Soybean [Glycine max (L.) Merrill] is apparently one of the most important cultivated crops worldwide in its agro-economic value and diverse utilities in both agriculture and industry. The legume family, which contains this crop, is composed of approximately 20000 species, which is the third largest group among flowering plants, and includes other agriculturally important legume crops such as common bean (Phaseolus vulgaris), mung bean (Vigna radiata) and pigeon pea (Cajanus cajan). Among many other evolutionary branches within the family, Phaseoloid clade harbors most of important crop legumes of agricultural importance, within which soybean is a member of this clade (Choi and Cook 2011).

It is generally known that distribution of the wild soybean (G. soja), which is the ancestor of current cultivated soybean, is limited to the East Asia regions including China, Korea and Japan. Historical records have addressed that the first cultivation of soybean originated in China, which was approximately 4500 years ago (Qiu and Chang 2010). “Shu”, which is ancient Chinese character meaning soybean, has been frequently found in ancient Chinese books. In addition, carbonized remain of soybean seeds, which was estimated to be 2600 years old, were discovered in an excavation site of the Eastern Zhou Dynasty (Qiu and Chang 2010). Although its cultivation history dates back to ancient age, soybean actually came to prominent crop during the last 200 years (Singh and Shivakumar 2010). Because of such a long history of cultivation and natural/artificial selection, which was the natural breeding process itself, China had been the world’s top producer until the first half of the 20^th century. However, the situation had been reversed by the USA in the 1950s. It is known that soybean was first cultivated in the USA as early as 1765 (Hymowitz 1984). During 1927?1931, the USA sent scientist to collect soybean germplasm in China, Korea and Japan, and some of germplasm played a pivotal roles as primary parents to breed the current USA cultivars (Qiu and Chang 2010). Using those collected accessions, the USA rapidly developed breeding program and accelerated the production of soybean, and thereafter the country has now become not only the highest producer but also largest exporter all across the world. In the USA, soybean is now the second largest crop, right after corn, in production.

Korea also has a significantly long history of domestication and cultivation of soybean, which is comparable to that of China and dates back to the ancient Chulmun period (8000-1500 BC). A recent archeological study shows that charred soybean seeds discovered in ‘Pyeonggeodong’ around the Nam River valley are as old as 4840-4650 years (Lee et al. 2011). Thereafter soybean continued to remain an important crop in Korea throughout the ancient Mumun period followed by the Three Kingdom periods. Interestingly, the study also addresses that soybean domestication occurred independently in multiple regions of the East Asia and landraces with larger seeds were adapted in Korea and Japan far earlier than in China (Lee et al. 2011). Although Korea is one of the East Asian countries of soybean origin and cultivation, start of systematic breeding programs were relatively delayed. The first Korean cultivar bred by hybridization method ‘Kwangkyo’ was recently developed in 1969. Soybean is still important as one of main crops in Korea, which is the third after rice and wheat, and its production in1975 could almost fill up 98% of domestic need. Since then, the rate of self-sustenance has gradually decreased and is now only 8.7%, thereby causing severe dependence on the GMO soybean imported from the USA.

Soybean is cultivated in enormous area of arable land worldwide, which accounts for 90.2 million ha, resulting in a total production of about 276 million ton (MT) in 2013 (FAO 2013). Among major producing countries, the USA is the top producer (28.2% and 32.2%), by both area and production, followed by Brazil (23.7% and 27.5%), Argentina (18.5% and 21.2%) and China (9.7% and 7.0%), respectively (Soytech Inc. 2007). In comparison to the total production of the world, Korea produced relatively small amount of soybean (0.16MT) in 2013 (FAO 2013).

Presumably, the most prominent purposes for soybean cultivation should be its high contents of protein and oil, which make up approximately 40% and 20%, respectively. In the Western world of North America and Europe, soybean is mainly regarded as oil crop or protein source for animal feed. In other countries of the Eastern World, this crop is consumed mainly by human in various forms of foods, such as bean curd (tofu), soy milk, sprout, soy paste/sauce. Soybean also has a diverse array of utilities in industry and has been used for the production of lubricants, toner ink, cosmetics and for many other purposes. In more recent years, this crop is increasingly drawing interest of scientists as a useful biofuel source.

Although soybean production in Korea has gradually decreased during the past 40 years, its importance as a major crop has not diminished probably because soybean has a long history of cultivation associated intimately with traditional food culture of Korea. Recent completion of soybean whole genome sequencing (Schmutz et al. 2010) may promote the development of new technologies for soybean molecular breeding. Moreover, recently developed technology of the next generation sequencing (NGS) and a variety of bioinformatic tools can play a critical role in processing and analyzing biological big data. NGS-driven bioinformatic data processing enables researchers to analyze population level of resequencing and transcriptome data. Hence, collection and organization of useful germplasm and their phenotypic characterization in precision are becoming more and more important. To effectively select germplasm for the data analyses, it is essential to know relationships among a variety of cultivars, landraces and wild accessions. Towards this end, this review intends to integrate all available data and information, and provide a pedigree-based view of Korean soybean varieties, with a high expectation for making a constructive use of the pedigree information to breed a diverse array of soybean varieties with superior traits in the future.

A brief history of soybean breeding in Korea

According to a recent archeological study, it has been known that the cultivation and domestication of soybean in Korea was evidenced back in ancient era of the Chulmun period, which was approximately 4000 year ago. In spite of such a long history of cultivation, a memorable moment towards the development of cultivation methods and systematic crop improvement initiated officially with the first cultivar ‘Jangdanbaekmok’, which was derived from landrace through pure line selection breeding, and was recommended for farmers to grow and cultivated nationwide until the first hybridization-bred variety ‘Kwangkyo’ was developed. This cultivar was grown and used for soy sauce/paste and beared medium-to-large seeds. Before cross breeding technology was actively used, most varieties were developed by selecting pure lines or directly introduced from other countries such as the USA and Japan. At the early developmental stage of new cultivars, breeding purposes mainly focused on improvement of qualities associated with yield and easiness in cultivation. In 1960s, 6 Korean landraces (‘Haman’, ‘Chungbukbaek’, ‘Keumgangdaelip’, ‘Buseok’, ‘Iksan’, including ‘Jangdanbaekmok’) and ‘Yukwoo3’, which was introduced from Japan, were released for the purpose of cultivation. Thereafter, two cultivars, ‘Hill’ and ‘Shelby’, were introduced from the USA in 1967 and 1968, respectively. In the next year, ‘Kwangkyo’, which was the first variety developed by hybridization-based breeding, was released to farmers. ‘Kwangkyo’ was developed using ‘Jangdanbaekmok’ as maternal line and ‘Yukwoo3’ as paternal line, and thereafter was frequently employed to develop many other cross-bred varieties. In 1970s, both Korea-bred and introduced varieties were supplied for the cultivation in parallel manner, and breeding came to further focusing on development of cultivars adaptive to mechanized farming because of a social phenomenon of the rural exodus. Influenced by such change in social environment, ‘Muhankong’ (Hong et al. 1989), ‘Jangkyongkong’ (1988), ‘Jangsukong’ (Hong et al. 1991), which were resistant to lodging and indeterminate type with relatively longer stem length, were developed for mechanized farming and released in 1980s. In 1986, ‘Eunhakong’ (Shin et al. 1988), which was first bred for bean sprout, was released. Until 1990s, soybean breeding mainly focused on qualities in appearance, stress resistance and useful chemical ingredients in seeds. Since 1995, soybean breeding has gradually switched to upgrading crop qualities and diversifying uses to compete cheaper GMO soybeans imported from open international market of the agricultural products. As a result, many varieties with desirable qualities were developed and released, some of which included ‘Danbaek’ (Soybean Breeding Team 1994a), ‘Kwangankong’ (Soybean Breeding Team 1994b) with higher protein content, ‘Jinpumkong’ (Kim et al. 1995) with no fishy smell and ‘Geomjeongkong 1’ (Soybean Breeding Team 1994c) suitable for cooking with rice. After 2000s, soybean breeding has more focused on the improvement of quality for processing, functional ingredients, seed appearance and mechanized farming.

Previous pedigree studies of soybean varieties in Korea

It appears that pedigree analyses for Korean soybean varieties were intermittently made by a limited number of breeding researchers. Jong et al. (1999) conducted CP (coefficient of parentage)-based clustering analysis for 75 soybean varieties released during 1913?1998. The analysis ended up with separation of the varieties into nine clusters. In the next year, the same group of researchers performed a similar pedigree analysis for 53 varieties bred by hybridization method and released during 1960?1998 (Jong et al. 2000). It is currently thought that the pedigree study in 2006 was the most recent one done by the same research team (Jong et al. 2006). In that study, they analyzed pedigree using a total of 86 cultivars bred between 1933 and 2002. They also attempted to compare genetic relationships among cultivars by pedigree analysis and DNA finger printing method (Jong et al. 2006). However, any other noticeable research of pedigree analysis for Korean soybean varieties has not been found, although considerable amount of time has elapsed since then. Thus we intended to update and reconstruct the pedigrees by integrating data and/or information on recently released soybean varieties in this review.

Data/information resources and current status of soybean breeding in Korea

In an attempt to reconstruct the Korean soybean pedigree, previous analyses done by Jong et al. (1999, 2000, 2006) played a skeletal role in rebuilding the pedigree. The majority of relevant information was obtained from the RDA-Genebank Information Center (GBIC; http://www.genebank.go.kr) and the Korea Seed & Variety Service (KSVS; http://www.seed.go.kr). Basic information on all registered cultivars/varieties is provided in Table 1 and Table 2.

The data reveals that a total of 178 soybean varieties have currently been registered at the two national institutes. Fig. 1 demonstrates their basic information for all these varieties according to their uses, breeding methods and timelines at which they were developed. It is apparent that soybean varieties for soy sauce and tofu predominate in Korea while ones for vegetables are relatively minor (Fig. 1A). Of 178 registered accessions, the majority of varieties (155, 87.1%) have been developed by hybridization-based breeding methods (Fig. 1B). Since the first hybridization-bred Korean variety ‘Kwangkyo’ was released, seven more varieties were added in 1970s as recommended cultivars. These include ‘Bong-Eui’, ‘Kang-lim’, ‘Baegcheon’ (Choi et al. 1978), ‘Jangyeobkong’, which were developed by hybridization method and three introduced varieties, ‘Eundaedu’, ‘Dongpuk-tae’, ‘Danyeobkong (Essex)’. It is noteworthy that new varieties have been increasingly developed almost exclusively by hybridization methods and registered since 1980s (Fig. 1C). It is also noticeable that more than one third of varieties (68 cultivars, 38.2%) were emerged during the period of 2000s, which implicated that previously developed varieties played nodal roles to provide a diverse array of progenitors for soybean breeding.

Integration and reconstruction of soybean pedigree

Starting from ‘Jangdanbaekmok’ the first cultivar registered in 1913 till now, we attempted to reconstruct a fully integrated pedigree including all the 178 varieties currently recorded in Korea. Thereby, it has ended up with a total of four pedigrees (Fig. 2). Of 178 registered cultivars, 10 varieties (5.6%) could not be integrated into any pedigrees because they lacked information on which they were employed as elite parental lines to cross. They include ‘Chungbukbaek’ (1948), ‘Iksan’ (1948), ‘Haman’ (1960), ‘Keumgangsorip’ (1960), ‘Buseok’ (1948), ‘Heugcheongkong’ (1999), ‘Galmikong’ (1999), which were derived from Korean landraces, and introduced accessions, such as ‘Shelby’ (1967), ‘Eundaedu’ (1970), ‘Hwasongputkong’ (1993). Of 168 varieties included in the pedigree network, 136 varieties (76.4%) and 32 varieties were involved in one or two pedigrees, respectively.

As seen in the Pedigree I (Fig. 2A), ‘Kwangkyo’, ‘Hwangkeumkong’, ‘Paldalkong’ and ‘Sinpaldalkong 2’ (Kim et al. 1994) served as central crossing parents, which is the largest among four pedigrees and resulted from the integration of three previously reported pedigrees (Jong et al. 2006), Cluster 7 (‘Kwangkyo’ group), Cluster 8 (‘Hwangkeumkong’ group), Cluster 9 (‘Paldal’ group). This pedigree contains 85 hybridization-bred varieties, 49 breeding lines, 7 landraces and 24 introduced varieties (Fig. 2A). Nodal breeding points of this pedigree start with ‘Jangdanbaekmok’ (the first landrace-derived cultivar), ‘Kwangkyo’ (the first hybridization bred variety) and three Japan-introduced accessions such as ‘Yukwoo3’, ‘Baekmokjangyeob’, ‘Dongpuk-tae’.

‘Kwangkyo’ was intensively employed as parental line to breed other useful progenitors such as ‘Hwangkeumkong’ (1980), ‘Milyangkong’ (1983), ‘Baegunkong’ (1984), ‘Saealkong’ (Shin et al. 1985), ‘Dankyeongkong’ (1986), ‘Jangkyongkong’ (1988). Of ‘Kwangkyo’-derived progeny, ‘Hwangkeumkong’ and ‘Paldalkong’ were recognized by their superior traits and frequently used as parental line. For one example, ‘Hwangkeumkong’ was developed by using USA-introduced ‘Clark63’ as the parental line and another introduced line ‘Baekmokjangyeob’ (SMV-resistant) as maternal line, and thereby resulted in a combined trait of SMV resistance as well as large seed size (RDA 2012a). Likewise, ‘Hwangkeum’ was employed as crossing parents to develop following varieties; ‘Muhankong’ (1988), ‘Geomjeongkong 1’ (1993), ‘Kwangankong’ (1993), ‘Alchankong’ (Kim et al. 1997), ‘Seonheukkong’ (1998), ‘Jinyulkong’ (Yun et al. 2000), ‘Hojang’ (2002), ‘Cheonsang’ (Kim et al. 2012). ‘Paldalkong’ was bred using USA-introduced ‘Elf’ and a ‘Kwangkyo’-derived breeding line ‘SS74185’. ‘Paldal’ is highly tolerant not only to the lodging but also to major biotic stresses with small-to-medium seed size (RDA 2008a), and was employed to develop following varieties; ‘Sinpaldalkong’ (1991), ‘Sinpaldalkong 2’ (1992), ‘Dajangkong’ (Shin et al. 1997), ‘Tawonkong’ (Kim et al. 1996), ‘Ilmikong’ (Shin et al. 1998), ‘Saeolkong’ (Baek et al. 1998), ‘Geomjeong4’ (2001).

‘Sinpaldalkong 2’, which occupies a significant part of the Pedigree I and is strongly resistant to soybean mosaic viruses, was developed by employing ‘Togyu’ as maternal line and ‘Paldal’-derived breeding line ‘SS79186’ as paternal line. It bears medium-sized seeds and is known as tolerant or resistant to many diseases such as soybean mosaic virus, purpura, downy mildew and necrosis. Thereafter, ‘Sinpaldal2’ was used as breeding parent to develop following varieties; ‘Daepung’ (Park et al. 2005), ‘Shingi’ (2003), ‘Cheongdu1’ (Yun et al. 2005a), ‘Daemang #2’ (Yun et al. 2006), ‘Cheonga’ (Ha et al. 2013), ‘Geomjeong5’ (Han et al. 2013), ‘Uram’ (2010), ‘Chamol’ (2011). One of ‘Sinpaldal2’-derived progeny ‘Daepung’ was bred using ‘Baegun’ as maternal line and its yield was 305 Kg/10a, which was 20% higher compared to ‘Taekwangkong’. ‘Daepung’ was further improved towards higher yield, and resulted in ‘Daepung2ho’, which is currently a cultivar of the highest yield (345 Kg/10a) in Korea. This cultivar is also tolerant to lodging, fire blight and seed shattering. ‘Uram’ is another high yield cultivar (327 Kg/10a), and has merits of suitability to mechanized farming and fire blight resistance (RDA 2011).

Pedigree II (Fig. 2B) shows that ‘Baegunkong’, ‘Jangyeobkong’ and ‘Keunolkong’ take central positions as basic parental lines, which comprises three pre-existing clusters 3 (‘Dongpuk-tae’ group), 5 (‘Jangyeob’ group) and 7 (‘Kwangkyo’ group) (Jong et al. 2000). This pedigree harbors 61 hybridization-bred varieties, 33 breeding lines, 1 landrace and 26 introduced varieties. ‘Baegun’ was developed using ‘Kwangkyo’ as maternal line and the USA-introduced ‘Dt1-long receme’ as paternal line, and has following characteristic traits; resistance to mosaic viruses, susceptibility to necrosis viruses and cyst nematodes, tolerance against lodging with medium height of 70 cm on average. ‘Taekwangkong’ (1991), ‘Myeongjunamulkong’ (1995), ‘Daepung’ (2002) and ‘Shingi’ (2003) were developed using this cultivar as parental line. Of these, ‘Taekwangkong’ is known to have resistance against causal insect agents of mummification, brown spot and leaf spot, and can also tolerate soybean mosaic virus and microbe-derived brown spot (RDA 2012b). ‘Jangyeobkong’ was bred by crossing Japan-introduced ‘Miyagi Sirome’ with ‘SS7023’ derived from between ‘Kwangdu’ and ‘Baekmokjangyeob’. The height of this variety is approximately 50 cm with medium-to-large seed sizes. ‘Jangyeobkong’ has contributed to developing following varieties; ‘Deokyu’ (1983), ‘Mallikong’ (1990), ‘Myeongjunamulkong’ (1995), ‘Hojang’ (2002), ‘Daemang’ (Kim et al. 2005), ‘Daeyang’ (Kim et al. 2010a), ‘Daeha’ (Baek et al. 2013a). ‘Keunolkong’ was originated by selecting pure lines collected in ‘Chilgok Kyungbuk’ in 1986, which bears large seeds with high yield. It belongs to short stem type with 37 cm average height and is very strong against the lodging. This variety was reported to have resistance to mosaic/necrosis viruses, downy mildew and anthracnose (RDA 2008b). Following varieties, which were bred using ‘Keunolkong’ as crossing parent, are involved; ‘Saeolkong’ 1998), ‘Daehwangkong’ (Baek et al. 2001a), ‘Seonnogkong’ (2000). ‘Daol’ (2002), ‘Dajin’ (Oh et al. 2004), ‘Geomjeongsaeol’ (Baek et al. 2005), ‘Mirang’ (Baek et al. 2006), ‘Nokwon’ (Ko et al. 2008), ‘Sangwon’ (2007), ‘Chamol’ (2011).

Pedigree III was constructed by extending previous cluster 4 (‘Pangsa’ group) and cluster 6 (‘Danyeobkong’, Hill group) (Jong et al. 2000). This pedigree is composed of 47 crossing/mutation-bred varieties, 23 breeding lines, 2 landraces and 15 introduced varieties, in which ‘Danyeobkong’, ‘Pangsakong’ (Hong et al. 1985) and ‘Eunhakong’ take major parts as the parental lines (Fig. 2C). ‘Pangsakong’ is the first mutation-derived variety developed by treating the USA-introduced ‘CB27’ with gamma ray. This variety has a suitable property for bean sprout due to its uniformly small seed size. ‘Pangsa’ possesses a high necrosis resistance, whereas it shows a medium level resistance to the soybean mosaic viruses. ‘Danyeobkong’ was actually introduced from the USA in the name of ‘Essex’, which was developed using salinity-tolerant variety ‘Lee’ as the maternal line, and registered as Korean cultivar in 1978 (Table 1). Like ‘Pangsakong’, ‘Danyeobkong’ belongs to small seed type, resists necrosis/mosaic viruses, and is strong against the lodging with stable yield patterns. ‘Eunha’ (1986), ‘Namhaekong’ (Shin et al. 1989), ‘Iksannamul’ (1995) and ‘Wongwang’ (Oh et al. 2009) have derived from ‘Danyeobkong’ using as crossing parent. Of these, ‘Eunhakong’ was developed by employing the USA-introduced ‘D69-7816’ as maternal line with an aim to make a new sprout cultivar. This cultivar has suitable features, i.e., small seeded with high yield, for the bean sprout, and is known to show resistance to necrosis, root rot, downy mildew and purpura, whereas it is weak to mosaic viruses. Owing to these qualities, ‘Eunha’ has frequently been employed to develop other useful varieties such as ‘Alchankong’ (1996), ‘Somyeongkong’ (1998), ‘Sowonkong’ (Park et al. 2000), ‘Anpyeong’ (Yun et al. 2005b), ‘Dachae’ (Shin et al. 2003), ‘Jangki’ (Oh et al. 2006), and ‘Jonam’ (Oh et al. 2007).

Pedigree IV is a kind of orphan ones, because most varieties lack information available for breeding pathways (Fig. 2D). Actually it consist of three small pedigrees containing 6 introduced variety, 11 Korea-bred varieties, 12 breeding lines and 3 landraces. Except for ‘Jungmo3004’ and ‘jungmo2010ho’, other major varieties in the pedigree including ‘Ilpumgeomjeong2’ (2005), ‘Heugmi’ (2006), ‘Socheong’ (Baek et al. 2000) and ‘Soheuk’ (2009) are all black-seeded and developed in relatively recent years with favorable traits. For example, ‘Ilpumgeomjeong2’ has a varietal property of medium-to-large seed, short stem, stress tolerance with high yield, and its average height is 49 cm, which offers a desirable strength against the lodging, high moisture content/rainfall and drought. This cultivar is also popular for cooking-with-rice due to its color and sweetness (RDA 2006). ‘Socheong’ and ‘Soheuk’ were bred using ‘Milyang78’, which derived from collected black soybean landrace, as the maternal line and ‘Peking’ (small seeded and lodging-resistant) as paternal line. ‘Socheong’ is 62 cm in its height, bears relatively more pods, and is very strong against the lodging. ‘Soheuk’ is a black seeded variety with green cotyledon, and was developed to replace a popular landrace ‘Jwinunikong’. This variety has been further improved, compared to pre-existing small-seeded black soybean varieties, for higher absorption rate of moisture, and for larger pod numbers per hill (Baek et al. 2013b).

Pedigree for crops is truly essential to effectively manage breeding programs, and may provide breeders with pivotal information for selection of parental lines and design of crossing strategies. In this review, we intended not only to integrate all available information on currently registered Korean soybean varieties, but also to reconstruct an entire breadth of their pedigree. Such an attempt has resulted in a total of four pedigrees (Fig. 2), which is a lot more extended compared to the latest pedigree analyses (Jong et al. 2006). Out of 178 soybean varieties officially recorded at the GBIC and the KSVS, a total of 168 (94.4%) cultivars could be connected within the context of pedigrees. These pedigrees were reconstructed almost purely based on publically available information accessible to literatures and public databases.

Soybean has long been cultivated for major source of protein and oil in the human history. In more recent years, it is becoming more important as a source of well-being for mankind in terms of medicinal use as well as food. Additionally its unique biological property of symbiotic nitrogen fixation is beneficial to enrich the soil, thereby protecting environment and enabling sustainable agricultural practice. Another benefit of soybean on human health has been widely explored during the last two decades (Ali 2010). Thanks to its wide range of beneficial aspects, it is likely that the cultivation and breeding of soybean will be increasingly promoted in the future.

In recent years, we are facing a transition stage in breeding technologies, from conventional (phenotype-first) breeding to genomics-driven (genomic information-based design) molecular breeding. Genome-wide understanding and its application to breeding are being significantly facilitated by cutting edge NGS technologies in conjunction with high throughput bioinformatic analyses. Such situation may lead to opening of a new phase of molecular breeding in the near future, so called ‘breeding-by-design’. In parallel with technological advancement, well-organized crop resources and information will become more and more important. In summary, as crop pedigree is one of the most important information for breeding, we anticipate that the pedigree rebuilt in this study will play a constructive role in breeding a diverse array of new soybean varieties with desirable traits when it is synergized with new findings and knowledge driven by ever advancing technologies and ‘omics’-based bioinformatic tools.