Durum wheat (DW) (Triticum turgidum L. ssp. durum) (Desf.) is the 10^th most cultivated cereal worldwide, with a total production of about 38 million tons (Xynias et al., 2020). DW is grown primarily in the Mediterranean basin, accounting for 75% of global production, supported mainly by Italy, Spain, France, Greece, Morocco, Algeria, Tunisia, Turkey and Syria ( Table 1 ). In addition, outside the Mediterranean basin, the major producers are Canada, Mexico, the USA, Russia, Kazakhstan, Azerbaijan, and India (De Vita and Taranto, 2019; Martínez-Moreno et al., 2022). Although DW production constitutes only 7% of total wheat production, the rest is produced from bread wheat (Triticum aestivum), its importance for the countries of the Mediterranean basin is pivotal. DW is considered a staple food as it constitutes the dominant part of the diet for many populations in this area. The main products derived from DW include pasta, cous cous, burghul, and bakery products. Durum wheat-based products have a low glycemic index which makes them healthy products that can be recommended for low-carb diets (Di Pede et al., 2021). Furthermore, DW constitutes the main food source for 1.2 billion poor people, providing 20/50% of daily calories, 20% of protein, and is considered a strategic crop for food security. Regarding the economic importance of DW, Italy is the world’s largest producer of pasta with over 3.36 million tons/year of pasta produced, and the leading country for exports with 1.9 million tons/year (Altamore et al., 2020). On a cultural level, DW and its ancestor wild emmer (Triticum turgidum ssp. dicoccoides) have been at the foundations of food, from the Neolithic period to the Greeks and the Roman Empire, up to the present day (Martínez-Moreno et al., 2020). Its cultivation and processing constitutes a cultural heritage.

However, the on-going climate change threatens DW production and is subjecting this crop to stresses rarely experienced. In the Mediterranean area and western Europe, climate projections for the 2040-2070 interval warn that extreme drought events will become more frequent and severe due to decreased winter precipitation and increasingly dry springs (Spinoni et al., 2018). In a recent study, it was estimated that global warming may reduce the world’s suitable areas for DW cultivation by 19% in 2050 and by 48% in 2100, with the greatest losses occurring in the Mediterranean basin which is recognized as a climate change hotspot (Shayanmehr et al., 2020; Martínez-Moreno et al., 2022). The main environmental constraints influencing the yield of DW in this area are drought, high temperatures, and salinity (De Vita and Taranto, 2019). These stresses, if occurring in growth stages such as flowering, pollination, and grain-filling, can strongly affect crop productivity.

This review aims to highlight the important role of durum wheat landraces as a useful genetic resource to improve the sustainability of Mediterranean agroecosystems, with a focus on adaptation to environmental stresses ( Figure 1 ). We described the most recent molecular techniques and statistical approaches suitable for the identification of beneficial genes/alleles related to the most important traits in landraces and the development of molecular markers for marker-assisted selection. Finally, we outline the state of the art about landraces genetic diversity and signature of selection, already identified from these accessions, for adaptability to the environment.

Durum wheat (tetraploid, genome AABB, 2n=4x=48) is a cereal grain evolved from the tetraploid domesticated emmer wheat Triticum turgidum ssp. dicoccum (Schrank ex Schübl.) Thell (Özkan et al., 2002). Domestication of wild emmer (Triticum turgidum L. ssp. dicoccoides) occurred in the Fertile Crescent (Israel, Jordan, Lebanon, Syria, south-eastern Turkey, northern Iraq, and western Iran) about 8000 years BCE (Before Common Era) from limited founder lineages (Özkan et al., 2002; Wang et al., 2022). Emmer wheat has been the main cereal together with einkorn and barley during the Neolithic period and the Bronze age. Then, it was progressively replaced by the tetraploid naked DW, spread by land through the Balkans and the maritime routes to the Mediterranean regions of Southern Italy, France, Spain, and Greece (Martínez-Moreno et al., 2020). Finally, DW became a prominent crop at about 300 BCE during the Hellenistic period. DW was commonly cultivated in the Roman Republic where the writers started to call it Triticum in Latin. Later, the early Islamic world and then the Arab empire further promoted the spread of the cultivation of DW through the Mediterranean areas by introducing several food types based on semolina (dry pasta and couscous). Until 1950-1955 most of the DW Mediterranean areas were cultivated with DW landraces, local accessions adapted to their place of origin (Martínez-Moreno et al., 2020). However, the first breeding activities were started in Italy in the early 1900s, resulting in the release of the most renowned cultivar Senatore Cappelli in 1915, by the breeder Nazareno Strampelli (Scarascia Mugnozza, 2005). Since then, the cv. Senatore Cappelli appeared in the pedigree of almost all new DW varieties due to its repeated use in DW programs until the end of the 1960s (Laidò et al., 2013). Afterward, the introduction of the law n. 580/67 for Pasta Pureness gave the impulse to start the development of private seed companies and international research centers, such as CIMMYT and ICARDA. In particular, the development of the CIMMYT-derived semi-dwarf wheat varieties led to the Green Revolution in several countries. During the period between 1960’-90’, breeding activities coupled with mutagenesis generated new genetic variability (Xynias et al., 2020) and efforts were made to release every year improved varieties with high yield potential and other interesting traits. Such more productive cultivars were preferred for cultivation over large areas. Therefore, the multitude of DW landraces planted for centuries was progressively replaced. This replacement has led to an important erosion of the environmental adaptation traits evolved during the years by the landraces.

The climate characteristics of the Mediterranean region have played a significant role in shaping the phenotypic (and the genotypic) configuration of both DW wild relatives and cultivated varieties. This basin is characterized by hot and dry summers, followed by cold and wet winters. Climate change, particularly important in the last decades, points to an increased variability, in which drought events, often coupled with heat waves, can hamper growth and development, eventually affecting crop yield. For example, yield is reduced of about 5% per Celsius degree of increase in temperature, with a gross loss reaching 24% under a growth temperature of 31°C during flowering (Liu et al., 2016). Clearly, the negative effect of the abiotic stress depends on its duration and the phenological phase of the plant. For instance, sudden and extremely high temperature (T > 32°C) for a short duration (3 to 5 days) is referred to as a heat shock, while moderately high maximum temperature (20 to 30°C) for a longer duration is known as chronic heat stress (Li et al., 2013). In DW, the most sensitive stages to heat stress are anthesis and grain filling (Chaparro-Encinas et al., 2021). Heat stress alters membrane fluidity and enzyme activity which in turn hamper respiration and photosynthesis, and related processes (e.g. electron flow and carbon fixation metabolism, starch accumulation and stability, architecture and functioning of thylakoids), as well as water assimilation and nutrient absorption and allocation in the plant body. After phase transition, this results in compromised pollen viability, aberrant macrosporogenesis, starch synthesis, and grain filling, responsible for the reduction in yield. Reduced water availability, due to both erratic or deficient rainfall and limited irrigation, worsens the negative effect of heat stress, hindering grain yield (in terms of seed number and weight) and technological quality and protein composition (Flagella et al., 2010). Drought stress is induced also by soil physical characteristics, which significantly affect water holding capacity and supply, influencing water and nutrient absorption by roots. Plants respond to drought and heat stress by enacting similar physiological mechanisms. Transcriptome analysis of heat susceptible and tolerant wheat revealed the involvement of multiple processes associated with tolerance to heat shock and drought stress, including the formation of deeper roots, synthesis of heat shock proteins, stomatal control, coordination of transpiration rate, and enhancement of osmoprotective response (Kulkarni et al., 2017). Also, the use of genome wide mapping approaches is providing abundant information about genomic regions associated to heat tolerance (Sukumaran et al., 2018).

Soil geo-biochemistry, geographical localization (sea proximity, with seawater intrusion into freshwater aquifers), and events of rising groundwater table can increase the amount of salts in soils. Moreover, anthropogenic activities, such as inappropriate irrigation and drainage practices or irrigation with brackish water, determine salt accumulation in the soil surface or within the solum, causing salinity stress in plants (Annunziata et al., 2017). Soil salinity is usually referred to the increased amount of Na⁺/Cl^¯ in the soil upper layer, but a variety of ions, mainly K⁺, Ca²⁺, Mg²⁺, and SO 4 2 − , HCO 3 − , NO 3 − , can also accumulate. On one hand, salinity leads to permanent modifications of the soil structure by decreasing soil aeration, leaching, and infiltration rate, and increasing runoff and soil erosion (Edelstein et al., 2010). On the other hand, salinity affects plant physiology and growth. Stress effects harbored by salinity are usually due to both (i) cell toxicity of the accumulated ions, which often results in nutrient imbalance and enhanced oxidative stress, and (ii) osmotic stress, due to the extremely low water potential of soil along with a reduction in water assimilation. Therefore, cellular, and metabolic processes involved to counteract salt stress are comparable to those induced by drought (Munns et al., 2006). Interestingly, DW is conventionally considered a tolerant crop enduring up to 5.9 dS/m (De Santis et al., 2021). In field conditions with 10 dS/m NaCl, DW produced a reduced yield, compared with rice that died before maturity (Munns et al., 2006). Lower level of salinity reduces the leaf area and shoot biomass, while grain yield is not affected. Tolerance to salinity is associated with low rate of Na⁺ root-to-shoot transport and higher selectivity for K⁺ than for Na⁺. Indeed, a correlation between grain yield and Na⁺ exclusion from the vegetative organs, together with the enhanced K⁺/Na⁺ discrimination in root absorption and xylem loading has been identified in tolerant genotypes (Munns et al., 2000), which confirmed xylem transport, as one of the main discriminants between sensitive and tolerant species (Davenport et al., 2005). Tolerance to salinity is a quantitative trait controlled by many genes. Moreover, it appears that the expression of genes responsible for salt tolerance depends on plant age and ontogeny. Also, environmental factors contribute to the large phenotypic variation reported, enhancing the difficulty of breeding programs aimed to improve salt tolerance (De Santis et al., 2021). In wheat, a QTL mapping approach has identified the locus Nax1 (involved in limiting Na⁺ translocation to the above-ground tissues and inducing salt tolerance), mapped to the long arm of chromosome 2A, responsible for almost 38% of phenotypic variation in low Na accumulation in the mapping population, and this locus, together with closely linked markers, are commonly adopted to select salt tolerant durum genotypes (Lindsay et al., 2004). Other characteristics of salt-tolerant genotypes include differential ion partitioning between aged and young leaves, cell osmotic adjustment contrasting osmotic stress, and early phase-transition, leading to a shorter growing season (Colmer et al., 2005).

Drought, heat, and salt stress, being linked to each other, induce the generation of reactive oxygen species (ROS), including hydrogen peroxide, superoxide, or hydroxyl radicals, which are continuously formed mainly in the cytosol, chloroplasts, and mitochondria (Laus et al., 2022). ROS have a significant role in signaling but, under stress conditions, their over-accumulation may be responsible for the oxidative stress characterized by membrane peroxidation, protein degradation, and DNA mutation, eventually leading to the death of the plant cell. Plant cells are usually equipped with a great variety of ROS scavenging enzymes including superoxide dismutase, catalase, and glutathione peroxidase, and antioxidants, such as ascorbic acid, tocopherol or glutathione, which also contribute to ROS detoxification (Dvorak et al., 2021). Interestingly, the tolerance of DW genotypes to environmental stresses leading to ROS production has been widely associated with higher activities of scavenging enzymes, pointing to a role of these mechanisms in the drought and salt tolerance in particular genotypes (Laus et al., 2022). Therefore, they are a good candidate to be considered in DW breeding programs. Anyway, it should be stressed that as the DW sensitivity to stress is influenced by the phenology, also the antioxidant performance depends on the stress characteristics (severity and duration), on the stage of development at which the stress acts, and on the plant organs targeted. Finally, (as shown by the increasing literature on the topic, Li et al., 2013; De Santis et al., 2021), breeding programs should also keep the attention on the stress effect on quality traits of DW grains. Indeed, changes in grain content and composition, affecting technological and health quality (e.g. protein, starch and dietary fiber accumulation and composition, phytochemical, and health-related micronutrient accumulation), are incredibly susceptible to environmental clues and stress and must be taken into account when implementing the breeding programs.

Many efforts are made by researchers and breeders to constantly look for new sources of genetic variability to improve the elite varieties for adaptation traits, to face climate change. The exploitation of the existing genetic variability, still available in landraces, is one of the best approaches to adopt. According to Camacho Villa et al. (2005), a landrace is “a dynamic population of a cultivated plant that has a historical origin, distinct identity, and lacks formal crop improvement, as well as often being genetically diverse, locally adapted, and associated with traditional farming systems”. It is the result of natural and/or farmer-mediated evolutionary forces that generated plants better adapted to the local climate/environmental conditions (Zeven, 1999).

They are considered a reservoir of useful alleles that can be exploited to broaden the genetic basis of important adaptation traits. Landraces are rich in micronutrients and have high concentrations in total phenol and antioxidant content, as well as in tocols, carotenoids, and lutein (Azeez et al., 2018). Since the landraces can tolerate abiotic and biotic stresses, their yield is lower than modern varieties (Tan, 2002). For this reason, landraces are no longer cultivated over large areas where the more productive cultivars are preferred. Anyhow, several landraces have been rediscovered and reused thanks to their adaptation to sustainable and low-input cropping systems. They produce a great amount of straw, which, when used for animals, can make them economically more convenient than modern varieties, or preferable for organic farming because of their greater competitive ability against weeds (Lemerle et al., 2001; Annicchiarico et al., 2005).

Indeed, thanks to the efforts of farmers and scientists, wheat landraces and old cultivars have been collected and conserved by in-situ or ex-situ strategies. The in-situ strategy relies on individual farmers who traditionally cultivate landraces for their production or are sponsored by the government or private companies. The ex-situ conservation is managed by international institutions such as CIMMYT, ICARDA, and USDA or by national projects led by local universities (Adhikari et al., 2022). With the advance of modern technologies, phenotyping and genotyping are extremely affordable, and the landraces can be studied both for their conservation and for molecular markers development (Nazco et al., 2012; Marone et al., 2021). The exploration of genetic variability in landrace germplasm has become an issue of great global interest, mainly during the last two decades. Figure 2 shows how the number of publications related both to “plant breeding” and “landraces”, and “plant breeding” and “climate change”, have had a strong increase since 2005, when “The 2005 United Nations Climate Change Conference” took place and marked the entry into force of the Kyoto Protocol. From 2005 to 2021 the number of publications related to the plant breeding strategies, to cope with climate change, has grown exponentially. Also for “durum wheat” and “landraces”, it can be noted a similar trend, although the number of publications is scarce. Some works aimed to characterize specific DW landraces are reported in Table 2 , which includes a list of the most important European landraces specially studied both for stresses and quality related traits.

Because of the genetically heterogeneous nature of landraces, which are in a constant state of evolution due to different factors such as ecogeographical area and conventional or modern breeding techniques (Casañas et al., 2017), the establishment of core collections represents an affordable cost approach to reduce the degree of co-ancestry redundancy and the genetic stratification in the landraces whole collections. The goal of creating core collections is to maximize the allelic richness at molecular markers and best represent variation at phenotypic traits, in order to define the smallest possible number of individuals that represents a more compact and manageable population. Core collections were made for Spanish, Indian, Iranian, and Israeli-Palestinian landraces (Etminan et al., 2017; Ruiz et al., 2012 and 2013; Frankin et al., 2021; Phogat et al., 2019; Chacón et al., 2020). In addition, a core set of landraces was developed starting from the Global Panel of Durum Wheat (GPD), reducing their number from 416 to 192 (Mazzucotelli et al., 2020 and Kabbaj et al., 2017). This approach is useful not only to represent whole genetic diversity but also to enquire and identify new sources for interesting traits. For example, SNP markers associated to resistance to leaf rust, tan spot and Stagonospora nodorum blotch were identified using the core collection created from the Watkins collection (Martínez-Moreno et al., 2021, Halder et al., 2019). Moreover, the core collections have also been used for unlocking the genetic and morpho-physiological adaptation traits to semi-arid environments (Abu-Zaitoun et al., 2018) and to study agronomic and quality traits, such as root architecture, stem cross section, height, heading date and carotenoid content (Ruiz et al., 2018; Ávila et al., 2021; Requena-Ramírez et al., 2021).

The DW sequencing project has been carried out by Maccaferri et al. (2019) using the modern cultivar Svevo. The annotation led to the identification of 66,559 genes, where the gene density distribution reflects the QTL density distribution. The comparison between wild emmer Zavitan and Svevo genomes identified putative loci under domestication and selection, and localized the reduction in diversity mainly in the pericentromeric regions of the chromosomes (Maccaferri et al., 2019). However, in the evolution of DW landraces, the reduction of diversity was more spread along the genome as a consequence of breeding programs.

Resequencing techniques, such as whole genome resequencing, are not suitable for species with complex genomes, for which a reduction of genomic complexity prior to NGS-based SNP discovery is preferred (Borrill et al., 2019). In polyploid species such as DW, with a large genome size (about 10.45Gb) and a large proportion of repetitive sequences (> 85%), the presence of paralogous and multi-copy sequences adds complexity in classifying the correct scoring of SNPs at a single locus for SNP discovery (Sandve et al., 2010).

In the last decades, the number of sequenced genomes in crop species has continued to increase exponentially, highlighting the presence of large structural variations between individuals of the same species. Therefore, relying on the single reference genome cannot represent the entire sequence diversity present in a population (Golicz et al., 2016). This observation led to the concept of “pangenome”, that describes the landscape of genetic variation within a species, in order to capture a comprehensive view of genetic diversity that include the entire crop gene pool (Bayer et al., 2020; Khan et al., 2020). In the pangenome development, it is pivotal to consider the genetic variation brought by the crop wild relatives, in order to include as much genetic variability as possible (Khan et al., 2020). In fact, wild species preserve important genes related to tolerance to various types of stress that were lost during the domestication process. The increasingly less expensive sequencing approaches have allowed to deepen the genetic architecture of the crop wild relatives leading, in the past decade, to several de novo sequencing projects developed also in crop wild relatives. In soybean, 14 cultivated and 17 wild accessions were resequenced, confirming the great allele diversity present in wild accessions (Zhou et al., 2015). In maize, 75 lines including cultivated, wild and landrace accessions were resequenced, highlighting the genes linked to selection and providing evidence for introgression from wild relatives (Hufford et al., 2012).

In wheat, the first pangenome has been constructed using the bread cv. Chinese Spring as suitable reference assembly, followed by the expansion of this reference with 16 additional sequences from other bread wheat varieties (Montenegro et al., 2017). Graph pangenomes based on 16 public assemblies (Wheat Panache) was developed with the aim to discover genome variation between cultivars and to mine the diversity present in the large and complex wheat genome (Bayer et al., 2022). However, the mathematical modeling of pangenome expansion revealed that all these wheat varieties have a closed pangenome; therefore, the inclusion of more distant accessions such as wild relatives and landraces could harbor yet unexplored sequence variants that may further increase the gene content of the pangenome. The use of divergent individuals may increase the number of novel gene variants as well as improve the accuracy of the read mapping for SNP discovery. The use of landraces can support the breeding of cultivars better adapted to diverse environments and more resilient to climate change; indeed, plant pangenome assemblies have shown that genetic variations are often associated with biotic or abiotic stress.

No pangenome has yet been assembled for durum wheat, although several projects are underway. Indeed, the use of Svevo genome as suitable reference may accelerate the sequencing of new durum cultivars enabling the pangenome construction. As far as is known, at the moment the only reference genome for durum wheat remains Svevo. Given the growing interest in some European landraces ( Table 2 ), the release of new genome assemblies from landraces is expected in the next few years.

Until a few years ago, DW was well adapted to the Mediterranean environment. More recently, due to the climate crisis, drought, salinity, and low nutrient inputs occurring during flowering, pollination, and grain-filling represent the major stresses which adversely affect crop yield and quality, thus hampering agricultural productivity. Landraces coming from the Mediterranean basin are considered a particularly important group of genetic resources thanks to their high variability and tolerance to drought, pests, and adaptability to low farming systems (Lopes et al., 2015). Nowadays, the recovery, conservation, and enhancement of landraces are becoming central to increase the resilience of agricultural systems. However, how to exploit the genetic diversity of landraces to deal with environmental stress resilience is unclear and scattered (Lopes et al., 2015).

With the advance in genomic sequencing technologies and the release of the DW genome (Maccaferri et al., 2019), there has been a growing interest in comparing the patterns of genetic variation observed in landraces and modern varieties. These analyses were often focused on panels of landraces with a specific geographical origin. Population structure analysis was conducted for example in Iranian, Ethiopian, Tunisian, Turkish, and Italian germplasm revealing that, in most cases, landraces clustered separately from modern cultivars (Fayaz et al., 2019; Alemu et al., 2020; Alsaleh et al., 2022; Miazzi et al., 2022). Interestingly, a high level of genetic variation within landrace populations was detected, according to their geographical and climate of origin, revealing the importance of these factors in shaping wheat genome (Alipour et al., 2017; Taranto et al., 2022). On the other hand, cases of synonyms or homonyms as well as the presence of higher admixture of accessions between different populations of landraces were discovered, probably due to the exchange of seeds associated with human migration over time.

Examining wider collections, including landraces from different geographic origins, opens the possibility of investigating relationships on a wider global level and provides also a more precise estimation of the genetic diversity within each group. Genotyping the already mentioned core GPD by means of the iSelect 90K SNPChip followed by structure analysis showed comparatively limited genetic diversity in modern cultivar and a closer relationship to specific landrace population (North Africa and Transcaucasia). Landraces from Ethiopia appeared instead as the more isolated and distant to modern cultivars, while a high admixture level within landrace populations was confirmed (Maccaferri et al., 2019).

In addition, genome-wide population structure uncovers divergent selection during modern wheat breeding, suggesting the existence of untapped gene pools which will provide a basis for DW improvement in the next future. Many hotspots of selection were detected in the genomic regions where there are located the genes for adaptation, quality, grain yield, and stress response (Taranto et al., 2020; Soriano et al., 2021). These hotspots included important loci such as the photoperiod (Ppd), the vernalization (Vrn), and the dwarfing (Rht) genes, as well as loci associated with nitrogen use efficiency, plant architecture and grain yield (TaASN3, asparagine synthetase 3; NR1, nitrate reductase 1; Fd-GOGAT, ferredoxin-dependent glutamate synthase; GS, glutamine synthetase; Sus2, sucrose synthase 2 and TEF, transcript elongation factor). In addition, genes related to quality such as pasta-making quality and color of semolina and other durum wheat-end products were also divergent between landraces and modern cultivars. In detail, loci for gluten composition (HMW/LMW, high/low molecular weight, and α, β, γ, and ώ gliadins), as well as loci involved in the carotenoid pathway (Psy) and polyphenol oxidase reaction (Ppo) were identified in hotspot regions (Requena-Ramírez et al., 2022; Taranto et al., 2022). Other divergent loci with implications in disease resistance, plant-microbe interactions, abiotic stresses, and plant development corresponded to gene models involved in important biological functions (Soriano et al., 2021). However, the identification of these genes and their allelic variants in the germplasm of indigenous DW varieties has been mainly carried out by in-silico analysis, at least for now. Future studies will be needed to validate the potential of the new allelic variants discovered in landraces.

In DW, domestication and, more lately, selection and fixation of favorable alleles had led to genetic erosion, lowering the buffering capacity of modern elite cultivars towards varied climatic conditions and strongly reducing potential for improvement (Tanksley and McCouch, 1997; Lopes et al., 2015). The systematic search and discovery of genetic resources from landraces by means of genotypic and innovative phenotypic profiling of genetic resource collections and the introduction in elite crops through pre-breeding efforts are being currently implemented in bread wheat and could be a promising strategy also for DW (Reynolds et al., 2021; Sharma et al., 2021; Schulthess et al., 2022). Improvement of yield potential of landraces, by considering them as recipients, can be undertaken as an alternative strategy for developing better cultivars adapted to climate change. Inherent agronomic inferiority and disease susceptibility would hinder the direct utilization of landraces for breeding programs. However, the genomic tools and approaches we have described (i.e. genomic prediction and selection) could also strengthen pre-breeding efforts aiming at the improvement of genetic backgrounds of landraces, thereby attempting to achieve agronomic superiority starting directly from landraces as recipients and making this second alternative approach a possible and viable option (Adhikari et al., 2022). Finally, more complex breeding approaches, to develop new “synthetic” wheat crops exploiting genetic resources from wild species instead of landraces, also exist and have been already undertaken in the past (Reynolds et al., 2007; Balla et al., 2022).

Desirable genes were already identified in landraces and exploited for DW cultivars improvement by classical breeding. The most renowned DW cultivar is Cappelli, which is assumed to have been selected from the North-African landraces (De Cillis, 1942). However, the cv. Cappelli has a height of about 1.80 meters; therefore, several breeding activities were focused to create new variability by crossing the cv. Cappelli with Syriacum landraces (Aziziah, Eiti, Sinai, Tripolino). The result was the introduction of cultivars such as Capeiti 8 and Patrizio 6 which had slightly lower height, higher yield, earliness, and lodging resistance compared to Capelli, while preserving grain quality. Other similar examples of superior DW varieties, obtained by introgressing traits from landraces, were released in the frame of other breeding programs conducted in different countries (Kabbaj et al., 2017; Martínez-Moreno et al., 2020).

Molecular mapping technologies such as bulk segregant analysis (BSA), gene/QTL mapping, and genome-wide association studies (GWAS), supported by the high-throughput genotyping tools and strategies previously described, importantly increased the rate discovery of genes/QTLs regulating biotic, abiotic stress resistance, agronomic and quality traits from landraces. Several studies have been already undertaken so far to identify new genes/traits in landraces useful for breeding (see Table 3 for a comprehensive but not exhaustive list). As can be clearly appreciated, in most recent studies traits and genes have been often also mapped by taking advantage of the newly released genomic tools for DW here described, to further support their prompt exploitation in breeding.

Besides introducing new abiotic stresses, climate changes are shaping the dynamics of plants and pathogens resulting in more complex biological interactions difficult to predict and characterized by new outbreaks. Therefore, landraces are being widely explored as a potential source of new resistance traits, in regions where plant and pathogens co-evolved. Resistance sources to Fusarium head blight, rust, common bunt, stem sawfly, tan spot, and Septoria tritici blotch disease have been discovered within different DW landrace collections and majority of these have also been successfully mapped in the latest years. As pivotal examples leaf and stem rust resistance sources were mapped in the Portuguese DW landraces PI 192051 and Aus26582, by developing RIL mapping populations, beside alternative contribution from other sources (Qureshi et al., 2017; Qureshi et al., 2018; Aoun et al., 2019). Similarly, a resistance gene to Zymoseptoria tritici was mapped in the Tunisian DW landrace ‘Agili 39’ (Ferjaoui et al., 2022).

Concerning abiotic stress resistance, it is well recognized that Mediterranean DW landraces represent a particularly important group because of their documented better adaptation to drought (Moragues et al., 2006; Royo et al., 2014). Therefore, landraces from this region potentially include adaptive traits that could be exploited to boost the breeding for heat/drought tolerance and promote cultivars adaptation to stress-prone environments. Up to now, drought, heat, and salinity resistance traits have been studied in landraces coming from the Mediterranean basin, such as in Jordanian, Israeli-Palestinian, Tunisian, Italian, and Spanish, but also Afghan landraces (Al Khateeb et al., 2017; Shamaya et al., 2017; Hamdi et al., 2020; Frankin et al., 2021; Naranjo et al., 2022; Taranto et al., 2022). However, works concerning the mapping analysis of resistance traits to abiotic stresses are less in number than the biotic stresses ones. Just in two cases, abiotic stress resistance sources have been genetically mapped. A salinity resistance trait has been identified in an Afghan DW landrace and mapped. Moreover, using a GWAS approach on a worldwide collection of DW landraces, drought stress tolerance was associated to a locus of DW genome found to be collinear with a previously identified QTL in bread wheat (Shamaya et al., 2017; Wang et al., 2019). In order to improve the genetic characterization of abiotic resistance traits from landraces, an important contribution is expected from the development of adequate protocols for the abiotic stress evaluation, following similar strategies to those applied in bread wheat for high-throughput and accurate stress response phenotyping in large collections (Rasheed and Xia, 2019; Langridge et al., 2021; Shan et al., 2022).

Landraces are typically low yielding and can show lower agronomic attributes. Therefore, several studies enquired agronomic traits variability in landraces, focusing mainly on yield, phenology and morphological traits ( Table 3 ). GWAS studies, based on high-throughput genotyping tools, helped in defining genomic regions affecting such agronomic traits in landraces highlighting available superior alleles. Among others, the contribution to the yield of root system architecture traits and phenology were highlighted (Mengistu et al., 2016; Kidane et al., 2017a; Desiderio et al., 2019; Gupta et al., 2020; Alemu et al., 2021; Ávila et al., 2021; Royo et al., 2021).

DW semolina is considered the ideal raw material for the production of pasta or macaroni products, especially in Italy, which is the first producer and consumer of DW in Europe (http://www.internationalpasta.org, accessed on 12 November 2022). The aptitude of the raw material to be transformed into a high quality end-product mainly depends on grain protein content (GPC) and composition that directly affect wheat’s market price and end-use value (Shewry, 2019).

Grain protein content, mainly above 12-13%, is highly related to the amount and composition of glutenins and gliadins proteins, that are the principal components of gluten and are responsible for the viscoelastic properties and extensibility of semolina, respectively. Past breeding activities aimed at improving grain yield resulted in a loss of genetic variability for quality-related traits, probably because of the negative relationship between yield and GPC (Nazco et al., 2014a; Subira et al., 2014). As proof of this, Roncallo et al. (2021) observed a decreasing trend in GPC over the last 85 years using a DW collection including accessions representative of the Argentina, Italy, Chile, France, CIMMYT and other countries breeding programs.

Previous studies suggested the potential quality-enhancing landraces as reservoir of new allelic variants for gluten quality improvement ( Table 3 ) (Moragues et al., 2006; Nazco et al., 2014b; Roselló et al., 2018; Ruiz et al., 2018; Hernández-Espinosa et al., 2018). Roselló et al. (2018) performed a pasta-making quality QTLome using a Mediterranean collection of DW landraces and observed how landraces had higher GPC than modern cultivars but lower gluten strength. This result is due to very few allelic combinations of glutenin subunit loci in modern cultivars (Nazco et al., 2014b), while landraces showed a higher genetic variability useful to recovering and broadening allelic variation of gluten composition.

Other parameters can affect pasta production such as color ( Table 3 ). Semolina and pasta color are constituted by yellow (desirable) and brown (undesirable) pigments (Colasuonno et al., 2019). Usually, DW landraces showed lower total carotenoid contents and higher values of browning compounds compared to commercial cultivars (Digesù et al., 2009; Subira et al., 2014; Taranto et al., 2015). However, the first DW landraces with carotenoid esterification ability were identified by Requena-Ramírez et al. (2021) and could represent donor sources in DW biofortification programs.

Although DW is mostly used for pasta production, it is an ingredient in typical breads in some areas of Southern Italy. It is the case of “Pane nero di Castelvetrano” and “Pane di Monreale” which are two traditional breads constituted by two Sicilian landraces, Timilia and Russello (Palumbo et al., 2008; Melini et al., 2021; Visioli et al., 2021). The most notable characteristic of Timilia is the dark color of semolina, due to the high content of antioxidant phenolic compounds (Giancaspro et al., 2016; Bianco et al., 2017; Taranto et al., 2022). To preserve these landraces and derived-products, a traceability approach was developed using the high molecular weight glutenins, suggesting a method to verify the varietal identity from the seed to the final product (Visioli et al., 2021).

In conclusion, recent findings unveiled the strategic role of landraces in the genetic improvement of durum wheat. Studies on genomic divergence among T. turgidum sub-species indicated that the allelic variations of domesticated accessions and their wild relatives, lost during the domestication and breeding processes, were and will be recovered by exploring and exploiting landraces genetic diversity. In particular, in a context of climate changes, understanding the environmental and genetic factors behind the adaptation of landraces can help to introduce beneficial alleles in elite varieties to overcome stress and increase yield. The availability of durum wheat reference genome and the increasingly precise molecular techniques at affordable costs are giving a big boost to accurately identify the genetic determinants underpinning resistance/tolerance against biotic and abiotic stresses.

The recent application of genomic technologies (i.e. genome-wide association and genomic prediction analysis) on durum wheat landrace resources paves the way to accelerate the next-generation breeding programs to overcome the gap of knowledge of these underexplored resources and identify advantageous alleles that have been lost in modern varieties.

CB: writing original draft. DB and FT: conceptualization and writing. FT, GD, AF: writing, review, and editing. All authors contributed to the article and approved the submitted version.