Study the relationships between yield and yield component of faba bean (Vicia faba L.) genotypes by multivariate analyses

Faba bean (Vicia faba L.) is a major crop legume that is used as food owing to the high nutrient components in seeds. Yield improvement is a major breeding objective of most crop breeding programs. Multivariate analyses are useful for characterization, evaluation and classification of plant genetic resources when a number of accessions are to be assessed for several characters of agronomic, morphological and physiological importance. Different types of multivariate analysis such as regression analysis, path analysis, principal component analysis (PCA) can be used to identify groups of genotypes that have beneficial traits for breeding and instructing the patterns of variation in genotype accession, to recognize relationships among accessions and possible gaps. Correlation coefficients describe the mutual relationships between different pairs of characters without providing the nature of cause and effect relationship of each character. Path analysis was also performed to determine the direct and indirect contribution of each character to seed yield. Principal component analysis has been widely used in the studies of variability in germplasm collections of many species. The objective of the present study was to estimate the correlations and partition of the coefficient of correlation between seed yield with its primary components, into direct and indirect effects to determine the relative importance of each one in faba bean seed yield. The other aims of the present study are to assess the genetic diversity present in the morphological and agronomical traits in 26 faba bean genotypes by principal component analysis.

This study was carried out during 2015-16 growing season in Lorestan province, Iran (longitude, 48° 45´ E; Latitude, 35° 55´ N; Altitude, 1629 m above sea level). Experimental material comprised 26 genotypes of faba bean. Field experiments were conducted in a randomized complete block design with three replications. Each plot consisted of four rows with 4 m long and distance between rows and plants were 50 and cm, respectively. The characters included days to germination, days to flowering, days to maturity, plant height, number of stems per plant, number of node per stem, number of seeds per pod, number of seeds per plant, hundred seed weight, pod length, dry seed yield, biological yield and harvest index were measured before and after harvesting. SAS 9.2 used to analyses the correlation and regression coefficients. Path coefficients were estimated by path analysis software. Principal components analysis was done using SPSS and the graphs were drawn via Minitab.

 The analysis of variance indicated significant differences between genotypes for all of the studied traits. Correlation analysis indicated there were positive correlation coefficients between seed yield and number of days to germination, number of days to flowering, plant height, number of pods per plant, number of nodes per stem, hundred seed weight, pod length, biological yield and harvest index. Regression analysis indicated seed yield as dependent variable, while plant height, number of pods per plant, number of stems per plant, number of nodes per stem, number of seeds per pod, hundred seed weight and pod length were considered as casual variables. Path coefficient analysis indicated plant height (0.74), number of pod per plant (0.51), number of nodes per stem (0.11), number of seeds per pod (0.20), hundred seed weight (0.34) and pod length (0.41) had positive direct effects on seed yield. The results of principal component analysis showed that the eigenvalues were reduced with the increase of the number of PCs. Principal components analysis, identified six components that explained 80% of total variation. The maximum values of eigenvalues were obtained for first three PCs, which accounted for a cumulative percentage of total variance of 24.34%, 17.57% and 14.18%, respectively. The remaining percentage of the total variation decreased sharply. The eigenvalue of the first principle component had a variance of 4.14, while the other two components had much smaller variances. It could be said that the first principal component is by far the most important of the three and that the first principal component include the largest variance of any one unit the length linear combination of the observed variable. The first two and three PCs was used to grouping the genotypes in two and three-dimensional plots, respectively.

Attention should be paid to traits such as plant height, number of pods per plant, number of nodes per stem and pod length for augmentation of seed yield and these traits could be used as selection criteria in faba bean breeding programs. Tow dimensional plot based on first two principal components showed genetically different genotypes by the pattern on scattering. It could be concluded that the high yielding genotypes, such as 9 and 22 could be used to improve seed yield of faba bean and making possibilities of extending production of this legume crop.