Agro-morphological characterization of durum wheat accessions using pattern analysis.(Report)

Introduction

Genetic variability present in collection and preserved germplasms are important resource in generating new plant ideotypes having desired traits that help to increase crop production and thus improve the level of human nutrition (Singh, 1991). The germplasms of diverse plant species are maintained in gene banks around the world with collections holding anywhere from hundreds to thousands of accessions (IBPGR, 1992). Together with the important role of conserving genetic resources, gene banks also provide accessions for incorporation into plant breeding programs to develop new cultivars of crop and pasture species (Shands and Wiesner, 1992). In breeding programs, characterization of accessions based on multiple traits can be used as a management tool in regenerations to allow validating the identity of an accession. Evaluation data is used when searching the bank for useful or sets of useful germplasm (Delacy et al., 2000). Studies of the variation present in germplasm collections have been carried out frequently using characterization of plant morphological attributes for white clover (Caradus et al., 1990; Jahufer et al., 1997; Rosso and Pagano, 2001), alfalfa (Rumbaugh et al., 1988), wheat (Pecetti, 1992), white lupin (Rubio et al., 2004), Fenugreek (McCormick et al., 2009), apricot (Ruiz and Egea, 2008), watermelon (Szamosi et al., 2009), sesame (Morris, 2009), safflower (Elfadl et al., 2010) and vineyard peach (Nikolic et al., 2010). Germplasm collections continue to play a vital role in providing the genetic resources needed for improving durum wheat. The ICARDA gene bank preserves more than 19600 accessions of durum wheat (including 15,020 accessions of landraces), 849 accessions of primitive wheats, 1585 accessions of wild Triticum and 301 accessions of Aegilops species with S genome (out of 1905 accessions of Aegilops). CIMMYT gene bank has also large number of accessions including 14,835 accessions (36% of which is breeding material of T. durum and accessions of T. diccoccom and T. carthlicum). Iran national gene bank holds a unique set of landraces of durum wheat (700 accessions) and wild relatives collected mainly within the country. International centers, annually collect, regenerate and conserve the genetic resources and they also evaluate and compare the morphological characteristics of accessions in a common environment. In fact, these aspects of breeding programs are conducted throughout the target region every year in which multiple traits are usually recorded. Effective interpretation and utilization of these breeding programs data is important at all stages of plant breeding, particularly when it is only possible to select on yield components. Numerous methods have been used to understand of the data patterns, although strategies may differ in overall appropriateness, different methods usually lead to the same or similar conclusions for a given dataset (Flores et al., 1998; Rubio et al. , 2004). Different statistical procedures, ranging from simple univariate to the more complex multivariate techniques, have been used in the analysis of characterization data in the germplasm collection. Although statistics such as means, ranges and variances are helpful in providing information on the diversity of accessions in germplasm collections, they do not enable the simultaneous comparison of the accessions and the plant attributes (Harch et al., 1995). Pattern analysis techniques such as clustering and ordination, have been used extensively to study the diversity among accessions for various plant species (Harch et al., 1995; Harch et al., 1996; Jahufer et al., 1997; DeLacy et al., 2000; Rosso and Pagano, 2001). In addition to clustering technique, the genotype-by-trait (GT) biplot has been applied to study relation among studied traits in a set of genotypes (Yan and Rajcan, 2002; Rubio et al., 2004; Peterson et al., 2005; Yan and Fregeau-Reid, 2008). It is an application of the GGE (genotype plus genotype-by-environment) biplot technique to study of the genotype-by-trait data, and to examine its usefulness in visualizing crop trait relationships, and its application in genotype evaluation comparison, and selection (Yan and Rajcan, 2002). However, little is known about the characterization of durum wheat accessions maintained in the gene bank of Iran based on a large number of qualitative and quantitative traits as well as the interrelationships among the traits which are more affect on genotype discrimination. More information, however, is needed to find out adapted accessions that are suitable for durum wheat breeding program in Iran. Therefore, the main objectives of this investigation were to (i) evaluate 85 durum wheat accessions on the basis of multiple agro-morphological traits which could be used to describe the genetic variation of a subset of accessions from several countries preserved in Iran's national gene bank and (ii) study the relationships among recorded traits using biplot analysis techniques.

Materials and Methods

Plant Materials

Genetic materials consisted of 85 durum wheat accessions maintained at the national gene bank of SPII (Seed and Plant Improvement Institute), Karaj, Iran. The accessions studied were collections from 11 countries worldwide (Supplementary 1). Four out of 85 accessions were the check cultivars (Vee/Nac and Soissons as bread wheat checks, Dena and Zardak as durum wheat checks). The field experiment was carried out under irrigated conditions at the research station of SPII (ordination: latitude 35' 48[degrees] N; longitude 51' 10[degrees] E, altitude 1321 m) in north central Iran during the cropping season 2008-09. The plant materials were sown in 2 rows of 2.5 m long in a non-replicated trial. The checks were repeated every 10 genotypes, intervally. Weeds were controlled manually. Fertilizer application was 135 kg N [ha.sup.-1] and 90 kg [P.sub.2][O.sub.5] [ha.sup.-1] at planting.

The traits recorded

Durum wheat accessions were examined for (i) five qualitative traits, i.e., seed color (SC), glume color (GC), lodging (Lg), leaf type (LT) and shriveled seed (SS) (Supplementary 1) and (ii) fourteen quantitative traits, i.e., days to heading (DH), plant height (PH), 1000-kernel weight (TKW), weight of seeds per spike (WSS), number of seeds per spike (NSS), number of spikelet per spike (NSPS), sterile spiklet number (SSPN), spike length (SL), number of nodes (NN), stem thickness (ST), peduncle length (PL), number of effective tillers (NET), chlorophyll rate (Chl) and grain yield (YLD) per plot (Supplementary 2).

Statistical analysis

Several simple univariate statistics including minimum, maximum, means, ranges and standard deviation (SD) were used to describe the variability among the accessions, which were obtained for each trait based on the accessions (Maggs-Kolling et al., 2000; Prosperi et al., 2006; Morris, 2009). Coefficients of variation (CV %) (Francis and Kannenberg, 1978) was also calculated from the variance components and the overall means for all the investigated traits. Combination of the mean and SD for each attribute was also used to identify superior accessions. In this case five categorizes of accessions for each attribute can be characterized. The accessions with values > (mean + SD) and > (mean + 2SD) can be identified as desirable accessions for each attribute. The accessions with values between mean [+ or -] SD are average in their performance, and those with values < (mean - SD) and < (mean - 2SD) can be identified as undesirable ones (Shakhatreh et al., 2010). This simple methodology can be useful to preliminary selection of desirable accessions base on each attribute. Pattern analysis, defined by Williams (1976) as the joint use of classification and ordination methods, was applied to characterization of durum wheat accessions. This approach attempts to identify accessions that have similar performance for a set of traits, and the traits that have similar pattern for discriminating among genotypes. Clustering of accessions based on the morphological traits was carried out using an agglomerative hierarchical clustering procedure with squared Euclidean distance as a measure of dissimilarity and incremental sums of squares (Ward, 1963) as a grouping strategy. Dendrograms were constructed on the basis of fusion level to examine similarities in pattern of performance among accessions (in reaction to morphological …