Agro-morphological diversity of Nepalese naked barley landraces

Naked barley is a traditional, climate-resilient and highly nutritious crop of the high mountains of Nepal. Knowledge of agro-morphological diversity among the naked barely genotypes is fundamental for their efficient utilization in plant breeding schemes and effective conservation programs. The primary objective of this study is phenotypic characterization and diversity assessment of naked barley landraces in Nepal for pre-breeding purposes. Data on quantitative and qualitative traits of 25 naked barley landraces from diversity blocks established across the four mountainous locations of Nepal (Humla, Jumla, Dolakha and Lamjung) were subjected to calculation of descriptive statistics and multivariate analysis (UPGMA cluster analysis/principal component analysis). Frequency distribution of various categories of qualitative traits and Shannon–Weaver diversity indices were computed. The UPGMA cluster analysis using both quantitative and qualitative traits individually categorized the 25 naked barley landraces in five clusters in each case with no distinct regional grouping patterns in such a way that the landraces from same or adjacent regions of origin and collection amassed in different clusters. NGRC04894, Lamjung Local, NGRC02306, NPGR1579, NGRC02327 and NPGR1579 exhibited morphoagronomical superiority and potentiality for utilization as genitors in crop improvement programs. Principal component analysis revealed the quantitative traits, viz. grain yield, plant height and earliness, and qualitative traits, viz. grain color, overall phenotypic performance, lemma awn/hood and lemma awn barbs, to be the principal discriminatory characteristics of the Nepalese naked barley landrace collection. The Shannon–Weaver diversity index (H′) ranged from 0.32 to 0.99 with a mean value of 0.73, inferring tremendous diversity in the collection for the qualitative traits. The marked diversity observed among the Nepalese naked barley landraces could be utilized in crop improvement for various traits. The information generated complements the robust breeding program of competitive, stable and climate-resilient varieties of end users’ preferences in different mountainous agro-ecologies and also bolsters the employment of innovative and proven participatory plant breeding approach using diversity kits and informal research and development kits to expand and promote the varietal choice options for expeditious benefits to the farmers in the high mountains, considering that only one naked barley variety, viz. Solu Uwa, has been released in Nepal to date.

Background Hordeum L. is a widely distributed genus of the tribe Triticeae with 45 species and subspecies consisting of mostly weedy annual or perennial grasses adapted to the temperate zones of both the northern and southern hemispheres. Barley (Hordeum vulgare L. subsp. vulgare) has historically been the prevalent cereal grown in marginal agricultural areas [40]. It is the world's fourth most important cereal crop after wheat, rice and maize. A salient botanical feature of domesticated barley is that most cultivars have covered caryopses in which the hull (outer lemma and inner palea) is firmly adherent to the pericarp epidermis at maturity; but few cultivars are of the freethreshing variant called naked (hullless) barley.
Naked barley/hulless barley (Hordeum vulgare var. nudum Hook. f.; Nepalese vernacular name: Uwa/Mudule Jau) differs from hulled barley in that the loose husk cover of caryopses is easily separable upon threshing [5]. The domestication of naked barley is believed to have occurred after the hulled type around 6500 BC [40]. The hulless grain trait is governed by the single recessive gene 'nud' located on the long arm of chromosome 7H [14]. A monophyletic origin of naked barley has been suggested [41] as a single mutation event either from wild barley (H. vulgare subsp. spontaneum) or from domesticated hulled barley. It is mainly cultivated in the higher Himalayan ranges of East Asia (India, Nepal, Bhutan and Tibet), China, Korea and Japan [42,43] and also at low frequency in Ethiopia [44]. Its cultivation has been practiced for ages in the high mountains of Nepal due to its adaptability and its traditional and cultural importance. The barley landraces from the Himalayas, particularly from the highlands of Nepal, share a significant part of world barley germplasm resources [32]. As naked barley is extensively grown in the Himalayas of Nepal from the east to the west (about 800 km), along the Himalayas, it is frequently represented in Himalayan barley germplasm collections [23].
The importance of barley as a food crop increases with ascending altitude toward the north, where other cereals cannot be cultivated successfully due to poor edaphic factors and extreme environmental stress, particularly drought [21]. Naked barley is a traditional, culturally important, climate-resilient and highly nutritious winter cereal crop of the high mountainous region of Nepal. The unique geographical distribution pattern of naked barley in the higher altitudes (> 2500 masl) implies its tremendous food value and wide utility in the high mountains. Furthermore, it does not require additional processing, contrary to hulled barley which has tightly adhered grain husks. It is rich in soluble ß-glucans content, which acts as an inhibitor of cholesterol synthesis and thus is salubrious for heart patients [4,5,9]. Its integration in daily diet can help to reduce blood glucose and cholesterol [20]. Gladiators also popular as hordearii (the eaters of barley) were known for their strength and stamina [1,15]. In the highlands, its flour is mostly consumed as Satu (a flour prepared from roasted barley grains which can eaten as it is or can be used in several cuisines), flat breads, dumplings, Timmok and Thukpa (Tibetan noodle soup) and also mixed with flour of other crops like wheat, buckwheat and finger millet. Additionally, a special traditional fermented drink Chhyang and distilled liquor are prepared from the grains for different cultural and religious occasions by Lama, Karmarong and Sherpa tribes residing in mountainous region of Nepal [11,22,24,25].
Due to its hardy nature, ability to grow well in lowinput conditions and potential for use in climate change adaptation, naked barley renders itself as one of the future-and climate-smart crops in the wake of bulging food insecurity and climate change in the fragile mountainous region of Nepal. In Nepal, around 3.7 million people are estimated to be food-insecure [34]. Further, the food insecurity is critically alarming in far and mid-western mountains of Nepal [8] and likely to be aggravated more in days to come due to rapidly changing climate. In this context, the wider cultivation of this highly climate-resilient and nutritious crop could play an important role in climate change adaptation and ensuring food security in mountainous regions of Nepal. To date, only one naked barley variety, viz. Solu Uwa, has been released in Nepal with yield potential of 1.9 Mt/ha [18]. The inadequacy of promising and adaptive naked barley cultivars in national agricultural system has left mountainous farmers with nominal varietal selection options compelling them to shift to comparatively highyielding food crops such as wheat and hulled barley, effecting a sharp decline in the cultivation trend of naked barley. Despite its tremendous potential and prospects, the crop has not been able to engross satisfactory attention regarding research and development in Nepal and remains underutilized and neglected.
The existence of the marked diversity in morphology and agronomic traits between and within the landrace population and the genetic uniqueness of Nepalese naked barley has been reported by various authors [3,23,29], which signifies abounding room for robust breeding and crop improvement program. Pandey et al. [23] reported a high level of genetic diversity and complex population structure in Himalayan hulless barley landraces. Sharma et al. [29] found a vast variation in morphology between and within landrace populations. Besides this, differences in agronomic performance and disease resistance were detected [3]. Xu et al. [38] collected 562 hulless barley accessions with worldwide geographic origins and evaluated their genetic variability and relatedness based on 93 SSR markers. However, precise and detailed information on the pattern and structure of the phenotypic and agro-morphological diversity of Nepalese naked barley landraces is still lacking and paltry information is available for pre-breeding purpose. Advances in molecular techniques have opened doors to genetic screening of germplasm collections. Yet, accurate phenotyping is of paramount importance for identifying useful traits [19], visualizing the pattern of variation in agro-morphological traits, realizing the relative significance of the germplasm in breeding programs and facilitating effective management, conservation and exploitation of the gene pool. Hence, the present study, which is a part of sourcing local crop diversity to match farmers' needs through rapid diversity assessment, on-farm evaluation and dissemination of choice varieties under a GEF/UNEP funded project on local crops in Nepal (www.himal ayanc rops. org), was undertaken to generate explicit estimates of the agro-morphological and phenotypic diversity of Nepalese naked barley landraces for pre-breeding purpose and efficient deployment of the functional diversity to broaden the genetic base of community richness using a participatory plant breeding approach. In addition, the local crop project aims for large-scale testing of the candidate plant materials in the wider areas of Nepal using diversity kits for deployment of new diversity in order to expand the varietal choice options of the farmer and enhance agricultural system resilience, appraisal of the agronomic performance via crowdsourcing feedback [33] or a quick sample household survey as those employed by Joshi et al. [13] and Witcombe et al. [35], and use of evolutionary plant breeding [26] or simple grassroots breeding [31] of the crop for rapid development of a competitive and stable variety to provide better varietal choices than the existing one in a particular agro-ecology.

Plant materials, site description and field experiment
The materials used in this study consisted of a total of 25 naked barley landraces collected from different agroecological zones of Nepal (Table 1). Of all landraces, 19 were provided by National Agricultural Genetic Resource Centre (NAGRC), 5 were collected from farmer's field of different locations, and 1 (Solu Uwa) was collected from Hill Crop Research Program (HCRP). On-farm diversity block trials were laid out in four mountainous locations, viz. Lamjung, Dolakha, Humla and Jumla in 2016. Diversity block is an experimental block of farmers' varieties managed by a local institution for research and development purposes [39]. The block is not only used for measuring and analyzing agro-morphological characteristics, but also used to validate farmers' descriptors by inviting farmers to watch the diversity block in the field and determine whether farmers are consistent in naming and describing varieties. The four experimental locations are designated in Fig. 1. The geo-climatic description of the experimental locations is depicted in Table 2. Each landrace was continuously sown in 4 rows of 3 m length with the plot size of 3 m 2 with distance of 25 cm between row to row. All agronomic and management practices were followed as per the farmer's practice. Well-rotten farm yard manure @ 6-8 t ha −1 was applied in the field during land preparation. No chemical fertilizer and pesticides were used. Two hand weedings were performed 30 and 60 days after sowing. The lack of seeds of various landraces brought about asymmetrical distribution/ inclusion of the study materials in the experimental locations such that 19 landraces were tested in Lamjung, 18 in Humla, 20 in Dolakha and 22 in Jumla (Table 1).

Data collection and data analysis
Data on 5 quantitative traits such as days to emergence, days to heading, days to maturity, plant height (cm), grain yield (gplot −1 ) and 10 qualitative characteristics such as growth habitat, stem pigmentation, tillering capacity, spike density, lemma awn or hoods, lemma awn barbs, glume color, grain color, grain size and overall phenotypic acceptance were recorded as per the descriptors of barley [12]. Grain yield was recorded after sun drying in gplot −1 and later extrapolated to kgha −1 . A modified scale (3 = good, 6 = intermediate and 9 = poor) for tillering capacity and overall phenotypic acceptance was employed. For analysis of quantitative data, mean values were used for the common landraces tested across the experimental locations while data of those landraces which were assessed in only one of the experimental locations were used as such. The qualitative data of 22 landraces tested in Jumla, 2 landraces (Rato Humla and Jungu Local) tested in Dolakha and one landrace (Lamjung Local tested) in Lamjung were further analyzed. The descriptive statistics for quantitative traits of the clusters and individual landraces were calculated. Multivariate analyses based on 5 quantitative and 10 qualitative data such as Unweighted Pair Group Method with Arithmetic Mean (UPGMA) clustering based on similarity distance and principal component analysis (PCA) were performed in order to assess the degree of divergence and relatedness among the landraces and estimate the relative importance and contribution of traits to the overall variation using Minitab V17. Shannon-Weaver diversity indices [28] were calculated in order to estimate the phenotypic diversity for each qualitative trait with Microsoft Excel using the formula: Standard H′ = [∑(n/N)*{log2(n/N)*(− 1)}]/log2k, where H′ is the standardized Shannon-Weaver diversity index, k is the number of phenotypic classes for a character, n is the frequency of a phenotypic class of that character and N is the total number of observations for that character.

Pattern and structure of polymorphism for quantitative traits
Multivariate analyses, viz. UPGMA clustering and principal component analysis, were performed on a set of 25 Nepalese naked barley landraces by 5 quantitative traits, viz. days to emergence, days to heading, days to maturity, plant height and grain yield. The naked barley landraces were clustered into 5 distinct groups based on 5 quantitative traits (Fig. 2). The descriptive statistics of the distinct clusters are presented in Table 3. The number of landraces belonging to distinct clusters varied from 1 in Clusters IV and V each to 15 in Cluster I. Cluster I was the largest cluster comprising of 15 landraces (60%) followed by Cluster II with 6 landraces (24%). The landraces embraced in Cluster I depicted higher mean values for traits such as grain yield, plant height and moderate mean values for other agronomic traits. Similarly, the members of Cluster II portrayed relatively lower mean values for grain yield and plant height while moderate mean values for other agronomic traits. Clusters III and IV consisted of 1 landrace each while Cluster V contained 2 landraces. Of all, Cluster IV consisting of only landrace, Lamjung Local revealed agronomic superiority entailing relatively higher grain yield (3610 kgha −1 ) and lower values for phenological traits. Lamjung Local transcended all other landraces in Lamjung conditions for having preeminence of yield and earliness (Table 4). Earliness is one of the crucial traits that is highly valued by farmers dwelling in dry and rain-fed conditions in the mountains that is useful to escape drought, hailstone damage and terminal heat stress. In addition to this, the cultivation of early cultivars of naked barley ensures timely planting of rice crop and having rice intact in the cropping pattern. Women prefer early cultivars as it ensures that the food is available during lean period. The selection of the breeding material from cluster I holds more reliable and rational provided that the mean values of the quantitative traits for the cluster member across the multi-locations were used in the study. While considering the multienvironment trial data, the mean overall performance of landraces in Cluster I was appreciably better than the others. The UPGMA clustering of the landraces revealed ample variation for the agronomic traits and facilitated the selection of preferable landraces for pre-breeding and further evaluation. Several workers have deployed cluster analysis to depict the heterogeneity and similarity among the accessions and interrelationship among characters [7,16,36]. The analysis also revealed that the landrace aggregation was independent of the geographic source of collection, which is similar to the findings of Zakova and Benkova [37], Abebe et al. [1] and Mekonnon et al. [16]. Characterization of accessions and clustering of them on the basis of their morphological and genetic similarity aids in identification and selection of the best parents for hybridization [30]. Hence, grouping of landraces using multivariate analysis such as UPGMA clustering would be valuable for the breeders in such a way that the most promising landraces in the population may be selected from different clusters for crop improvement.

Pattern and structure of polymorphism in qualitative traits
Similar multivariate analyses as employed in the former section, viz. UPGMA and PCA, were performed to uncover major qualitative traits contributing to the diversity of the landrace. The qualitative traits taken into account for analyses were growth habitat, stem pigmentation, tillering capacity, spike density, lemma awn or hoods, lemma awn barbs, glume color, grain color, grain size and overall phenotypic acceptance. The UPGMA cluster analysis based on similarity distance generated 5 distinct clusters with similar patterns of qualitative traits in each ( Fig. 3 and Table 6). The number of landraces belonging to each cluster varied from 1 in Cluster II and IV each to 16 in Cluster I. Cluster I (64%), the largest of all, containing 16 landraces predominantly delineated erect growth habit, green and purple stem pigmentation, high tillering capacity and intermediate grain size with yellow glume color. Cluster III consisting of 5 landraces was principally characterized with purple stem pigmentation, low tillering and black/purple pericarp. Cluster V comprising 2 landraces demonstrated purple stem pigmentation, low tillering and small grain size. Similarly, Clusters II and IV comprising 1 landrace each were distinguished by revealing a unique set of quantitative traits as illustrated in Table 6. Although the cluster analysis grouped the naked barley landraces with greater phenotypic similarity, the cluster did not necessarily contain all the landraces from the same or adjacent niche/source of collection. This finding is in harmony with Dejene   [16], who reported that the clustering of accessions based on the agronomic characters revealed no distinct regional grouping patterns in which accessions from same or adjacent regions appeared in different clusters. The discrimination of landraces into discrete clusters concluded preponderance of polymorphism among the germplasm of Nepalese naked barley landraces for qualitative traits. Additionally, a good fit between the graphical representation of distances and their similarity matrices was noticed in the dendrogram (Fig. 3). Knowledge on degree of genetic relatedness among the genotypes provides robust information to address breeding programs and genetic resource management [27].
The PCA based on the qualitative traits revealed that the first four PCs with Eigen value ≥ 1 accounted for 69.80% of the total variance (Table 7). For the first PC accounting for 27.5% of the total variance, the traits such as grain color, overall phenotypic acceptance, lemma awn/hood and stem pigmentation posed the largest contribution in the positive direction while lemma awn barbs and growth habit loaded heavily in the negative direction. The second PC accounting for 18.3% of the total variance was mainly influenced by growth habit, tillering capacity and lemma awn barbs while the traits loading heavily on the third PC (13.7%) were spike density and stem pigmentation. Similarly, the fourth PC accounted for 10.30% of the total variance for which stem pigmentation and growth habit posed the largest load. The comparison between the clusters and groups developed from PCA allowed us to deduce that the set of qualitative traits contributing substantially to divergence in naked barley landraces included grain color, overall phenotypic acceptance, lemma awn/hood and lemma awn barbs.

Estimate of phenotypic diversity
Substantial polymorphism was revealed across all the qualitative traits with varying degrees of contribution to the phenotypic diversity. Descriptor states and their frequency as well as Shannon-Weaver diversity indices for each qualitative trait are depicted in Table 8. The diversity index (H′) ranged from 0.32 to 0.99 with a mean value of 0.73, inferring tremendous diversity in the collections for the qualitative traits. The estimate of Shannon-Weaver diversity indices (H′) considers both richness and evenness of the phenotypic classes of the traits. High H′ was inferred for most of the traits such as grain size (H′ = 0.99), phenotypic acceptance (H′ = 0.99), spike density (H′ = 0.94), stem pigmentation (H′ = 0.92), growth habit (H′ = 0.89) and tillering capacity (H′ = 0.83) while moderate values of H′ were revealed for lemma awn barbs (H′ = 0.62) and glume color (H = 0.53). However, relatively lower values of H′ were found for pericarp color (H′ = 0.37) and lemma awn/hood (H′ = 0.32). The notable phenotypic diversity in Nepalese naked barley landraces may be attributed to high agro-ecological heterogeneity, propitious for the cultivation of crops in Nepal. In addition to this, the findings showed that the structure of morphological variation in Nepalese naked barley landraces is significantly influenced by natural selection factors. Tremendous agro-morphological variation in Nepalese barley landraces has been reported by several authors [10,17]. The existence of high phenotypic diversity in the Nepalese naked barley landrace increases the space for selection and pre-breeding purpose of the elite varieties with earnest emphasis to the farmers' preferences of the decisive morphological traits in the selection and adoption process. Morphological traits have been extensively taken into account by farmers inter alia to discriminate varieties regarding selection and adoption of a particular variety [17]. High mean value of Shannon-Weaver index (H′ = 0.74) for the qualitative traits confirmed huge genetic diversity in the collection and more or less balanced frequency classes for individual traits with implications for utilization in robust crop improvement and enhancement of genetic potential of naked barley.

Conclusion
The present study engendered clear insights on the pattern and structure of agro-morphological and phenotypic diversity in collection of 25 Nepalese naked barley landraces, which is very crucial for efficient pre-breeding, management and utilization in the crop improvement program and enhancement of genetic potential of The diverse landraces which were identified as superior in various agronomic traits in the present study could be effectively utilized as genitors in hulless barley breeding. Furthermore, the information we generated complements the robust breeding program of highyielding and climate-resilient varieties of end users' preferences and bolsters employment of innovative and proven participatory plant breeding approaches using diversity kits and IRD kits to expand and promote the varietal choice options for expeditious benefits and impact, fostering agro-biodiversity and food security in the high mountains.