Rice (Oryza sativa L.), as a staple food, has played an essential role in Korean culture and life (NICS, 2017). Despite a 3,000 year history of rice cultivation, the application of modern rice breeding techniques began relatively recently. The evolution of Korean rice breeding can be divided into several stages according to advances in breeding technology and social demands (Nongsaro, 2019). In the first stage, the Agricultural Demonstration Station was established in 1906 for modern crop research. Rice researchers continued to introduce and disseminate Japanese rice varieties until the first Korean-bred rice variety “Namseon 13” was released in 1933 (Nongsaro, 2019). Rice breeding was discontinued during colonial liberation and the Korean War of the 1950s. The second stage, from 1964 to 1970, marked the introduction of modern rice breeding methods, such as pedigree breeding, interspecific hybridization, and rapid generation advances. The first high-yielding variety Tongil was created from an interspecific hybrid of Indica × Japonica, resulting in the self-sufficiency of Korean rice production in the third stage (1971-1980). The fourth stage was characterized as a transition period from productivity to cultural stability from 1981 to 1990. One significant achievement in this period was the integration of the semi-dwarf gene and disease resistance. Biotechnology in rice breeding was adopted during this period and the first variety, “Hwaseong” was developed by anther culture (Nongsaro, 2019). In the fifth stage (1991-2000), cost-effectiveness and global competitiveness were emphasized. The primary breeding goal to reduce inputs shifted from high-yield to short growth duration. Multiple varieties with resistance to biotic and abiotic stresses were developed to reduce the cost and environmental impact of chemical applications. The adoption of molecular approaches to breeding, such as marker-assisted selection and gene transformation, was also investigated during this period. In the last stage (2000-present), grain quality and stress resistance were refined to improve rice farming production. Rice varieties were diversified following reductions in rice consumption and to address increased demands of processing industries.

Rice breeding and varietal improvement in Korea has transformed with farmers’ requests and socio-environmental circumstances. However, there are few reports on the pedigree (Song et al. 2002; Kwon et al. 2017), genetic simi-larity (Jong 2003), and agro-morphological traits (Kim et al. 2016) of Korean rice varieties. Furthermore, long-term analyses of genetic resources and agronomic trait shifts in breeding schemes are lacking. This study analyzed the parental origins and hybrid types used in the genetic development of 325 Korean rice varieties released by the National Institute of Crop Science (NICS), Rural Development Administration (RDA) over the last 40 years. Additionally, we summarized the major shifts in agronomic traits and biotic stress resistance of released varieties over the decades.

In Korea, rice research is conducted mainly by the NICS under the RDA. More than 40 rice varieties were developed by the NICS before 1979; however, some agronomic data were not openly accessible. Thus, we analyzed 325 Korean rice varieties released after 1980 in this study (Table 1). Agronomic and parental information were collected from the “New cultivar” section of the Korean Journal of Plant Breeding (http://www.breeding.or.kr) and Nongsaro (https://www.nongsaro.go.kr) operated by the RDA. Data descriptions, such as breeding purpose, followed by technical questionnaires were submitted to the Korean Seed and Variety Service (KSVS) when request testing. Data matrices were constructed using Microsoft Excel for detailed analyses of the parental information and agronomic traits of each variety.

Table 1 . List of rice varieties developed by Rural Development Administration in Korea from 1980 to 2017.

Year	No. of Variety	Name of variety
1980’s	33	Milyang42, Cheongcheong, Dongjin, Sujeong, Gaya, Samgang, Sobaek, Sinseonchal, Yeongpung, Gwangmyeong, Weonpung, Odea, Chilseong, Daecheong, Yeongdeok, Yeongsan, Yongmun, Unbong, Jangsung, Hwaseong, Namyeong, Yongju, Palgong, Hwacheong, Geumo, Donghae, Sangnambat, Tamjin, Hwajin, Gyehwa, Jangan, Jinmi, Cheongmyeong
1990’s	85	Namwon, Seoan, Obong, Jinbuchal, Sinunbong, Anjung, Mangeom, Jinbu, Jinbuol, Gancheok, Yeongnam, Hwayeong, Daeya, Sangju, Hwaseonchal, Dunae, Hwajung, Daerib1, Joryeong, Nongan, Shinkeumo, Hyangmi1, Daean, Sambaek, Unjang, Juan, Geumnam, Sangsan, Hwanam, Samcheon, Ansan, Hyangnam, Dasan, Yangjo, Seojin, Ilpum, Geumo1, Namcheon, Naepung, Daejin, Dongan, Hwasin, Junghwa, Hyangmi2, Aranghyangchal, Yeonghae, Hwadong, Sangjuchal, Guru, Nampyeong, Daesan, Hwasam, Geumo2, Namgang, Heugnam, Sangmi, Sura, Inweol, Hoan, Dongjinchal, Mihyang, Gwangan, Anda, Undoo, Weonhwang, Manan, Hwabong, Nongho, Heugjinju, Hwamyeong, Seolhyangchal, Saechucheong, Areum, Munjang, Sampyeong, Sobi, Sujin, Ilmi, Jungan, Jinpum, Sindongjin, Jungsan, Goami, Manpung, Anseong
2000’s	115	Haepyeong, Jeogjinju, Jinbong, Taebong, Hwaan, Junam, Hojin, Heughyang, Saegyehwa, Dongjin1, Manweol, Manchu, Saesangju, Seokjeong, Yeongan, Jongnam, Baekjinju, Seolgaeng, Manmi, Geuman, Namil, Daepyeong, Manho, Samdeok, Seogan, Taeseong, Hanareum, Goami2, Hopyeong, Samgwang, Sangok, Seopyeong, Joan, Pyeongan, Hwarang, Heugkwang, Josaengheugchal, Pungmi, Goun, Gopum, Unkwang, Cheongho, Hanmaeum, Boseogchal, Heapyeongchal, Geumo3, Manna, Baegjinju1, Seoan1, Sinunbong1, Odae1, Juan1, Pungmi1, Hanam, Onuri, Hwasin1, Dongjin2, Keunun, Nunbora, Malgmi, Junamjosaeng, Cheonga, Hwangeumbora, Gangbaek, Dami, Donghaejinmi, Keunseom, Hanganchal1, Haechanmulgyeol, Hongjinju, Hwangeunuri, Cheongdam, Hopum, Sandeoljinmi, Nogyang, Pyeongwon, Dasan1, Handeol, Goami3, Baekseolchal, Borami, Saenuri, Unmi, Jokwang, Cheongam, Hwanggeumnodeul, Heugseol, Danmi, Chilbo, Boseok, Joami, Dacheong, Boseogheugchal, Jinsumi, Haiami, Haeoreumi, Deurechan, Baekogchal, Jinbaek, Yeonghojinmi, Jinbo, Cheongnam, Goami4, Dasan2, Manjong, Mogwoo, Migwang, Boranchan, Seomyeong, Saegaejinmi, Joun, Hanseol, Geumyoung, Honong, Cheonghaejinmi
2010’s	92	Geonganghongmi, Dongbo, Weolbaek, Mipum, Gangchan, Seolbaek, Suryeojinmi, Suan, Jopyeong, Cheongbaekhal, Chinong, Hanareum2, Sinbaek, Jeogjinjuchal, Mogyang, Anmi, Geonyangmi, Jungsaenggold, Daebo, Saeilmi, Seonhyangheukmi, Sodami, Sugwang, Joeunheukmi, Huimangchan, Saegoami, Seolemi, Geonyang2, Hwawang, Hyunpum, Dabo, Saeodea Subo, Jinseolchal, Cheongun, Cheongcheongjinmi, Chindeul, Heugsujeong, Sanhomi, Mimyeon, Palbangmi, Ondami, Dodam, Misomi, Sinbo, Asemi, Jinok, Haepum, Bodrami, Haedam, Baegilmi, Manbaek, Saesin, Yeongbo, Unilchal, Nunkeonheugchal, Seonpum, Anbaek, Nokwoo, Asemi1, Saemimyeon, Danpyeong, Hanareumchal, Boramchal, Yeongwoo, Jopum, Jinhan, Cheongpum, Cheonghyangheugmi, Heugjinmi, Joil, Samgwang1, Unbaegchal, Hanareum3, Miho, Jingwang, Sangbo, Sangsan, Shinpyeong, Hangaru, Hanareum4, Hyangcheola, Sinjinbaek, Aromi, Misiru, Miwoo, Saeilpum, Saechilbo, Yeachan, Cheongwoo, Keunpum, Haedeul
Sum	325	-

Chronological development and diversification of variety types

The NICS conducts most rice research and varietal improvements in Korea; 325 varieties have been developed since 1980. The number of varieties released per year increased continuously from 1980 to 2006, and then gradually declined (Fig. 1). The average number of varieties released in the 1980s was 3.3 variety/year, which increased to 8.5/year in the 1990s, 11.5/year in the 2000s, and then decreased to 9.2/year in the 2010s. The ideal number of yearly variety releases remains controversial. Despite the number of varieties cataloged yearly, KSVS currently produces 24 cultivars of farmer’s seed. Among 325 varieties, 232 (71.4%) were categorized as good quality for eating in rice bowls, 27 (8.3%) were high yielding, and 21 varieties (6.5%) were for waxy rice (Fig. 2). High-yielding varieties were dominant in the 1980s (36.4%) but decreased dramatically after the 1990s (7.1%). The proportion of high quality for eating varieties gradually declined as they were replaced by varieties with other purposes in the 2000s. The percentage of special-purpose varieties such as food processing, colored, fragrant, and forages almost doubled from 11.8% in the 1990s to 22.8% in the 2010s. The main reasons for the diversification of varieties were the reduction of domestic consumption per capita and changes in people’s preference trends. Colored rice then advanced to colored waxy (Song et al. 2010; Park et al. 2015) and colored aromatic varieties (Kim et al. 2014). The release of the first cattle feed variety “Nokyang” (Yang et al. 2013) in 2006 against increased international grain market prices for animal feeds changed rice breeding trends. Nutrient-fortified high-lysine rice “Yeonganbyeo” (Song et al. 2007), high essential amino acid “Haiami” (Hong et al. 2011), and high iron “Hyangcheola” (Jeong et al. 2018) were also released during this period.

Figure 1. The number of rice varieties released yearly by the Rural Development Administration (RDA) from 1980 to 2017.

Figure 2. Classification of Korean rice varieties grouped by breeding purpose in decade intervals.

Parental origins and hybrid types

Germplasm is an important factor for maintaining genetic diversity and hybrid vigor in plant breeding. The parental origins of 325 Korean varieties show that 66% of the varieties originate from two Korean-bred stocks and 26% of the varieties have at least one parent from Japanese stocks (Fig. 3). However, only 8% of genetic stocks were used as cross parents from other countries. Genetic stocks from the Philippines were most frequently used in the 1980s as hybrid parents for high-yielding rice. Chinese germplasms were used after the 1990s, especially in parental sources for colored rice. Decadal shifts in the origins of hybridization parent stocks show that the proportion of Korean varieties increased from 45.5% in the 1980s to 72.8% in the 2010s. Japanese genetic stocks decreased from 39.4% to 17.4% during the same period. The most frequently used Japanese varieties were “Koshihikari,” “Hitomebore,” and “Kinuhikari,” which improved the eating quality. Japanese varieties were also used to breed waxy rice varieties.

Figure 3. Analysis of the hybridization type of Korean rice varieties in decade intervals.

Our analysis of the crossing type revealed that the single-cross hybrid method was dominant in rice breeding programs. Single-cross varieties accounted for 79% of hybrids, and 16% were bred by double-crossing (Fig. 4). The most frequently used parental combinations for hybridization were cultivar × elite line (33%) and cultivar × entry (24%) (Fig. 5). Varieties bred from cultivar × cultivar crosses were relatively low at 19%. This suggests that higher proportions of entry and elite lines have been used extensively as cross parents by breeders.

Figure 4. Analysis of the origin countries used in parents for genetic cross of Korean rice varieties in decade intervals. *Korea represents both parents are Korean origin and Japan, China, Philippines represent at least one parent.

Figure 5. Analysis of the hybridization parent identity of Korean rice varieties in decade intervals.

Decadal agronomic trait shifts

The main shifts in agronomic traits, including culm and panicle length, seeds per panicle, flowering date, and yields of the high eating quality varieties were analyzed (Fig. 6). Among 232 varieties, culm length was continuously reduced from 79.1 cm to 74.8 cm from the 1980s to the 2010s. Culm length decreased significantly from 79.1 cm in the 1980s to 75.5 cm in the 1990s. Panicle length increased slightly from 19.7 cm in the 1980s to 20.7 cm in the 2010s. The number of panicles per hill decreased from 16.1 in the 1980s to 13.9 in the 2010s. The number of seeds per panicle increased from 91.1 to 103.3 during the same period. “Nongan” (Nongsaro, 2019) has the highest seed set per panicle among the 193 varieties with seeds. There were minor changes in the ripening ratio and 1,000-grain weight over the decades. The mean flowering date showed little variation, ranging from the 8^th to the 11^th of August. Milled rice yields have increased dramatically since the 1980s. The average yield increased from 491 kg/10a in the 1980s to 555 kg/10a in the 2010s. The increased number of seeds per panicle and reduced culm length are likely direct/indirect factors affecting the increase in yield. Li et al. (2019) analyzed the yield components of 7,686 Chinese rice varieties and reported that the number of panicles per unit area, grain filling ratio, and 1,000-grain weight all had direct positive effects on grain yields; plant height and growth period indirectly affected grain yields.

Figure 6. The decadal shifts of agronomic traits in 232 eating quality rice varieties. (A) Culm length, (B) panicle length, (C) number of panicles per hill, (D) number of seeds per panicle, (E) flowering date, and (F) milled rice yield (kg/10a). *Data source of agronomic traits used from the technical questionnaires submitted to the Korean Seed and Variety Service (KSVS) when request testing.

The major rice diseases in Korea are blast (Magnaporthe grisea, BL), bacterial leaf blight (Xanthomonas oryzae, BLB), and rice stripe virus (RSV). RSV is a member of the Tenuivirus genus transmitted by small brown planthopper (Laodelphax striatellus Fallen, sBph). In recent years, frequent outbreaks of migration insects such as brown planthopper (Nilaparvata lugens, Bph), white-backed planthopper (Sogatella furcifera Horvath), and sBph from south-east Asia threatened rice cultivation. Table 2 shows the number of disease- and insect-resistant varieties at decade intervals. Among them, 269 varieties (82.8%) have one or more resistance. Varieties that have resistance to major diseases, BL + BLB + RSV, account for 17.8%. Varieties without any disease or insect resistance increased to 24.4% in the 2000s. The number of varieties with multiple resistances to BL + BLB + RSV increased continuously from 1.9% in the 1980s to 17.8% in the 2010s. Among 325 varieties, four varieties had multiple insect and disease resistance to BL + BLB + RSV + DV + RBSDV + Bph. “Dasan 2” (Nongsaro, 2019) is a unique variety that is resistant to both sBph and Bph simultaneously with other diseases.

Table 2 . Number of rice varieties that are disease and insect resistant.

Year	Total	1980-1989	1991-2000	2001-2010	2011-
No. of varieties	325 (100)	33 (100)	85 (100)	115 (100)	92 (100)
None	56 (17.2)	4 (12.1)	11 (12.9)	30 (26.1)	11 (12.0)
BL	48 (14.8)	3 (9.1)	18 (21.2)	14 (12.2)	13 (14.1)
BLB	9 (2.8)	-	3 (3.5)	5 (4.3)	1 (1.1)
RSV	31 (9.5)	3 (9.1)	18 (21.2)	5 (4.3)	5 (5.4)
BL + BLB	19 (5.8)	-	1 (1.2)	13 (11.3)	5 (5.4)
BL + BLB + RSV	45 (13.9)	2 (6.1)	12 (14.1)	15 (13.0)	16 (17.4)
BL + BLB + RSV + DV	7 (2.2)	-	1 (1.2)	4 (3.5)	2 (2.2)
BL + BLB + RSV + DV + RBSDV	3 (0.9)	3 (9.1)	-	-	-
BL + BLB + RSV + DV + RBSDV + insect resist.	3 (0.9)	1 (3.0)	1 (1.2)	1 (0.9)	-
Other combinations of resistance	104 (32.0)	17 (51.5)	20 (23.5)	28 (24.4)	39 (42.4)

BL: Rice blast, BLB: Bacterial leaf blight, RSV: Rice stripe virus, DV: Dwarf virus, RBSDV: Rice black-streaked dwarf virus.

Korean rice breeding shows dynamic changes according to social and economic demands over the last 40 years. The type of rice variety diversified from high eating quality rice to colored, aromatic, and processing types. Plant architecture and resistance to biotic stresses also transformed. Plant height was reduced to increase lodging resistance and the number of seeds per panicle increased continuously over the last four decades. As a consequence of extensive breeding efforts, milled rice yields increased by 13% from 491 kg/10a to 555 kg/10a. More than 80% of released varieties had resistance to single or multiple diseases and insects. However, varieties resistant to three major diseases, rice blast, bacterial leaf blight, and RSV, remained at 17.8%.

One of the potential risks in Korean rice breeding stems from the narrow pool of genetic diversity. Except for special-purpose varieties, many varieties use Korean origin germplasms as crossing parents. The limited genetic diversity has a negative influence on hybrid vigor and problems may arise in the future. The use of more Korean genetic stocks in crossing might be a double-edged sword for rice breeding. It might produce a unique Korean varietal group; however, it might also reduce the quality and genetic diversity of hybrid varieties. The genetic similarity of Korean rice has previously been calculated using phylogenetic analysis at the molecular level (Kwon et al. 2017). Song et al. (2002) reported that the four leading varieties that share “Milyang 95” as a parent accounted for 40% of rice acreage in Korea in 2001. The question of where to acquire virgin germplasms for future rice breeding in Korea remains. The utilization of Japanese genetic resources with high breeder acceptability can be one solution. However, Yi et al. (2006) reported that more than 90% of 17 high eating quality rice varieties widely accepted in Korea had Japanese varieties as their ancestral pedigree. Furthermore, abiotic stresses are potential risks for future rice breeding in Korea. The frequency of high and low temperatures, drought during the reproductive stage, and rainy harvest seasons have increased due to global warming and climatic changes in recent years. We suggest the use of European rice resources to combat abiotic stressors. The EU has a unique European Rice Germplasm Collection from the EU, USA, Argentina, and tropical regions (Sleper and Poehlman 2006; Courtois et al. 2012). The advantages of European rice are that most of them are Japonica varieties with good agronomic traits such as short maturation, cold tolerance, and direct seeding (Yi 2017).

This work was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture, Forestry, and Fisheries (IPET) through Golden Seed Project (213009-05-4-WT311). The author would like to thank to Dr. Jun-Hyun Cho, National Institute of Crop Science, Rural Development Administration, providing pedigree information of rice varieties.