Material and Methods
Experimental Site and Material
The experiment was conducted during winter season 2011–12 at CCS HAU, Hisar, located at 29°–10' N latitude, 75°–46' E longitude and an elevation of 215.2 m asl. The soils were sandy loam having pH range of 7.5–8.0. The study material included entire gene bank holding of cultivated wheat accessions. For the purpose of core set development, bread wheat (T. aestivum), durum wheat (T. durum), and emmer wheat (T. dicoccum) were grown for agronomic characterization. A set of 22, 663 accession of wheat were grown in Augmented Block Design (Federer 1956) with 8 checks representing different species viz, C 306, PBW 343, DBW 17, RAJ 3765, DWR 1006, UAS 415, DDK 1025, and DDK 1029. The checks were replicated in each of the 114 blocks of 200 accessions each. Each accession was grown in three rows of 2 m length and plant to plant spacing of 25 cm. Standard agronomic practices were followed to raise a healthy crop.
All the accessions were characterized for 34 important traits, 22 qualitative and 12 quantitative, as outlined by NBPGR minimal descriptors and complete set of observation were recorded for 22,469 accessions. The qualitative characters included early growth vigour (EGV), growth habit (GH), flag leaf angle (FLA), foliage colour (FC), waxiness on leaf blade (WLB), waxiness on leaf sheath (WLS), waxiness on peduncle (WP), waxiness on spike (WS), glume pubescence (GP), auricle colour (AC), auricle pubescence (AP), awnedness (WA), awn length (AL), awn colour (AC), glume colour (GC), spike shape (SS), spike colour (SC), spike density (SD), grain colour (GC), grain shape (GS), grain texture (GT) and grain width (GW). The quantitative traits included, days to 75 % spike emergence (SE), days to 90 % maturity (DM), plant height (PH), effective tillers per plant (EF_T), spike length (SL), number of spikelets per spike (SLS), no. of grains per spike (GRS), grain weight per spike (GRW), 1,000 grain weight (TGRW), dry matter yield per m row length (DMY), grain yield of 1 m row length (GY) and harvest index (HI).
The “PowerCore” (genebank.rda.go.kr/powercore/) software developed by the Rural Development Administration (RDA), South Korea, that uses the advanced M (maximum) strategy with a heuristic search for establishing core sets possessing the power to represent all alleles or classes, was used in the present study. It creates subsets representing all alleles or observation classes, with the least allelic redundancy, and ensures a highly reproducible list of entries. This approach has recently been used in developing core set from large rice and foxtail millet collection (Chung et al. 2009; Gowda et al. 2013). It effectively simplifies the generation process of a core set while significantly cutting down the number of core entries, maintaining 100 % of the diversity as categorical variables. Core collections are considered to represent the genetic diversity of the initial collection if the following two criteria are met: (1) no more than 20 % of the traits have different means (significant at α = 0.05) between the core collection and the entire collection and (2) Coincidence Rate (CR) is retained by the core collection in no less than 80 % of the traits (Hu et al. 2000). The design, concept and implementation strategy of “PowerCore” and the validation on the outcome in comparison with other methods have been well described by Kim et al. (2007). PowerCore by default classifies the continuous variables into different categories based on Sturges rule (Sturges 1926), which is described as: K = 1 + log 2 n, where n = number of observed accessions. However, the software also allows modification of this rule to make desired number of classes for the continuous variables. Once classifi of the continuous variables is performed, the software takes into account all classes, without omission of any of its variables. It thus, possesses the capability to cover all the distribution ranges of each class.