Evaluation and identification of wild lentil accessions for enhancing genetic gains of cultivated varieties

Domesticated lentil has a relatively narrow genetic base globally and most released varieties are susceptible to severe biotic and abiotic stresses. The crop wild relatives could provide new traits of interest for tailoring novel germplasm and cultivated lentil improvement. The primary objective of this study was to evaluate wild lentil accessions for identification of economically viable agro-morphological traits and resistance against major biotic stresses. The study has revealed substantial variations in seed yield and its important component characters. Further, the diversity analysis of wild accessions showed two major clusters which were bifurcated into sub-clusters, thereby suggesting their wider genetic divergence. However, principal component analysis exhibited that seed yield plant-1, number of seeds plant-1, number of pods plant-1, harvest index and biological yield plant-1 contributed significantly to the total genetic variation assessed in wild lentil taxa. Moreover, some of the wild accessions collected from Syria and Turkey regions showed resistance against more than one disease indicating rich diversity of lentil genetic resources. The identification of most promising genotypes carrying resistance against major biotic stresses could be utilized in the cultivated or susceptible varieties of lentil for enhancing genetic gains. The study has also identified some trait specific accessions, which could also be taken into the consideration while planning distant hybridization in lentil.


Introduction
Lentil is a self-pollinating true diploid (2n = 2x = 14) grain legume crop having genome size of 4063 Mb [1]. The cultivated lentil species (Lens culinaris) encompasses two groups, established on the basis of distinct morphological characters, the small-seeded (microsperma) and largeseeded (macrosperma) [2]. The production and productivity of lentil has increased from an average yield of 565 kg ha -1 in 1961-63 to 1153 kg ha -1 during 2016-2017 [3]. Despite the a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 tremendous improvement, the current lentil yield is much below as compared to other pulses. In India, most of the lentil varieties which were developed through pure line selection showed low yield potential. In cultivated lentil varieties, the yield limiting factors are lack of seedling vigour, very high rate flower drop and low pod setting, lack of lodging resistance and exposure to major biotic and abiotic stresses [4]. The pedigree analysis of some important released varieties exhibited that a few donors have contributed substantially into the background of those varieties [5]. Therefore, to attain further breakthrough for enhancing genetic gains, new target traits are needed to be identified and introgressed into cultivated gene pool for widening the genetic base of cultigens. Crop wild relatives (CWRs) are an invaluable reservoir of productivity enhancement related characters having resilience to climate change and farming system, and are source of novel traits [6][7]. The global wild lentil germplasm introduced from the International Centre for Agricultural Research in Dry Areas (ICARDA), has been multiplied, characterized and evaluated against target characters [8]. The promising 96 wild lentil accessions selected from 405 global wild collections were validated under multilocational evaluation for identifying stable donors against the target traits. The present study was therefore, undertaken (1) to evaluate promising wild lentil accessions for various target agronomical characters and major biotic stresses and (2) to identify the potential accessions (gene sources) carrying important traits for enhancing genetic gains of cultivated gene pool.

Agronomic evaluation
All 96 wild lentil accessions were evaluated under two agro-ecological locations viz. Experimental Farm of NBPGR Regional Station, Shimla (28 0  , number of pods plant -1 , number of seeds plant -1 , number of seeds pod -1 , 100-seed weight, seed yield plant -1 , and biological yield plant -1 ) in both locations. Further, numerical data was subjected to statistical analysis using SAS software [9]. The coefficient of variation for various characters was determined using the formulae of Burton [10] and broad sense heritability was also calculated as per the method suggested by Lush [11]. The expected genetic advance was performed as per the procedure of Johnson et al. [12].

Phenotypic diversity and association study
The phenotypic diversity analysis was carried out using quantitative data of 96 promising accessions of wild lentil. Euclidean distances were calculated using quantitative data which was used to establish relationship between interspecific accessions. DARwin 5.0 software of Perrier et al. [13] was used for hierarchical clustering of accessions using UPGMA mode based upon Jaccard's coefficient. Correlation among various parameters was determined using SAS software [9]. The test accessions were frequently inoculated by spraying suspension of local isolate of U. viciae fabae (1x10 6 spores/ml). The data were recorded using 1-9 rating scales at vegetative and reproductive stages as followed by Mayee and Datar [14].

Screening against rust, powdery mildew and Fusarium wilt
Powdery mildew (Erysiphe trifolii). Another experiment was undertaken at CSKHPKV Research and Extension Centre, Dhaulakuan during winter season of 2014-2015, 2015-2016 and 2016-2017. The plants of test accessions were raised in plastic pots in a mixture of soil and farmyard manure mixture (10:1) under green house conditions. The disease symptoms started developing at flower initiation stage and susceptible checks developed disease earlier and faster. The pots with heavily infected susceptible plants were transferred near the test accessions raised in the green house. The heavily infected plants were shaken with the wooden stick to dislodge the conidia which dispersed and infected the healthy plant parts of the test accessions as also suggested by Tiwari et al. [15]. The full infection of susceptible accessions was observed near green pod stage and scorings were done at different intervals using 1-9 rating scales [15].  [16][17][18][19]. The performance of disease pressure was compared with the resistant (PL639) and susceptible (L-9-12) checks from India, which were repeated after 15 accessions of each replication. The data were taken using 1-9 rating scales as followed by Bayaa et al. [17,19].

Diversity for quantitative characters
The analysis of variance revealed that the wild lentil accessions differed significantly in terms of seed yield plant -1 and its important component characters (significant at p = 0.05) and also manifested by the range, mean and coefficient of variation for all the metric characters assessed. However, the frequency distribution of important quantitative characters depicted in

Diversity analysis
Wild lentil accessions formed two major clusters I and II in hierarchical clustering as shown in Fig 2. Major cluster I grouped only three accessions including ILWL418 (L. ervoides) from  and ILWL65 of L. ervoides including first from Montenegro and other two from Turkey; and ILWL20 of L. odemensis from Palestine. Cluster B was occupied by two sub clusters BI and BII, each of which were further divided into two groups a, b and c, d, respectively. Group d further formed outgroups (d1 and d2) and finally clutches (d2a and d2b). As Cluster II grouped subclusters were occupied by accessions from different wild lentil species belonging to different countries of origin, they showed highest diversity (S1 Table). Euclidean dissimilarity matrices varied from 6.01 between ILWL429 (L. lamottei) from Spain and ILWL90 (L. tomentosus)

PLOS ONE
Evaluation of wild Lens taxa from Turkey to 2156.52 between ILWL18 (L. nigricans) from France and ILWL476 (L. orientalis) from Turkey (S1 Table). First two principal coordinates in factorial analysis explained 99.07% and 0.52% variability, respectively. In both type of groupings, the role of geographical or country of origin of accessions was not observed.

Correlations among different parameters
The correlation coefficient among different agro-morphological traits of wild lentil accessions is shown in Table 2. Correlation indices highlighted a number of significant inter-relationships (both +ve and -ve) among studied traits. In the present study, significant positive correlation was observed between days to flowering and maturity (>0.4 r-value). Similarly, significant positive correlations were obtained between seed yield plant -1 , biological yield plant -1 and harvest index. These traits also showed positive correlations with plant height (r 2 <4.0 in case of biological yield plant -1 ), number of branches plant -1 (r 2 <4.0 in case of harvest index), number of pods plant -1 and seeds plant -1 . Seeds plant -1 showed significant correlation with plant height, branches plant -1 and pods plant -1 . The number of pods plant -1 exhibited significant positive correlation with branches plant -1 . Significant negative correlation was observed between days to flowering and seed yield plant -1 , seeds plant -1 , pods plant -1 , and harvest index.

Principle component analysis
Maximum variability (66.31%) was explained by only three components with Eigenvalue >1.0 (Table 3). Of the total variation, 42.37% was exhibited by PC1, and the contributing traits with high coefficients included seed yield plant -1 , seeds plant -1 , pods plant -1 , harvest index and biological yield plant -1 . Similarly, 15.04% variability was explained by PC2, and the contributing traits with high coefficients included days to maturity, days to flowering, rust incidence and plant height. Similarly, PC3 contributed 8.89% of overall variability, and the traits contributing to diversity included 100-seed weight, Fusarium wilt and Powdery mildew severity. Fig 3A  and 3B shows the two dimensional variability pattern and contributing traits towards variability, respectively.   lamottei with disease reaction 2 were found to be resistant. The remaining accessions were either moderately resistant or susceptible against the pathogen (Fig 5).

Discussion
The precise evaluation of crop wild relatives is pre-requisite to identify target traits of interest followed by their introgression into the background of cultivated varieties for enhancing genetic gains [20][21][22]. A wide range of variation in wild lentil accessions was observed against the target traits viz. earliness, high pod number and seed yield suggesting diverse genetic makeup and geographical origins of wild lentil collections. These promising accessions belonging to different taxa and ecological niches could be useful resource for enhancing genetic gains of cultivated varieties. Accessions ILWL 18 and ILWL 19 of L. nigricans were found promising for high seed yield plant -1 , suggesting their broader genetic base as also reflected in clustering pattern of wild accessions based on quantitative data analysis. Euclidean dissimilarity matrices of wild lentil accessions ranged from 6.01 to 2156.52, which revealed very high variability among interspecific lentil taxa. The observed variability could be due to the presence of diverse origins of these wild accessions. Among the different countries of origin, accessions from Turkey showed maximum variability for the studied traits suggesting more exploration and collection of lentil germplasm. Viera et al. [23] characterized wheat germplasm based on phenotypic traits and obtained distances up to 196.61. The report of existing variability among germplasm accessions provides an idea about the expected heterosis for starting breeding programme. A broad range of Euclidean distances obtained in the present study shows the importance of wild lentil accessions as a source of diverse heterotic material [24]. However, the diversity pattern further suggested that no role was played by the geographical distributions of global accessions in their grouping. This grouping pattern can be described on the basis of different genetic constitution of accessions. Further, it might be due to continuous gene flow occurrence among wild lentil taxa [8]. No role of geographical distributions was observed in grouping pattern of accessions clustered on the basis of morphological, biochemical and molecular markers in guar [25]. Similar observation in wild lentil accessions was also reported by Singh et al. [8] and Kumar et al. [26]. Further, significant positive correlation of seed yield plant -1 with seeds plant -1 , pods plant -1 , branches plant -1 was observed in lentil accessions and similar correlations have also been reported by other workers [27][28][29]. They reported that seeds plant -1 , biological yield plant -1 , pods plant -1 , and harvest index are the important yield component traits in lentil contributing towards yield enhancement. Such correlations are useful for lentil genetic improvement and breeding programmes. These observations were found to be in accordance with the results of Toklu et al. [29] and Abo-Hegazy et al. [30]. The traits which contributed heavily towards variability included seed yield plant -1 , seeds pod -1 , pods plant -1 , biological yield plant -1 and harvest index. These findings highlighted the significance of yield related component characters generating substantial variations in wild lentil accessions.
Furthermore, wild taxa of lentil are known to be resistant against major biotic and abiotic stresses [8]. They are the carrier of important characters and have potential value in diversification of cultivated gene pool for enhancing genetic gains. Some wild lentil accessions revealed resistant disease reaction against more than one disease (Table 5) viz. accessions ILWL230 and ILWL476 of L. orientalis (rust and powdery mildew); accessions ILWL9 and ILWL37 of L. nigricans (rust and powdery mildew); accession IG136639 of L. ervoides (powdery mildew and Fusarium wilt) and accession ILWL308 of L. tomentosus (rust and Fusarium wilt). It is important to mention that the northwestern part has been the traditional lentil cultivation area in India. But in late 1980s, among various factors including the epidemics of rust, powdery mildew and Fusarium wilt resulted in drastic reduction of lentil area and production as well. In this context, gene sources identified against rust, powdery mildew and Fusarium wilt in the present study can prove to be useful donors which could be exploited through pre-breeding with the aim to develop usable germplasm for developing resistant lentil varieties, especially for the epidemiologically important Indian conditions. Furthermore, the development of lentil cultivars with combined resistance against major biotic stresses may help in the development of appropriate management strategies in future in conjunction with allelism studies to identify and pyramid novel genes for resistance breeding of lentil.

Conclusions
During the past few years, little efforts were made on evaluation of wild lentil genetic resources in India. However, this is the first attempt on characterization and evaluation of all available global wild lentil germplasm against the target traits of interest under diverse agro-ecological conditions. Further, the study has helped in identifying the confirmed and stable gene sources (donors) across intra and interspecific accessions viz. ILWL203 (EC789113) of L. odemensis for rust and high pod number; ILWL230 (EC718515), ILWL476 (EC718605) of L. orientalis for rust and powdery mildew; ILWL191 (EC718682), ILWL9 (EC718236), and ILWL37 (EC718257) of L. nigricans for rust and powdery mildew, IG136639 of L. ervoides for powdery mildew and Fusarium wilt and ILWL308 (EC728782) of L. tomentosus for rust and Fusarium wilt. These gene sources could be taken in lentil wide hybridization programme for enhancing genetic gains of cultivated varieties and could also be shared among lentil breeders/researchers in the country or elsewhere under Standard Material Transfer Agreement (SMTA) for strengthening the on-going lentil genetic improvement programmes.
Supporting information S1 Table. Clustering pattern of wild lentil core accessions based on agro-morphological characters. (DOCX)