The assessment of genetic diversity for traits of interest is a critical step in the breeding process. National and international gene banks harbor diverse germplasm collections that are inexpensive source of beneficial alleles of agronomic and economic significances. These genetic resources are used in breeding programs to develop new varieties having superiors traits such as high-productivity and nutritional quality, and tolerance or resistance to abiotic/biotic stresses. In lentil, up to 58,000 germplasm accessions are stored in different gene banks around the globe (Khazaei et al. 2016). The largest collection of lentil germplasm (14,597 accessions) is stored in the International Center for Agricultural Research in the Dry Areas (ICARDA).
In this study, we assessed the genetic variability in a Mediterranean lentil collection of 119 accessions including landraces (from Morocco, Turkey, Italy, and Greece), advanced breeding lines, local cultivars, and improved cultivars. The accessions were grown in four trials × growing season environments and screened for four important agronomic traits namely days to 50% flowering, plant height, hundred-seed weight, and grain yield. The results unveiled significant variation among Mediterranean lentil collection accessions for traits examined, highlighting the possibility for the selection of promising genotypes that may be exploited in lentil breeding program. Similarly, previous studies reported considerable genetic variation for agronomic and grain quality traits as well as adaptation to biotic and abiotic stresses in different lentil germplasm (Erskine 1983; Tullu et al. 2001; Fernández-Aparicio et al. 2009; Choukri et al. 2020, 2022; El Haddad et al. 2021). The Mediterranean region holds an important lentil genetic diversity thanks to the history of domestication and cultivation and the frequency of abiotic and biotic stress that act as selection pressure (Idrissi et al., 2016). Hence, it can be expected that genetic resources from this region may offer useful adaptive traits for lentil breeding programs. Several studies reported considerable genetic diversity among lentil accessions from Mediterranean region using agro-morphological and phenological traits (Toklu et al. 2009; Bacchi et al. 2010; Torricelli et al. 2012; Zaccardelli et al. 2012; Baggar et al. 2023a), and an arsenal of molecular markers (Lombardi et al. 2014; Idrissi et al. 2015, 2016, 2018; Khazaei et al. 2016; Mbasani-Mansi et al. 2019).
Genetic variability for relevant agronomic traits in different environments
Flowering time is an important trait for lentil adaptation to different agro-ecological conditions. In this study, the results showed significant effect of genotype on time to 50% flowering in all trials except at Adana during the 2021 season (Table S4). In addition, significant effects of genotype, environment, and genotype × environment interaction on this trait were obtained (Table S7). These results are consistent with those reported in previous studies (Bermejo et al. 2020; Hossain et al. 2023). The mean of days to 50% flowering varied from 87 days (Adana in the 2021 season) to 111 days (Adana during the 2022 season). Over environments, the lowest value was recorded at Adana during 2022 season (61 days); while the highest values were observed at Sidi El Aidi under no-till (153 days). Previously, ranges of 105–125 days (Erskine and Goodrich 1988), 98–118 days (Erskine and Goodrich 1991), 83–120 days (Toklu et al. 2009) 86–112 days (Alghamdi et al. 2013), and 73–85 days (Sharma et al. 2022) have been reported in different environments. In lentil, temperature and photoperiod are the major factors determining flowering time (Summerfield et al. 1985; Erskine et al. 1990). In addition, light quality can also influence the time to flowering in lentil with a specie- and genotype-dependent manner (Yuan et al. 2017); in fact, flowering time of some lentil wild accessions was less affected by the variation in light quality as compared to cultivated genotypes (Yuan et al. 2017).
Concerning the plant height, the analysis of variance in each environment indicated significant differences between accessions at Adana during 2021 season, while in others trials the genotype had no significant effect (Table S4). Across environments, there were significant differences between accessions and environments, but the genotype × environment interaction was not significant (Table S7). Another study found significant effect of genotype, season, and their interaction on plant height in lentil (Sharma et al. 2022). Furthermore, it was noted that plant height is influenced by genotype, environment and their interaction (Bermejo et al. 2020). In another investigation, the plant height in lentil crop was significantly varied according to genotype and environment, while the genotype × environment interaction was not significant (Balech et al. 2023). On average the plant height was higher in Adana during both seasons compared to Sidi El Aidi under both tillage systems. Over environments, the minimum value was observed at Sidi El Aidi under conventional tillage (19.54 cm) and the maximum value was observed at Adana during the 2022 season (55.84 cm). In lentil, earlier studies reported diverse ranges of plant height such as 10–45 cm and 18–40 cm (Hamdi et al. 1991), 26.7–38.8 cm (Erskine and Goodrich 1991), and 14–52 cm (El haddad et al. 2020).
Hundred-seed weight has great economic significance because it influences the market value of the lentil. In the present study, the effect of genotype on hundred-seed weight was significant in all environments with the exception of Adana during 2022 season (Table S4). Combined analysis over environments demonstrated highly significant effects of genotype and environment, whereas the genotype × environment interaction was not significant (Table S7). In accordance with these results, previous studies reported the variation of hundred-seed weight according the genotype and environmental conditions (El haddad et al. 2020; Choukri et al. 2020). In addition, others investigations, as in the present study, indicated that hundred-seed weight in lentil is not significantly affected by genotype × environment interaction (Abo-Hegazy et al. 2013); however, others researches reported that this effect was statistically significant (Bermejo et al. 2020; Baggar et al. 2023a). The highest average value of hundred-seed weight was observed at Sidi El Aidi under no-till and the lowest average value was observed at Adana during the 2022 season. Across environments, the minimum hundred-seed weight (1.56 g) was obtained at Adana during the 2022 season while the maximum value (6.47 g) was recorded at Adana during the 2021 season (Table S5). Prior investigations in lentil documented wide ranges of hundred-seed weight including 1.37–7.4 g (Tullu et al. 2001), 1.68–4.03 g (Toklu et al. 2009), 1.7–4.8 g (Choukri et al. 2020), and 1.93–4.97 g (El Haddad et al. 2021). Additionally, significant variation has been reported for hundred-seed weight between and within yellow and red-cotyledon types, with ranges of 1.7–7.4 g for yellow-cotyledon types and 1.3–5.2 g for red-cotyledon types (Tullu et al. 2001).
In the present study, grain yield was significantly influenced by the genotype at Sidi El Aidi under both tillage systems, while this effect was not significant at Adana in both seasons (Table S4). Combined analysis over environments revealed that grain yield was significantly affected by genotype, environment, and their interaction (Table S7); in the same sense similar results have been reported previously (Mohebodini et al. 2006; Sabaghnia et al. 2008; Mohammed et al. 2016; Abbas et al. 2019; Idrissi et al. 2019; Chen et al. 2022; Baggar et al. 2023b, a; Ghaffar et al. 2023; Hossain et al. 2023).
On average Adana during the 2021 season produced the highest yield followed by Sidi El Aidi under conventional tillage, while the lowest average yield was obtained at Adana during the 2022 season and this can be explained by the variability in weather conditions between environments; in fact, at Adana during the 2022 season the trial experienced cold stress which negatively impacted the performance of the accessions. Low temperature observed in January 2022 (2.06°C) and March 2022 (3.94°C) significantly disturbed the normal growth and development of plants, leading to significant yield reduction. In March 2022, the flowering stage of the majority of accessions was coincided with cold stress which hampered the normal development of flower and pod formation. In lentil, it has been reported that cold stress at flowering stage induces flower abortion, damages the vegetative tissue, and alters the development of seed, thereby compromising yield formation (Gupta et al. 2019). Furthermore, exposure to frost stress may raise the sensitivity to pathogens infections (Gupta et al. 2019). In food legumes, including lentil, exposure to cold stress may reduce significantly yield and quality of harvest or even lead to complete crop failure (Bhat et al. 2022). Overall, cold stress (i.e., chilling: 0°C to 15°C; freezing: <0°C) is adversely affecting the normal growth and development of plants as well as their geographical distribution (Ding et al. 2019). Cold stress reduces crop productivity by altering various physiological, biochemical and molecular processes (Raza et al. 2023); however, plants have developed sophisticated mechanisms at biochemical and molecular level that help them cope with adverse effects induced by cold stress (Ding et al. 2019).
Cold tolerance is an important trait to increase lentil production under winter sowing and in highland regions. In highlands of Central and West Asia and North Africa (CWANA) region, lentil crop experiences cold stress at seedling stage and its productivity in such a region could be increased by the development of cold/frost tolerant, winter-hardy cultivars (Kumar et al. 2013). Interestingly, the screening of 3,592 lentil accessions under field conditions in Central Anatolia allowed the identification of up to 238 cold-tolerant accessions which could be used to develop winter-hardy varieties (Erskine et al. 1981). Additionally, sources of winter hardiness have been identified among lentil wild relatives (Hamdi et al. 1996). Several cultivars with improved winter hardiness such as ‘Kafkas’ in Turkey, ‘Morton’ in USA, ‘Gachsaran’ in Iran, ‘Shiraz-96’ in Pakistan, and ‘Bichette’ in Morocco have been released (Sakr et al. 2004; Sarker and Erskine 2006; Aydogan et al. 2007; Muehlbauer and McPhee 2007; Sabaghpour et al. 2007; Sarker et al. 2009; Kumar et al. 2013). In a recent study conducted in Morocco across nine contrasted environments (combination of three locations and three growing years), the cultivar ‘Bichette’ showed adaptation to low temperature conditions (Benbrahim et al. 2021). ‘Bichette’ is a single plant selection from a Jordanian landrace ‘76TA 66005’ (Sakr et al. 2004; Idrissi et al. 2019); therefore, landraces may constitute an important source of useful traits enabling the development of improved cultivars with high productivity that can prosper in winter conditions. These genetic resources might provide a starting point for breeding programs targeting winter hardiness as well as others traits of interest.
Phenotypic diversity of accessions according to geographical origin
Considering the performance of accessions as function of their country of origin, the results indicated that the effect of origin varied depending of the environment and the trait concerned (Figs. 2 and 3; Table S6). The highest effect of origin was observed in the time to 50% flowering across all environments (Figs. 2 and 3; Table S6). Plant height was also affected by origin in all environments; hundred-seed weight was affected by origin in all environments, except at Adana in the 2022 season, while the effect of origin on grain yield was significant only at Sidi El Aidi under both tillage systems. In the same way, highly significant effect of country of origin on different agro-morphological traits, including time to 50% flowering, plant height, hundred-seed weight and seed yield, was observed in 615 lentil accessions from 13 lentil-producing countries (Erskine et al. 1989). Furthermore, significant effect of geographical origin on different phenological and morphological traits was observed in a lentil core collection of 287 accessions across two years under no-till and conventional tillage conditions (Tullu et al. 2001). However, in the present study, non-significant effect of origin, for yield in Adana during both seasons and for hundred-seed weight and yield in Adana during the 2022 season, can be explained by differences of the climatic conditions characterizing each growing season that have important effects on the performance of each accessions and, consequently, the average value of each origin. Likewise, the agronomic assessment of 25 lentil accessions, including landraces from Mediterranean Basin, indicated no relation between grain yield and geographic origins (Bacchi et al. 2010). Recently, the phenotypic evaluation of 986 accessions of mungbean and urdbean revealed significant effect of geographic origin on several phenological and agro-morphological traits (Burlyaeva et al. 2019). On micro-geographical level the examination of the genetic variation of morphological traits in 156 lentil landraces from different regions of Ethiopia revealed significant differences between geographic regions for the time to 50% flowering and maturity, plant height, number of seeds per pod and hundred-seed weight (Bejiga et al. 1996). On the other hand, non-significant correlations were observed between altitude of origin and agro-morphological traits in lentil landraces from Ethiopia and Yemen, except hundred-seed weight in Ethiopia (Erskine and Choudhary 1986; Bejiga et al. 1996). The evaluation of the genetic diversity and relationships between Moroccan lentil landraces showed no clear separation of landraces according to their geographic origins; however clear differentiation was obtained according to agro-environmental origin (i.e., dry area, favorable area and highlands) which highlights the importance of specific adaptation to agro-climatic conditions in the genetic differentiation among lentil landraces (Idrissi et al. 2015).
On average, Moroccan landraces were the earliest to flower in all environments (Fig. 2 and Fig. 3) as compared to landraces from Italy, Greece and Turkey. The same pattern was reported previously (Idrissi et al. 2018). It has been well established that phenology has played a pivotal role in the adaption of lentil to different agro-ecological regions after the spread from its center of origin (Erskine et al. 1989). The dissemination of lentil from its center of origin to new agro-ecological environments has been accompanied by selection for an appropriate phenological response to regionally-specific balance between photoperiod and temperature (Erskine et al. 1994; Neupane et al. 2023). The response of flowering time in lentil to photoperiod has proven to be related to latitude of origin (Erskine et al. 1990). The spread of lentil from its center of origin has resulted in reduced photoperiod sensitivity but increased temperature sensitivity (Erskine et al. 1994; Erskine 1997). Generally, lentil is sown in winter in the Mediterranean environments; however there is phenological variation according to the region. In fact, in Morocco late autumn-early winter sowing, characterized by cool temperatures and short day followed by an increase in the temperature and day length during spring, is the most frequent; however in northern Mediterranean region there is a tendency towards late winter-early spring sowing which is characterized by relatively warm temperatures and increased day length (Idrissi et al. 2018). In the same country of origin different growing regions can be present with different photoperiod and temperature regimes. Additionally, both early and later-flowering were found among genotype from different Mediterranean countries (Neupane et al. 2021). Early-flowering observed in Moroccan landraces is the result of successive selection for avoidance of terminal drought and heat which are considered as important limiting factors for lentil production in Morocco. Interestingly, advanced lines, selected under Moroccan climatic conditions, were the earliest compared to landraces in all environments, highlighting the efforts deployed in Morocco to improve lentil adaptation to drought and heat stress. In fact, it is well known that early phenology is an important trait to escape the end-cycle drought and heat stress (Erskine et al. 1993; Sarker et al. 2005; El haddad et al. 2020).
On average, advanced lines exceeded yield of landraces in all environments, except at Adana during the 2022 season; in the latter the highest mean grain yield was obtained in Turkish landraces. Importantly, Moroccan landraces outperformed others Mediterranean landraces in terms of grain yield in all environments with the exception of Adana during the second season where they were slightly surpassed by Turkish landraces (although no significant effect of origin was observed on grain yield at Adana during 2022 growing season). These results highlight the adaptation of Turkish landraces to climatic conditions as those observed in Adana in the 2022 season characterized by cold stress that adversely impacted the normal growth and development of the crop, which has led to significant yield reduction in comparison with others environments (Table S5). In fact, superior cold tolerance was reported in landraces from Chile, Turkey, Syria and Greece which can be attributed to natural and artificial selection for winter hardness in these countries (Erskine et al. 1981; Erskine 1997). However, when considering the performance of accessions regardless of their geographic origins Moroccan landraces were ranked among the top performing at Adana during 2022 season, as well as in the 2021 season, which clearly illustrates that they have also acquired a certain level of tolerance to cold stress in their growing areas. This is consistent with an earlier study indicating the differentiation of Moroccan landraces according to their agro-environmental origins with distinction of landraces from highland regions of Morocco, medium Atlas mountains, which are characterized by the occurrence of cold stress (Idrissi et al. 2015). Furthermore, there is evidence that Moroccan lentil landraces from highland zones and favorable areas may potentially share proportions of the genome with landraces from northern Mediterranean (i.e. Turkey, Italy and Greece) (Idrissi et al. 2018).
On the other hand, Turkish and Greek landraces displayed the highest mean values of plant height and hundred-seed weight, respectively, in all environments. Plant height assessed in 2,895 lentil accessions from different countries ranged from 10 to 45 cm and accession from Egypt, Greece and Turkey showed higher plant height compared to average value (Solh and Erskine 1984). In a recent study, Moroccan lentil landraces showed higher plant height compared to others accessions from 14 different countries (Preiti et al. 2024). Plant height is an important trait for selecting lentil accessions suitable for mechanical harvesting; it has been reported that plant height and time to maturity are positively associated with lodging in lentil crop (Erskine and Goodrich 1988); therefore, these traits could be taken consideration when selection for lodging resistance. On the other hand, hundred-seed weight is an important trait of significant market value in lentil crop; it is mostly affected by human-mediated selection processes and with low adaptive value (Erskine 1997).
Relationships between traits
The type and magnitude of correlation between traits of interest can inform on the selection strategy in a breeding program. In the present study, the type and magnitude of correlation varied according to environmental conditions (Fig. 7). Grain yield showed highly significant negative correlation with days to 50% flowering (Sidi El Aidi under both tillage systems and Adana during the 2021 season) and negative, yet not significant, correlation at Adana during the 2022 season. In agreement with these results, a previous study conducted in two locations in Morocco reported correlation with different type and magnitude in different conditions (El haddad et al. 2020); in fact, statistically significant negative, non-significant negative, and non-significant positive correlations between grain yield and time to 50% flowering were observed under normal, heat stress, and combined drought-heat stress, respectively (El haddad et al. 2020). Early flowering is an important trait that enables the adaptation to Mediterranean-type environment which is characterized by terminal drought and heat stress; this was confirmed by strong association of early flowering and yield in stressful environments (Lake and Sadras 2021). Grain yield exhibited highly significant positive correlation with hundred-seed weight at Sidi El Aidi under conventional tillage; however no-significant correlations were observed in others environments (Fig. 7). This is in accordance with results reported previously in lentil (El haddad et al. 2020); however, others studies reported strong positive association between grain yield and hundred-seed weight under normal, heat stress, and combined drought-heat stress conditions (Choukri et al. 2020, 2022). Plant height exhibited significant positive correlation with grain yield at Sidi El Aidi under no-till system and at Adana during the 2021 growing season and positive, but not significant, correlation in others environments (Fig. 7). In agreement with these results positive (El haddad et al. 2020; Choukri et al. 2020; Naik et al. 2024) as well as no-significant correlations (El haddad et al. 2020; Choukri et al. 2020) were reported.
Principal Component Analysis and clustering patterns of accessions based on agronomic traits
Results of Principal Component Analysis (PCA) (Fig. 8) and cluster analysis (Figs. 9 and 10) confirmed the wide genetic variability between accessions phenotyped in the present study in different environments and enabled the investigation of traits contributing to variation and the designation of clusters or groups with genetically similar accessions, thus increasing the likelihood to identify promising accessions to be included in crossing blocks in breeding programs. The accessions were grouped into four clusters in different environments. Generally, there was no clear relationship between clustering patterns and geographical origin of landraces as landraces from different countries were grouped in the same cluster in different environment which is in agreement with other reports in common bean (Rana et al. 2015), faba bean (Karaköy et al. 2014), and wheat (Mohammadi and Amri 2022). Furthermore, it has been observed that drought tolerance was not related to the origin of lentil landraces, indicating that selection based on the performance of individual accessions, irrespective of their geographical origin, may prove effective (Idrissi et al. 2016). In the first detailed assessment of variation in cultivated lentil two subspecies were considered, the small-seeded (microsperma) and large-seeded (macrosperma) (Barulina 1930); furthermore, lentil germplasm was classified into regional groups (or grex) based on several qualitative traits such as pubescence, number of flowers per peduncle, flower color, branching pattern, relative length of the calyx teeth and degree of pod dehiscence (Barulina 1930). In the same way, a world lentil collection was classified into three main regional groups namely a Levantine group including Egypt, Jordan, Lebanon, and Syria; a more northern group encompassing Greece, Iran, Turkey, and the USSR; and another group comprising genotypes from India and Ethiopia (Erskine et al. 1989). Further, using single nucleotide polymorphism (SNP) markers, lentil accessions were categorized into three major groups namely subtropical savannah (South Asia), Mediterranean, and northern temperate (Khazaei et al. 2016); these groups reflecte the agro-ecological regions where lentil is grown. Additionally, others reports indicated differentiation of Moroccan landraces (southern Mediterranean) from landraces from Italy, Turkey and Greece (northern Mediterranean regions) (Idrissi et al. 2016, 2018).
Tillage and genotype × tillage interaction effects
In the present study, effects of tillage and genotype × tillage interaction were evaluated based on experiments conducted at Sidi El Aidi under no-till and conventional tillage system. No-tillage is an important component of conservation agriculture practices that contributes in improving soil physical, chemical and biological properties, thus increasing crop yield and productivity (Mrabet 1993, 2011). No-tillage can maintain yield stability in dry seasons through the conservation of soil moisture that helps crops cope with adverse effects of drought and heat stress. In lentil, it has been reported that no-till could be a buffer system for adaptation to drought stress (Saha et al. 2020). However, to exploit its full benefits, it is essential to tailor no-tillage technology to specific contexts, which may differ according to different factors such soil type, crop and climatic conditions. In addition, the genotypes developed for conventional tillage system may not exhibit the same level of performance when cultivated under conservation agriculture, making it necessary to set up a breeding program for each system; nevertheless, the implementation of such specialized breeding program requires justifying the presence of significant genotype × tillage system interaction (Serraj and Siddique 2012; Roohi et al. 2022). In the present study, there was no-significant effect of tillage system on all studied traits, suggesting that to achieve more sustainability and resource use efficiency no-tillage can be applied without any detrimental effects on lentil productivity. In agreement with these results, another study reported similar yield between no-till and conventional tillage in 13 lentil genotypes under Moroccan rainfed conditions (Devkota et al. 2021), suggesting that lentil can produce acceptable yield under conservation agriculture compared to conventional tillage without yield penalty. Similarly, traits such as days to flowering and days to maturity were not influenced by tillage system; however, plant height was influenced by tillage system in wet year compared to dry one (Devkota et al. 2021). In another study, in lentil, tillage system effect was not significant on seed yield and seed weight (Das et al. 2019). In chickpea, the effect of tillage system was not significant on different traits, including early vigour, number of secondary branches, time to flowering and grain yield (Abderemane et al. 2023).
In the present study, there was no genotype × tillage system interaction effect observed for all traits, except hundred-seed weight. Previous studies have documented genotype × tillage interaction for grain yield in wheat (Roohi et al. 2022; Mohammadi et al. 2024) and chickpea (Devkota et al. 2021; Abderemane et al. 2023). Although frequent genotype × tillage system interaction was observed in chickpea and wheat compared to lentil and barley, the major variation in grain yield is attributed principally to the variation in rainfall in terms of amount and distribution (Devkota et al. 2021). Overall, no significant effect of genotype × tillage system interaction on grain yield highlights that the establishment of a specialized breeding program could not be justified which implicates that varieties developed for conventional tillage system could be adopted in conservation agriculture system. However, the efficiency of the lentil selection under conservation agriculture condition was not yet addressed and warrant additional investigations, particularly for grain yield that indicated high heritability under no-tillage as compared to conventional tillage.