Genetic diversity analysis of linseed (linum usitatissimum L.) in different environments

[摘要] English: The study was carried out to assess the genetic diversity of 60 linseed accessions mainly collected from Ethiopia in different environments using morphological and amplified fragment length polymorphism (AFLP) markers from 2000 to 2002. AFLP and morphological characterization were conducted under glasshouse conditions at the University of the Free State in South Africa. Similarly, morphological evaluations were undertaken under field conditions at Holetta Research Centre in Ethiopia during 2000 and 2001 cropping seasons. The main objectives of the study were to determine the levels and patterns of genetic diversity along with other genetic parameters using both morphological descriptors and the AFLP markers, and to compare the usefulness and relationships of these two methods in discriminating the accessions of linseed by applying univariate, bivariate and multivariate statistical analyses. The analysis of variance for the glasshouse experiment showed highly significant difference (P < 0.01) among the accessions for Il quantitative traits measured, indicating the presence of high genetic diversity. This result was also confirmed by the principal component analysis (PCA) and cluster analysis (CA). PCA displayed that days to flowering, maturity and seeds/boil accounted for 20% of the total variability. CA grouped the accessions into Il main clusters, consisting one to 24 accessions each. No correspondence was found between clustering and geographic origins of the accessions. Estimation of broad sense heritability and predicted genetic gains (as percent of mean) were also computed to forecast the possible genetic advance in the future. Heritability ranged from 15.60% for seed yield/plant to 85.82% for initial days to flowering, whereas predicted genetic gains varied from 3.16% for days to maturity to 24.26% for plant height. The analysis of variance for the field experiments revealed highly significant differences (P < 0.01) among the accessions for Il characters. PCA and CA also denoted the same thing, showing the presence of a wide range of genetic variations between the accessions studied. CA clustered the accessions into nine classes, each consisting one to 39 accessions. The clustering was independent of collection areas. Phenotypic coefficient of variation ranged from 6.26% (days to maturity) to 54.97% (seed yield/plant), while genotypic coefficient of variation differed from 5.46% to 50.18% for the same characters. Heritability (broad sense) was in the range of 67.38% to 91.38%, whereas the predicted genetic gain varied from 9.38% to 94.37%. In both cases, the highest value was estimated for seeds/ball, while the lowest was for days to maturity. A combined analysis of variance of 10 quantative traits across four environments (two localities and two years) displayed highly significant differences (P < 0.01) among the accessions for all characters. There was also significant difference between the localities, years and their interactions for most characters, indicating the differential responses of the accessions across environments. PCA and CA also confirmed the differences between accessions and the prevalence of diversity among the accessions. PCA displayed that secondary branches/plant, plant height, days to flowering and seed yield/plant played major roles in differentiating the accessions. CA grouped the accessions into nine major clusters based on their mean performance rather than their geographic origins. Broad sense heritability and expected genetic advance were found higher across multi-environments than for the single one, indicating the importance of evaluating germplasm under different and appropriate environments. Both parameters were highest for plant height and days to flowering, indicating the effectiveness of selection for these traits. Diversity analyses for oil content, oil yield and fatty acid profiles generally indicated significant (P < 0.01) variations among the studied accessions. Oil content varied from about 29 to 36%, while oil yield ranged between 1443 and 3276 g/m2. The highest oil yield was obtained from Belay-96, followed by CDCVG (3212 g/m2), an introduction from Canada, which was identified as one the most promising genotypes in this study. Fatty acids that are principally grouped into two main groups (saturated and unsaturated) showed wide ranges of variation. Saturated fatty acids (palmitic and stearic) ranged from about 8 to 12%, whereas the unsaturated (oleic, linoleic and linolenic) ones significantly (P < 0.01) varied from 85 to 91%. The highest variation was recorded for oleic acid (14-29%), followed by linolenic (47-59%) and linoleic (10-16%) fatty acids. The accessions were grouped into six cluster based on their oil content, oil yield and linolenic acid, but into 10classes based on all their fatty acid profiles. Associations of characters were analysed across different environments to determine the extent and consistency of correlations and to identify major yield attributes of linseed. The degrees of correlation varied across years and locations due mainly to climatic factors, such as temperature, moisture regime and disease incidences. Seed yield was significantly (P < 0.01) and positively correlated with yield/plant, boils/plant and 1000-seed weight. These three traits were also strongly and positively related with plant height, branches/plant and days to maturity. Oil yield was significantly and positively associated with polyunsaturated (linoleic and linolenic) fatty acids, while it was negatively correlated with saturated (palmitic and stearic) fatty acids. The results generally showed tremendous variations in correlations of characters in response to the growing environments. The current accessions of linseed were assessed for their genetic variation by using AFLP markers. The estimated genetic distance for the entire pairwise accessions varied from 0.29 to 0.71. The top three pairs of accessions with the highest genetic distance were 22 and 57, 25 and 57 and 20 and 45. High levels of polymorphism were also observed for the analysed accessions, indicating the prevalence of genetic diversity in both exotic and local collections. Collections from central and northwest regions of Ethiopia revealed considerable variations, implying further explorations in these areas. Cluster analysis grouped the accessions into 13 classes, each consisting one to 14 accessions. AFLP was found effective in discriminating and grouping the accessions for more predictable breeding and efficient management of genetic resource in the future. Combining phenotypic and genotypic assessment of genetic diversity could help in acquiring more reliable genetic information for discriminating germplasm and cultivars. To this end, morphological and AFLP data were employed to investigate the levels and patterns of variation existing in linseed accessions and to compare the two methods. Analyses of ANOVA, PCA and CA revealed the existence of wide range of genetic diversity among the accessions. The average genetic distance of all pairwise accessions was 0.6684 for morphology and 0.5734 for AFLP markers. Higher level of variation was noted for morphology than for AFLP. Correlation between the two distances was very weak and their clustering patterns were also different. In fact, clustering patterns were variable based number of traits and environments tested. In spite of this disparity, however, both methods were found independently adequate and useful in germplasm characterization and/ or cultivar identification based on circumstances, like the availability of research facilities, fund and other essential resources.

[发布机构] University of the Free State