| Abiotic stresses(low tempreture, salinity and drought, etc) are serious threats to rice production, leading to greatly reduced yields. To ensure food security, it is therefore crucial to screen rice germplasm for abiotic stress tolerance and to elucidate the genetic mechanisms. Abiotic stress tolerance is very complex trait controlled by polygenes. Association mapping as a powerful tool is widely used in plants for the genetic dissection of complex quantitative traits.Rice landraces are extremely important germplasms for enriching the gene pool. Many rice landraces were successfully employed in many rice-breeding projects in China. Although, with the development of modern agriculture, local rice landraces were largely replaced by genetically uniform modern varieties, many traditional rice landraces are still planted in Yunnan due to the beneficial characteristics of the traditional landraces, such as their better adaptability, built-in resistance, high quality, and relationship to the culture of ethnic minorities Therefore, the strategic conservation and use of rice landraces in Yunnan are of great importance.In this study, association mapping was performed in 347 rice accessions worldwide in order to identify the genetic marker loci/QTL associated with stress tolerance, and we also performed the analysis of the levels and patterns of nucleotide variation in Yunnan rice landraces under on-farm conservation conditions based on large-scale sequencing of 600 rice accessions. The main results were summarized as follows:1. Association mapping of stress tolerance in japonica rice germplasm(1) A total of 347 rice accessions from 11 provinces in China and 9 other countries were used in this study. 148 SSR markers were used to measure the genetic diversity of the population. All of them showed polymorphism with a total of 805 alleles which were detected across the 347 rice accessions, and the average number of alleles per locus was 5.44. The average gene diversity index was 0.3413, ranging from 0.0058 to 0.8752. The average PIC value was 0.3137, ranging from 0.0058 to 0.8643(2) Population structure analysis identified three main subpopulations for the accessions, including P1, P2 and P3, which corresponded to major geographic origins. Relative kinship showed that 57.3% of the pair-wise kinship estimates were equal to zero. Almost 41.1% of the estimates were less than 0.3. The kinship analysis indicated that most accessions had no or weak relationship with the other accessions in this rice panel.(3) Within the entire population, 64.1% of SSR locus pairs were in significant linkage disequilibrium(LD)(P<0.05). Most of this LD was due to the overall population structure, as the percentage of locus pairs in LD was much lower within each subpopulation, ranging from 21.2% to 32.4%. In the entire population, r2 values among locus pairs ranged from 0 to 0.7270. Within subpopulations, average r2 values ranged from 0.0400(P2) to 0.1633(P3). LD decayed with genetic distance, indicating that linkage was a main cause of LD.(4) In total, 24 markers based on Q+K model were identified that were significantly associated with cold tolerance, including five markers in Yunnan and nineteen markers in Jilin. Six of these identified markers were located either in or nearby the regions where the QTLs have been reported for cold tolerance. Moreover, RM282, RM252, RM335 and RM6824 which were less affected by environment backgrounds were identified in multiple locations or years.(5) A total of 40 markers were identi?ed based on Q+K model: 25 related to salinity tolerance and 15 related to alkalinity tolerance. Of these loci, 17 were located in or near regions of previously identified QTLs for these traits. Furthermore, RM475, RM567, and RM505 were indentified under both salinity and alkalinity stress conditions. It indicates that salinity and alkalinity tolerance may have common genetic base.(6) A total of 59 loci(11 for days to heading, 12 for plant height, 4 for effective panicles, 11 for grains per panicle, 9 for spikelet fertility and 12 for 1,000-grain weight) were identified as significant for all traits based on the Q+K model. Of these loci, 23 were located in or near regions of previously identified QTLs for these traits. Furthermore, as many as eleven markers were associations with more than two traits simultaneously, which indicated that one trait may be correlated with the other and that common QTLs might exist as regulatory factors for both traits.2. Diachronic analysis of genetic diversity in rice landraces under on-farm conservation in Yunnan,China(1) Ten unlinked gene regions, including Cat A, GBSSII, Os1977, STS22, STS90, S5, Pid3, Ehd1, GS3, and GS5 were sequenced in 332 accessions of Yunnan rice landraces collected in 1980, as well as 268 accessions collected in 2007 to analyze the levels and patterns of nucleotide variation in Yunnan rice landraces under on-farm conservation conditions. The average silent nucleotide variation across loci was not significantly different between rice landraces from 1980(π = 0.0061, θw = 0.0032) versus 2007(π = 0.0058, θw = 0.0033; P > 0.05 for both π and θw). The results indicated genetic diversity was successfully maintained under on-farm conservation.(2) In total, rice landraces from 2007 had maintained 90.7% of the haplotypes found in 1980. But seven haplotypes were lost, with an average frequency of 1.38%, ranging from 0.90% to 3.59%; five haplotypes were new, with an average frequency of 1.90%, ranging from 1.16% to 4.25%. The results showed that with the exception of a few lost and new haplotypes, rice landraces grown under on-farm conservation conditions in 2007 maintained almost all of the haplotypes found in 1980.(3) Among all ecological zones, rice landraces from EZI in 1980 had the highest percentage of private haplotype richness(33.1%) and the lowest percentage of private haplotype richness(16.5%) in 2007, whereas rice landraces in EZIII had the lowest percentage of private haplotype richness(18.3%) in 1980. This result revealed that rice landraces grown in the ecological zone with the richest genetic diversity(EZIII) lost less haplotype richness than those grown in the other ecological zones, and they better maintained the genetic diversity.(4) Population structure revealed that the rice landraces could be grouped into two subpopulations, namely, the indica and japonica groups. Interestingly, the alternate distribution of indica and japonica rice landraces could be found in each ecological zone, which may be the result of the unique vertical three-dimensional climate in Yunnan.(5) The results of AMOVA(considering only the two periods) showed that variance components attributed between the periods were significant for half of the loci(P<0.05). The results showed that on-farm conservation provides opportunities for continued differentiation and variation of landraces. Therefore, dynamic conservation measures such as on-farm conservation(which is a backup, complementary strategy to ex situ conservation) should be encouraged and enhanced, especially in crop genetic diversity centers.(6) The FST results reveal that genetic differentiation between periods(-0.0139 to 0.0537) was lower than that among ecological zones within periods(-0.0021 to 0.0644 for rice accessions from 1980, and-0.0138 to 0.1728 for rice accessions from 2007). The average FST value among ecological zones in 2007(0.0664) was higher than in 1980(0.0310), which demonstrates that genetic differentiation became stronger among different ecological zones during nearly 30 years of domestication. |
