Our study analyzed the genetic diversity of 95 cultivars and strains of loquat collected from all over the world. On the basis of the analysis of the population structure of 19 Chinese loquat cultivars and strains by RAD-Seq, Hubei Province in China has been suggested as the center of origin of loquat cultivation6. However, the authors have not analyzed the movement of loquats following their introduction from China to other countries and have not verified whether the hypothesis that the cultivation of loquat started in a small area in China24 is correct or not. On the basis of the analysis of samples from around the world, we here propose and discuss a more complex history of loquat cultivation (Fig. 1B).
Although genetic diversity analysis using SSR markers revealed no country-specific cultivar clusters14, the RAD-Seq analysis used in this study allowed for such clusters to be detected. This was possible because more markers are available in RAD-Seq analysis than in conventional SSR marker analysis. RAD-Seq analysis classified the loquat genetic resources into three groups: (1) Japanese and Chinese cultivars and some Japanese wild strains, (2) Vietnamese, Israeli, Greek, US, and Mexican cultivars and strains, and (3) other Japanese wild strains strains.
Group 2 had the highest mean values of π and He (Table 3). The Fst values (Table 2) showed that group 2 is well separated from groups 1 and 3, which may be due to differences in their place of origin, as discussed below. The Fis value of group 2 was higher than that of the Japanese wild strains (Table 3), which may reflect the fact that group 2 plants have been grown and crossbred by humans.
Group 2 was genetically separated from the Japanese and Chinese loquats. Blasco et al.25 reported that SSR and S-allele markers can distinguish European cultivars from other cultivars. Our results are in good agreement with the above study. Morton8 reported that loquat genetic resources were introduced to the West from China and Japan, but our study does not agree with this report. When plants are preserved in botanical gardens, records are well kept. However, for commercial use, records are not always kept, because the purpose of plant introduction is different. Group 2 plants may have been introduced to the West for commercial cultivation in a different way than they were introduced to botanical gardens.
The place of origin for group 2 may be in China or Japan, although our analysis failed to find any cultivars or strains in group 2 that originated from China or Japan. Unfortunately, the cultivars and strains from China analyzed in this study are not representative of all loquat cultivars from China, and we were unable to analyze cultivars and strains from southern China, such as Yunnan, Guangdong, and Guangxi. Wang et al.6 reported that ‘Younan’ from Guangdong has a close genetic relationship with cultivars from Spain, Italy, and the USA. This suggests that genetic resources from particular regions of China may have been introduced to the West.
The fact that the Vietnamese strains belong to group 2 suggests two possibilities. One is that group 2 originated from Southeast Asian countries other than China, including Vietnam, in particular because cultivars from the USA and strains from Mexico were genetically similar to strains from Vietnam. Interestingly, when group 2 was analyzed in detail, we detected the presence of an Israeli subgroup and a Greek subgroup. They may have originated in Southeast Asian countries other than Vietnam. The second possibility is that cultivars or stains introduced from China and other countries to the USA or France (the former colonial master of Vietnam) were further introduced to Vietnam. Researchers who collected the strains in Vietnam told us that these strains were not used for edible fruit, even though they grew near houses. These plants may have been cultivated in Vietnam as ornamental trees or to be offered to Westerners. Thus, group 2 cultivars are likely to originate from various places.
The differences in taste preferences between the Greek, Israeli, American, and Mexican people and the Japanese and Chinese people may have influenced cultivar differentiation. Indeed, many cultivars and strains from Greece, Israel, USA, and Mexico have higher acid content than those from Japan and China26,27. However, differences in taste preferences need to be studied in detail in the future. Although group 2 strains grow in Vietnam, the Vietnamese most likely prefer to eat group 1 fruits because they do not eat group 2 fruits. In Vietnam, loquat imported from East Asia is commercially available.
In group 1, the mean values of π and He were lower than in group 2 but higher than in group 3 (Table 3). The Fst values indicated that group 1 is genetically closer to group 3 than to group 2 (Table 2). The Fis value of group 1 was higher than that of the Japanese wild strains (Table 3), which may reflect the fact that group 1 plants, like group 2 plants, have been grown and crossbred by humans.
The PCA and MDS analyses placed Japanese and Chinese cultivars in the center of the group 1 cluster. The present Japanese cultivars originated from excellent Chinese cultivars introduced at the Edo period (1600–1868) and were further improved in Japan11. The detection of group 1 reflects this proposed history. The cultivars were believed to have been improved through crossbreeding and bud sport mutations, using mainly ‘Mogi’, ‘Tanaka’, and ‘Kusunoki’. However, some samples (22–24, 28, 38, and 48) were not related to these three cultivars. Other cultivars introduced at the Edo period or earlier or gene flow from Japanese wild strains may have contributed to the formation of these six cultivars.
With the exception of these six samples, the central part of the cluster in group 1 was divided into a subcluster of Japanese cultivars and a subcluster of Chinese cultivars. This separation may reflect differences in the breeding process between China and Japan, where mainly ‘Mogi’, ‘Tanaka’, and ‘Kusunoki’ were used. The slight difference in taste preferences between the Japanese and Chinese may have influenced the formation of the two subclusters, but further research is needed to address this issue.
The PCA and MDS analyzes placed the Japanese wild strains at the periphery of group 1. Because group 1 contains Chinese cultivars, the wild strains in this group are closely related to plants introduced to Japan from China. These wild strains may be related to plants introduced before the development of excellent cultivars in China, as well as being related to plants described in ancient Japanese documents. Although other possibilities exist for the five wild strains (52, 53, and 55–57) as discussed below, two strains (54 and 58) are possible descendants of plants introduced to Japan before the Edo period. Analysis of wild strains in China could help to solve this problem, but unfortunately we were unable to analyze them. Among these wild samples, sample 59 was closer to cultivars such as Mogi (32–36, 43, 44, 47), which may be related to cultivar escape.
PCA and MDS analyzes placed the Japanese and Chinese cultivars and the Japanese wild strains at the bottom of the cluster in group 1. Admixture analysis suggested that gene flow from isolated Japanese wild strains in group 3 may have occurred in these plants. The finding that gene flow from group 3 may have occurred in Chinese cultivars suggests that plants belonging to group 3 may be present in China. Alternatively, these Chinese cultivars may have been placed here under the influence of populations not analyzed in this study, rather than group 3. Gene flow from group 3 may have occurred to Japanese cultivars (22–24, 28, 38, and 48) and wild strains (52, 53, and 55–57). It is interesting that genetic resources belonging to group 3 are not suitable for edible use and do not have useful traits that can be used to develop new varieties, such as disease resistance.
There are two possibilities for the origin of group 3 plants growing in Japan. The first possibility is that these plants are indigenous to Japan that have not been introduced by human activities. Genetic diversity analysis using SSR markers has demonstrated that wild strains and cultivars in Japan are genetically different14, and our data support this conclusion. Group 3 plants growing in Japan had the lowest mean values of π and He (Table 3). The Fst values indicated that group 3 was genetically closer to group 1 than to group 2 (Table 2). A notable feature of group 3 was that the lowest Fis value (Table 3), suggesting that humans may not have played a role in the formation of this group. Admixture analysis detected no gene flow from other groups to group 3, even though groups 1 and 3 were genetically close, suggesting that the ancestors of group 3 were not introduced by humans from China, but were indigenous to Japan. Common plants often grow in laurel forest ecoregions in East Asia. Loquat is found in these ecoregions, and may have been growing in Japan since prehistoric times. If these plants are indigenous to Japan, the problem to be elucidated in the future is whether plants belonging to this group are present in China. The second possibility is that the group 3 plants growing in Japan were brought to Japan by human activities from China over the last few thousand years. Although modern people do not use group 3 plants for edible purposes, we cannot rule out the possibility that people in the past used them for edible purposes. In order to clarify these issues, it would be desirable to test and analyze individuals from southern China, such as those from Yunnan, Guangdong, and Guangxi provinces, as well as from the northern parts of Southeast Asian countries.
In this study, we also examined the contribution of breeding through asexual reproduction, as in the case of bud sport mutations (Table 4). We confirmed the available records of the asexual emergence of ‘Morimoto’ from ‘Tanaka’. We also found that ‘Moriowase’, ‘Amakusawase’, and ‘Amakusagokuwase’ were asexually propagated from a single tree. On the other hand, we rejected the possibility that ‘Moriowase’ originated asexually from ‘Mogi’. The above cultivars determined to have been born by asexual reproduction were very similar in fruit traits to their parent cultivars. Thus, our DNA-level analysis allowed to clarify the origin of some cultivars.
This study demonstrates that RAD-Seq analysis is applicable to the genome analysis of loquat, which has relatively low genetic diversity14. The information obtained here can be used for loquat cultivar identification and DNA profiling, and in genetic research and breeding programs.