japan where azuki bean and soybean seeds were excavated are shown by age in maps in Fig. 6, and the archeological sites in China and Korea where azuki bean and soybean seeds were excavated are shown by age in maps in Fig. 7. For the azuki bean, the figure was created using 62 data points from Japan (34 carbonized seeds, 28 replica seeds), 1 from China, and 4 from Korea. For soybean, the figure was created using 48 data points from Japan (19 carbonized seeds, 29 replica seeds), 17 from China, and 4 from Korea. The oldest excavated azuki bean and soybean were both from sites in Japan. The diversification of seed sizes that accompanied seed size increase first occurred in Japan. In particular, a large number of replicas of larger azuki bean and soybean seeds were obtained from impressions of pottery from sites in the Central Highlands of Japan, where many sites of Jomon period appeared in the period from 6,000 to 4,000 BC.

Archeology of azuki bean
Detailed archeological surveys carried out recently have discovered many carbonized seeds and pottery impressions of azuki bean in Japan (Table 1). The oldest azuki bean sample in Japan, a 3-4 mm carbonized seed dating back to approximately 10,000 years, was excavated from the Awazu lakebed in Shiga Prefecture (Table 1, site 55 in Fig. 5). As the carbonized seeds appear to be approximately of the length as that of the modern wild azuki bean, it is believed that prehistoric people who lived there used wild azuki bean (Table 1, Fig. 4, Minaki and Nakagawa 2000). The high number of archeological records in the period from approximately 6,000 to 4,000 years ago indicates a more intensive use of azuki bean during that period (Figs. 4A, 6B). Particularly, azuki bean seeds have been excavated from many sites in the Central Highlands (Koshin region: Yamanashi and Nagano prefectures) in Japan (Fig. 6B). Increased seed size was observed in both carbonized seeds and seed impressions on pottery during this period (Fig. 4A). Several seed impression data depicting mean seed lengths indicate a seed size (up to 7.00 mm, Table 1) larger than those of modern wild species after absorbing water (5.37 mm, Nasu et al. 2015b).
Interestingly, archeological records show that the use of azuki bean declined dramatically in the Central Highlands around 4,000 to 3,000 years ago (Fig. 6C). Instead, large carbonized azuki bean seeds and seed impressions on pottery were found in western Japan and Kyushu Island around this time (Table 1, Figs. 4A, 6C). Although the reason for this remains unclear, it has been suggested that a period of rapid cooling around 4,200-3,800 years ago (Mayewski et al. 2004) led to the decrease in chestnut and Takahashi, Nasu, Nakayama and Tomooka legume harvests in the Central Highlands, causing humans to migrate and spread the culture of legume cultivation to Kyushu Island located in southern Japan (Nasu 2018, Obata 2016. It is an interesting question whether the large beans that appeared in Kyushu Island around this time had come from the Central Highlands or were brought from Korea or China. Archeological records show the widespread use of azuki bean in Japan around 3,000 years ago (Fig. 6D). Moreover, rice and millet farming were introduced into Japan from China during this period (Nasu and Momohara 2016). Another interesting question is whether the widespread use of azuki bean in Japan (excluding Hokkaido Island) was related to the spread of rice and millet cultivation introduced from China. Apart from the report by Lee (2013), there are fewer records of azuki bean in Korea than those found in Japan. Presently, the oldest record of azuki bean in Korea include carbonized seeds from the Middle Chulmun period (about Fig. 6. Historical changes in the geographical distribution sites in Japan where azuki bean and soybean findings (carbonized seeds or seed impressions on pottery) were excavated. (A) 13,000-6,000 years ago: no excavations from the Kanto region to the Tohoku region (northern part of Japan). (B) 6,000-4,000 years ago: during a period when the climate was relatively mild, the number of sites in the Central Highlands (Yamanashi and Nagano prefectures) increased sharply, with azuki bean being found even in the northern Tohoku region. (C) 4,000-3,000 years ago: changes in vegetation owing to a rapid cooling event occurring approximately 4,200 years ago have temporarily reduced the quantity of legumes in the Central Highlands and Tohoku region. (D) 3,000-100 years ago: around 3,000 years ago, rice and grain (millet) farming were introduced into Japan from China. Azuki bean and soybean seeds were found in sites throughout Japan excluding Hokkaido Island. The spread of agriculture in the continent could be related to the distribution of azuki bean and soybean throughout Japan.  Fig. 7B). The mean length of these seeds is 3.2 mm (max 4.1 mm), which indicates that the size of azuki bean seeds in Korea around that time was still close to the size of modern wild azuki bean. Increase in seed size appears to have occurred during the Middle Mumun period, approximately 2,600 years ago. The mean seed size during this period was still small (3.8 mm), but some seeds were found to be over 7 mm (Table 1, site 88 in Fig. 7D). After this period, a gradual increase in seed size to a mean size of 4.6 mm has been observed during the Three Kingdoms period (about 1,700 years ago) and 5.4 mm during the Joseon period (approximately 500 years ago) (Table 1, Fig. 4A, site 87 in Fig. 7D). The increase in the size of azuki bean seeds occurred later in Korea than in Japan. Rice and millet farming were introduced into Japan from Korea via northern Kyushu Island around 3,000 years ago. The genetic relationship between the large azuki bean seeds excavated from Korea and those excavated from Japan would be an interesting topic to explore in future studies.
Although there are records of excavated azuki bean in China, there is only one report on their seed size (Table 1, Fig. 4A, site 86 in Fig. 7C). Starch grains from the seeds of species belonging to the tribe Phaseoleae found at the Shizitan site in Shaanxi Province (approximately 13,800-8,500 cal. BP) are believed to be those of the genus Vigna (Liu et al. 2013). Carbonized azuki bean seeds dating back to approximately 4,000 years were found at the Lianchengzhen site in Shandong Province (4,060-3,840 cal. BP) ( Table 1, site 86 in Fig. 7C) (Crawford et al. 2005). The maximum seed length was 4.2 mm, which is about the same as that of modern wild azuki bean seeds (Fig. 4A). These findings suggested that azuki bean was not as commonly used in China as in Japan or Korea.

Archeology of soybean
The oldest soybean reported in Japan is a seed impression on pottery from the Ojiyama site in Miyazaki Prefecture dating back to 13,000 years (Table 2, site 73 in Fig. 5). The seed is 3.8 mm long, equivalent to the size of the modern wild ancestor of soybean (Table 2, Fig. 4B, Obata and Manabe 2012). Reliable evidence collected from approximately 9,000-year-old sites in the Central Highlands (Nagano and Yamanashi prefectures) indicated 6.8-7.6-mm long seed impressions and 4-mm long carbonized seeds discovered in 2009 and 2013, respectively (Table 2, Fig. 4B, sites 27 & 35 in Fig. 5). Similar to the azuki bean, archeological evidence showed the presence of soybean at many archeological sites beyond 6,000 years, mainly in the Central Highlands; some seed impressions on pottery were found to be longer than those of the modern wild ancestors after absorbing water (10 mm) (Fig. 4B). Around this time, seeds with approximately 2-10 times larger volume than that of their wild ancestors became prominent; different seed shapes were also observed, ranging from flat to thicker oval types, showing increased morphological diversification (Nakayama 2015). However, similar to that of the azuki bean, archeological evidence of soybean in the Central Highlands decreased dramatically around 4,000-3,000 years ago, and larger soybean seeds started to appear in Kyushu Island (Fig. 6C). Around 3,000 years ago when rice and millet cultivation was introduced into Japan from China, soybean seeds were found throughout Japan (Fig. 6D).
In China, soybean seeds from the Neolithic period were found in excavations (Lee et al. 2011); carbonized seeds BS Breeding Science Vol. 73 No. 2 Takahashi, Nasu, Nakayama and Tomooka from approximately 9,000 years ago were found at the Jiahu site in Henan Province. These seeds were small, measuring an average 3.1 mm length (Table 2, Fig. 4B, site 75 in Fig. 7A). Soybean seeds dating back to 6,000-4,000 years ago were mainly found in the Yellow River Basin (Fig. 7B). During the Longshan cultural period around 4,500 years ago, mean soybean seed length increased to 4.5 mm; during the Shang Dynasty around 3,000 years ago, the mean seed length from site 81 was >5 mm (Table 2). Increasing soybean seed size was observed in China but not as early as that observed in Japan (Fig. 4B). Carbonized seeds from the Late Chulmun period, i.e., approximately 4,800 years ago, were found at the Pyeonggeodong site in Korea (Table 2, site 89 in Fig. 7B, Lee et al. 2011). The mean seed length was 3.2 mm, which is similar to that of the modern wild ancestor (Table 2,  Fig. 4B). The mean seed size doubled to approximately 6 mm in the Early Mumun period around 3,500 years ago ( Table 2, Fig. 4B, site 90 in Fig. 7C). By the Middle Mumun period around 2,600 years ago, the average seed size exceeded 7 mm (Table 2, site 90 in Fig. 7D), showing a faster rate of seed size increase than that in China (Fig. 4B).
As shown in the examples described above, East Asian archeological records indicate that azuki bean and soybean seed size increased in Japan earlier than in Korea and China. This refutes the conventional theory that azuki bean and soybean have a single origin in China; it could be said that they originated in Japan or have multiple origins, with an earlier origin in Japan. Moreover, the widespread use of azuki bean and soybean in Japan approximately 3,000 years ago indicates the possibility that additional azuki bean and soybean cultivars were introduced from China along with rice and millet. Alternatively, it could be said that only rice and millet were introduced from China, whereas azuki bean and soybean were introduced into Korea and China from Japan. The answers for the above issues could be obtained by integrating archeology and DNA analysis.

Molecular phylogeny of azuki bean
A chloroplast DNA study placed Chinese, Korean, and Japanese domesticated azuki bean in the same group as Korean and Japanese wild azuki bean, whereas wild azuki bean from Nepal, Bhutan, Myanmar, and China were placed in different groups (Ye and Yamaguchi 2008). This suggested that the azuki bean was domesticated in Japan or Korea. Additionally, it has been shown that wild Japanese azuki bean possesses most of the simple sequence repeat alleles of domesticated azuki bean, including those from Bhutan and China (Xu et al. 2008). However, their analysis did not include wild Korean azuki bean and included only five samples of wild Chinese azuki bean. Tomooka et al. (2014) suggested that Japan is the center of domestication for azuki bean based on the results of DNA analysis (Xu et al. 2008, Yamaguchi 1992 as well as on the facts of oldest archeological remains found in Japan. Molecular phylogenetic studies based on resequencing of large-sized samples are yet to be conducted on the azuki bean.

Molecular phylogeny of soybean
Recent advances in sequencing technology have led to the progress in research on the origins of soybean domestication. Early research on domestication using morphological and seed protein variations suggested that soybean originated in China (Broich and Palmer 1981, Hymowitz 1990, Hymowitz and Kaizuma 1981. Later, a theory of multiple origins in China and Japan emerged based on mutations in chloroplast DNA (Xu et al. 2002). The availability of several DNA markers once again supported the single origin theory from China (Guo et al. 2010. After the release of the complete soybean genome (Schmutz et al. 2010), molecular phylogenetic analyses based on resequencing supported the single origin theory (Chung et al. 2014, Lam et al. 2010. Wang et al. (2016), who proposed the northern and central Chinese origin theory using high-density single nucleotide polymorphism (SNP) arrays, only analyzed wild and domesticated Chinese soybean accessions. Therefore, their study only investigated regions within China as candidates for the origin for those accessions. In contrast, Jeong et al. (2019) analyzed about 4,400 wild and domesticated soybean accessions from Japan, South Korea, North Korea (only domesticated accessions), Far East Russia (only wild accessions), and China using SNP arrays. This study is the most comprehensive investigation to date, and the results suggest that soybean was most likely domesticated only once in eastern Japan.

Domestication genes
Evolution and domestication are both driven by spontaneous mutations, although the mutations necessary for the acquisition of domestication traits tend to be simpler than those necessary to acquire other evolutionary traits (Meyer and Purugganan 2013). Spontaneous mutations include large structural changes such as chromosome-level insertions, deletions, duplications, inversions, translocations, and reduplications, and smaller mutations include genelevel base substitutions, insertions, and deletions (Crow 1983). Organisms accumulate a variety of spontaneous mutations over long periods of time, where genes and gene networks with new functions are constructed, creating a diversity of living beings. In contrast, domestication was achieved within a span of 12,000 years since humans began farming, which is a relatively short period (Meyer and Purugganan 2013). Recently, genes controlling certain domestication traits in major crops have been discovered. Such domestication genes are often transcription factors that regulate genes, and nonsense mutations that impair certain gene functions are often the primary factors involved in the development of domestication traits (Meyer and Purugganan 2013).

Domestication genes of azuki bean
The mechanisms through which azuki bean acquired certain domestication traits differ from those of domesticated soybean. Although domesticated soybean with yellow (translucent) seed coat absorbs water throughout its structure, domesticated azuki bean with red seed coat absorbs water only through a specialized tissue known as the lens ( Fig. 1) (Kikuchi et al. 2006). The differences in mechanisms of absorption of soybean and azuki bean seeds indicated that the water absorption rate of domesticated azuki bean is much slower than that of the domesticated soybean (Uenaka et al. 2000). Among the five loci detected for seed dormancy, one of them was tightly linked to seed coat color (red or tan) in azuki bean (Isemura et al. 2007). However, genes responsible for seed dormancy of azuki bean are yet to be identified. Although the gene VaSDC1 encoding a putative R2R3-type MYB transcription factor was identified as the candidate gene for seed coat color (red or black), it has not been associated with seed dormancy (Chu et al. 2021).
The non-shattering pod phenotype of domesticated azuki bean is superior to that of domesticated soybean. Although wild soybean and azuki bean display the same pod shattering mechanism, i.e., through the helical tension generated on the dried pod ( Fig. 2A, 2B), domesticated soybean and azuki bean acquired non-shattering pod traits through different mechanisms. In the case of azuki bean, a phenotypic variation that greatly reduces the helical tension of the pod valves by thinning the lignin layer was selected (Fig. 8A,  8B), and the adhesion of the seam that connects the pod valves did not change between wild and domesticated azuki beans (Fig. 8E, 8F) (Takahashi et al. 2020). In contrast, wild and domesticated soybean did not vary in terms of thickness of lignin layer (Fig. 8C, 8D), and the adhesion of the seam that connects the pod valves did change (Fig. 8G,  8H). The non-shattering pod phenotype of domesticated azuki bean is controlled by a single locus (Isemura et al. 2007). Fine-mapping results suggest that a loss of function mutation of the VaMYB26 (Vigan.07g034400) encoding a putative R2R3-type MYB transcription factor is responsible for the thinning of the lignin layer in the pods of domesticated azuki bean (Takahashi et al. 2020). The putative MYB26 orthologs have also been identified as candidate genes for non-shattering pods in common bean (Di Vittori et al. 2021).
The mechanisms involved in seed size increase are yet to be clarified in azuki bean. Although six loci involved in seed size differences have been identified between wild and domesticated azuki beans (Isemura et al. 2007), the candidate gene is yet to be identified. Future research should focus on the molecular mechanisms involved in seed size increase in azuki bean.

Domestication genes of soybean
Eight quantitative trait loci (QTL) for seed dormancy were detected between wild and domesticated soybeans (Kuroda et al. 2013). Two genes responsible for the control Takahashi, Nasu, Nakayama and Tomooka of physical seed dormancy of soybean have been identified to date: (i) GmHs1-1 (Glyma.02G269500), which encodes a calcineurin-like transmembrane metallophosphoesterase involved in the regulation of calcium content of seed coat (Sun et al. 2015); and (ii) GmGH9B8 (Glyma.02G269400), encoding an endo-1,4-β-glucanase involved in the development of cracks on the seed coat surface (Jang et al. 2015). Unlike the two aforementioned genes, the G gene (Glyma.01g198500), encoding a CAAX amino-terminal protease protein, affects physiological seed dormancy through interactions with key enzymes that modulate abscisic acid synthesis (Wang et al. 2018). The G gene also determines differences in seed coat color (yellow or green) among varieties (Wang et al. 2018). Additionally, two QTLs responsible for water permeability of the seed coat are tightly linked to the I and T loci for seed coat color (Kuroda et al. 2013). These findings suggest that part of the signaling molecules and secondary metabolites that determine seed coat color may affect physiological as well as physical seed dormancy.
A genome-wide association study on soybean pod shattering identified loci harboring SHAT1-5, Pdh1, and NST1A as loci harboring genes (Zhang and Singh 2020). SHAT1-5 (Glyma.16G019400) encoding a NAC domain transcription factor was selected during soybean domestication (Dong et al. 2014). A gain of function mutation of the SHAT1-5 makes pods less likely to shatter, because the adhesion of the seam that connects the pod valves is reinforced by excessive lignification, which resists the helical tension of pod valves (Fig. 8G, 8H). A severe selective sweep was detected across the ~116 kb region including SHAT1-5 on chromosome 16, indicating a strong artificial selection on this gene. The PDH1 (Glyma16g25580) encoding a dirigent-like protein was found to be associated with controlling the helical tension of the pod valves through modification of the inner sclerenchyma cells (Funatsuki et al. 2014). Although PDH1 was initially identified based on differences among cultivars, its non-shattering allele was found in wild soybean from the Huang-Huai-Hai valleys of China, suggesting that its alleles may have originated owing to natural variation in this region (Zhang and Singh 2020). Recently, a gain of function allele of NST1A (Glyma.07g050600), a paralog of SHAT1-5, was reported as a possible contributor to the non-shattering pod phenotype (Zhang and Singh 2020); however, the SNP detected by genome-wide association analysis is 49 kb away from the other SNP that gives rise to a premature stop codon in NST1A; therefore, additional analysis including transformation are required.
Nineteen QTLs have been detected for seed size differences between wild and domesticated soybean, and two genes have been identified to date (Karikari et al. 2019). PP2C (Glyma17g33690), which encodes a type-2C protein phosphatase, contributes to seed size increase by enhancing the cell size of the integument and activating a subset of seed trait-related genes (Lu et al. 2017). The ln (Glyma20g25000), which encodes a zinc finger protein JAGGED, has a significant effect on the number of seeds per pod and seed size (Jeong et al. 2012). Owing to the involvement of several loci and their associated effects, albeit small, there is no clarity on how soybean seeds enlarged during domestication.

Need for DNA archeology
Studies focusing on the domestication genes of azuki bean and soybean are yet to be conducted using DNA from archeological findings. Yano et al. (2004) extracted DNA from carbonized legume seed remains dating back to approximately 4,660 years from the Sannai-Maruyama site in Aomori Prefecture, Japan. Based on chloroplast DNA markers, the specimen were identified as azuki bean seeds. If the genes that control domestication traits could be sequenced from carbonized azuki bean and soybean seeds, it should be possible to determine the degree of domestication of the individual plants, for example, whether they had seed dormancy or pod shattering traits. In conclusion, there is a need for close cooperation between archeologists and molecular geneticists to find answers to these questions.

Author Contribution Statement
YT conceptualized this review. YT, HN, SN and TN wrote the manuscript.