Analyses On Diversity Of Germplasm Resources And QTLs For Important Agronomic Traits In Oat

Oat?Avena L.?is an important source of nutritious food and feed.Two types of oat are cultivated in the world,distinguished by hulled?covered?grain and hull-less?naked?grain.The hulled type is widely distributed in the temperate zones of the world as feed.However,the hull-less type is mainly grown in China as food.Due to the benefits in preventing from dermatology,hypercholesterolemia,cardiovascular disease and diabetes,oat has been recognized as a health food in many countries in the world.Over 200,000 accessions of oat cultivars and wild relatives are stored in gene banks worldwide.More than 4800 accessions of oat germplasm have been collected and stored in the national genebank in China.In order to promote the utilization of oat germplasm in enhancement and breeding,it is necessary to carry out the researches on evaluation of oat genetic resources and identify useful genetic materials from collections.With 286 accessions of oat from five countries,namely Brazil,Canada,China,Mongolia and USA,as experimental materials,and phenotypic and genotypic analyses as tools,this research evaluated the diversity of morphological and molecular traits,reviewed genetic structure and relations of oat accessions from different countries,identified the molecular markers and QTLs associated with important agronomic traits.Follows are the major achievements:1.Phenotypic analysis:Considerable variation was presented on each of eight agronomic traits in oat accessions.The highest number of tillers?6.77?was presented by accessions from Brazil;the highest plant height?98.19cm?by China and the shortest plant height?72.13cm?by Brazil;the largest panicle?16.57cm?,number of spikelets?49.75?and flag leaf length?26.10cm?by Mongolia.The accessions from Canada presented a high yield potential,with relatively higher seed weight per plant?8.14g?and1000-seed weight?44.47g?.In the principle component analysis?PCA?,the first three components had an accumulated variability of 76.51%.Panicle length and spikelet number mainly contributed to the first component,which could be considered as the panicle factor.Effective tiller number per plant,grain weight per plant and 1000-seed weight contributed greatly to the second components,which could be considered as the yielding factor.Effective tiller number and 1000-seed weight mainly contributed to the third component with a negative role by 1000-seed weight.The most closely interrelated traits were plant height/flag leaf length?r=0.691?,panicle length/spikelets number?r=0.662?,and flag leaf length/flag leaf width?r=0.639?.The 1000-seed weight was only closely related to the number of spikelets and leaf size?length and width?.By cluster analysis,286 oat accessions were morphologically classified into five groups,which show no relationships to the origin of accessions.Group 1 has characteristics of short plant,but higher effective tiller number.Group 2 has the second highest effective tiller number.Group 3 has the lowest effective tiller number,but the second highest values for all other traits.Group 4 has the highest values for all traits except the effective tiller number.Group 5 has a middle level value for all eight morphological traits.2.Genetic structure and diversity analysis:A total of 290 alleles were identified at 83 SSR marker loci.The mean number of alleles per SSR locus was 3.49,ranging from of 2 to 8.Allele frequencies ranged from 0.002 to 0.709,averaging 0.262.The mean gene diversity of all SSR markers computed over all accessions was 0.67,and ranged from of 0.44 to 0.85.The heterozygosity of all SSR markers across accessions was 0.49,ranging from 0.00 to 0.98.PIC varied from 0.37 to 0.83.Hierarchical cluster analysis based on origins of accessions showed that Chinese accessions were closely related to those from Mongolia,whereas those from Brazil were closer to accessions from the USA.Accessions from Canada were relatively distinct from those of the other four countries indicating a more distinct genetic background in the Canadian population.Cluster analysis based on individual accessions showed that 286 oat accessions were classified into six clusters.Generally,the clusters demonstrated a close relationship between genetic diversity and geographical distribution of accessions.Overall genetic structure analysis showed that six subpopulations were identified and well matched with those clusters identified in the cluster analysis,indicating that Canadian oat germplasm have interchanged with the germplasm from China and USA.There was a flow of germplasm from Brazil to USA.Mongolian accessions were more isolated from the other countries and only exchanged with China.3.Association analysis:By the GLM model in TASSEL software,41 SSR allele markers were associated with eight agronomic traits at the level of p<0.001,and explained from 2.5%-10.8%of the phenotypic variation,including one marker for the effective tiller number,two for panicle length,eleven for spikelet number of panicle,five for plant height,seven for leaf length,three for leaf width,three for grain weight per plant and nine for 1000-seed weight.There were five co-associate allele markers?AM2212-4,AM2270-3,AM3221-3,AM3385-3 and AM810-1?identified,each of them was associated with 2 to 3 traits.It was also found that different allele markers of the same SSR were associated with a single trait,including AM2201-1/AM2201-2 for plant height,AM337-4/AM337-5 for 1000-seed weight,and AM3385-3/AM3385-3 for 1000-seed weight.4.QTL analysis:A genetic linkage map was constructed with SSR markers in the basis of 224individual plants of F₂ population generated by a cross between naked grain type and hulled grain type.The map contained 88 SSR loci,covered 975.3cM in length,and composed of 17 linkage groups.There were 19 SSR markers on the largest linkage group covering 129.3cM,and only 2 SSR markers on the smallest linkage group.The minimum mapping distance between the two markers on the linkage group was<0.01cM,and the maximum mapping distance was 39.6cM.The average mapping distance between the two markers was 11.6cM.The map will provide a useful tool for identifying the QTLs associated with naked gain type.Two QTLs identified on LG8 were defined as qGT-1 and qGT-2.The qGT-1 was located at 26.0cM on the map,explaining 10.77%of phenotypic variation of naked grain type.The qGT-2 was on the location of 43.1cM,coincided with the position of marker AM1094,explaining 7.07%of the phenotypic variation of naked grain type.The two QTLs might include a major gene controlling the naked type and a minor gene interacting with the major gene.