Multipurpose selection of some Iranian endemic coriander (Coriandrum sativum L.) populations for simultaneous improvement of traits under different irrigation regimes

Introduction and goals
Coriander (Coriandrum sativum L.) is an annual plant that drought stress can impact on its yield. Drought stress is one of the most common environmental stresses limit agricultural productivities. This stress effects on many metabolic pathways such as photosynthesis, water absorption and transfer, enzymes’ activity and organic material transfer and accumulation. It can also lead to the accumulation of secondary metabolites in the plant. In the case of some quantitative and qualitative traits, such as the yield and the amount of the essential oil, the selection can be done indirectly with the help of some statistical techniques. This research was conducted to investigation of interrelated traits with fruit yield and fruit essential oil content and also to suggest two multipurpose selection procedures for simultaneous improvement of several traits in all three irrigation regimes by the help of graphical analysis in coriander. Thereafter top landraces based on all attributes for selection and cultivation in the well irrigated, gradual drought stressed and severe drought stressed irrigation regimes were identified. Material and Three experiments were carried out based on randomized complete block design with three replications at the Research Farm of Tarbiat Modares University in 2015. In experiments, normal irrigation, severe stress, and gradual stress were applied and combined analysis of variance was performed for 16 physiological, morphological and phonological traits. Genotypic and phenotypic variances, genotypic and phenotypic variation coefficients, heritability in broad sense and genetic progress percentage were calculated. Genotypic and phenotypic correlations were calculated according to Holland, 2006 in combined analysis of variance. Factor analysis using the genotypic correlation matrix of traits and plotting ecotype × trait biplot based on the first two factors, which had the highest correlation with the studied traits, was performed by MATLAB software. Clustering of genotypes and traits based on average traits in three experiments and also based on replication means separately in each experiment using Ward method and Square Euclidean Distance were performed and the corresponding heatmap was plotted using metaboanalyst 3.0 software. Combined ANOVA showed relatively high variation for ecotypes in most traits. The genotype × environment interaction was significant in most of the traits at 1% probability level. The traits of days to 50% of fruit maturity, 1000 fruit weight, fruit yield and relative water content of the leaves had a relatively high environmental impact, and the indices related to the environment had fairly large difference with the ecotype indices. The effect of the environment on the fruit yield was so high that the genetic variance was estimated to be zero. Prematurity related traits had a relatively high genetic variance, and their genetic advance was relatively high. Therefore, selection may be effective to improve these traits under normal and stressed conditions. The highest percentage of genetic advance was observed for traits of leaf chlorophyll and base leaf number. Phonological traits, basal leaf number, plant height, plant dry weight, number of branches, number of umbels, number of fruits per plant, chlorophyll content and essential oil content had a high value of heritability and genetic advance. In factor analysis, the first two factors explained 51.13% and 23.71% of the total variation. The first factor showed existence high genotypic correlation between the most traits with each other, and on the other hand, the second factor showed the correlation between the relative water content of the leaves and the number of days to fruit maturity. According to the obtained biplot, the ecotypes with the highest genetic potential for most of the traits were identified, which were respectively ecotypes 4, 11, 15, 3 and 14. The ecotypes and traits were categorized into three groups according to the heatmap clustering. The observed differences between these two graphical representations in the ecotypic and traits grouping are due to the fact that the biplot is independent of the environmental effects and the genotype × environment interaction, while the heatmap is drawn based on the average of ecotypes in different moisture regimes. Selection above mentioned ecotypes and obtain a superior population by them may be useful in simultaneously improving the important economic traits such as the amount of essential oil, basal leaf number and dry weight of biomass in both normal and stress conditions. Prematurity and relatively high value of harvest index in ecotype 9 make it suitable for the second cultivation in endangered areas of drought at the end of the season to produce fruit and essential oil. Ecotype No. 10 had the highest amount of essential oil and number of base leaves, but the lateness of this ecotype caused it to be severely drought sensitive and its average fruit yield reduced in different irrigation regimes. A great similarity was found between the results of the two graphical analysis methods and the traits were well grouped in a method that the combination of irrigation regimes and ecotypes was not considered.