Cet article synthétise une réflexion collective sur les orientations génétiques des bovins laitiers et allaitants de France métropolitaine nécessaires à l’adaptation des animaux aux évolutions des systèmes de production à l’horizon 2040. Il présente les facteurs susceptibles de faire évoluer les productions bovines françaises (démographie des éleveurs, changement climatique, raréfaction des ressources, progrès techniques et évolutions technologiques, réglementations et politiques agricoles, libéralisation du marché, évolution de la demande et organisation des filières). Les systèmes de production bovins devraient néanmoins conserver une diversité importante dans un contexte économique et réglementaire plus contraint, de changement climatique structurant et de raréfaction des ressources non renouvelables. Cet article recommande d’axer la sélection sur la rentabilité économique globale du système plutôt que sur le rendement animal. Pour les bovins laitiers, il s’agira d’augmenter la longévité fonctionnelle, réduire le format des animaux et augmenter la persistance laitière. Pour les bovins à viande, il est souhaitable d’augmenter la précocité sexuelle et la précocité de développement, améliorer la capacité des animaux à valoriser l’herbe et les fourrages, l’aptitude laitière, l’autonomie des animaux et, pour les systèmes d’engraissement herbager, réduire le format. Enfin, pour les bovins laitiers et à viande, il est suggéré d’augmenter la capacité d’ingestion de fourrages grossiers. En regard du changement climatique, il est indispensable de prendre en compte les problèmes sanitaires associés (en particulier émergents), améliorer la tolérance au stress thermique, et réduire les émissions de méthane. La dernière partie de cet article discute des besoins des programmes de sélection induits par ces évolutions.

Genomic offsets are increasingly used to quantify the mismatch between a population's current genetic composition and the composition predicted under changed environmental conditions. While genomic offset is a promising tool for assessing climate maladaptation, the sensitivity of predictions to different methodological choices is not well understood. In this study, we compared two fundamentally different approaches to detect outliers before predicting genomic offsets: Gradient Forest (GF, non-linear, non-parametric) and Canonical Correlation Analysis (CANCOR, linear, parametric). To do so, we used 457 natural populations of perennial ryegrass (Lolium perenne L.), an important agricultural forage species throughout Europe. Using a data set of 189,968 SNPs and 75 climatic variables, we experimentally validated genomic offsets against 105 phenotypic traits measured across three common gardens during multiple years. We also assessed the sensitivity of outlier detection and genomic offset predictions to the number and spatial distribution of sampled populations. Both GF and CANCOR detected a substantial number of outlier loci associated with environmental gradients (2113 and 653, respectively), with 429 loci identified by both approaches. When used to model spatial variation in genetic adaptation and estimate genomic offsets, the different outlier sets produce spatially congruent projections. We also found significant correlations between experienced genomic offset in the common garden predicted by both outlier sets and phenotypic traits, identifying traits that could serve as good fitness proxies for assessing climate risk. Analyses based on different population subsamples revealed that GF was less sensitive to sample size and geographic biases than CANCOR. Our findings provide practical guidance for designing genomic offset studies in both agricultural and natural systems and suggest that non-linear, non-parametric methods like GF may be less sensitive to sampling design and therefore potentially more robust for predicting climate maladaptation.

Under the regulatory framework of the International Union for the Protection of New Varieties of Plants (UPOV) for plant variety protection, a new variety must be assessed for uniformity, as well as distinctness and stability (DUS). For cross-pollinated crops such as perennial ryegrass (Lolium perenne) and lucerne (Medicago sativa), where cultivars are heterogeneous populations, assessment of uniformity poses particular challenges. These challenges stem from the fact that such evaluations can be costly and require substantial labour and time, as they typically involve the measurement of numerous comparator characteristics across multiple individual plants representative of a candidate cultivar, often over a period exceeding 1 year. Although DUS testing is based on phenotypic characteristics, genetic markers are currently permitted for use in planning of trials, and for distinctness assessment in particular situations. However, genetic markers have not yet been employed for distinctness assessment. Similarly, under OECD Seed Schemes, marketed seed is assessed for varietal purity through the observation of plants in field plots. This study proposes a new measure of genetic heterogeneity based on genetic markers, along with a method to estimate this in pooled samples to reduce costs. The method relies on a system for estimating allele frequencies at SNPs, such as genotyping-by-sequencing or sequence capture. We demonstrate the potential of this approach with example data for perennial ryegrass. The approximation based on pooled samples provided satisfactory estimates of genetic heterogeneity compared with direct assessment based on individual plants, and proved to be sensitive to contamination of samples by other varieties.

There is growing interest in using temperate forage trees to alleviate the effects of summer drought and heatwaves on herbaceous forage. However, forage trees remain understudied in temperate climates. We studied the seasonal variation of the nutritive value of 16 tree species commonly found in Western Europe. We collected 285 samples of tree leaves between spring and autumn (June, August and October) over three years at 14 sites across France. We measured seven nutritive characteristics: in vitro dry matter digestibility (IVDMD) and the contents of crude protein (CP), dry matter (DM), neutral detergent fibre (NDF), acid detergent fibre (ADF), acid detergent lignin (ADL), and ash. We used linear mixed models to analyse their seasonal variation and then clustered the species based on CP and IVDMD. CP content and IVDMD generally decreased from spring to autumn (by 26% and 6 percentage points), while DM and ash contents increased (by 42 and 32%). Corylus avellana , Morus alba , and Robinia pseudoacacia had the greatest CP content (from 138 to 250 g.kg −1 ), and M. alba had the greatest IVDMD (84.7% on average). We observed a trade-off between CP and IVDMD among clusters. The order of clusters based on their nutritive value remained consistent across seasons. Our findings highlight the importance of carefully planning tree use, as their nutritive value varies substantially among species and across seasons. Results provide new opportunities for farmers to compensate for the lack of herbaceous forage in summer, even though yield and palatability aspects remain to be studied.

In France, INRAE (Institut national de recherche pour l'agriculture, l'alimentation et l'environnement - French National Research Institute for Agriculture, Food and Environment) maintains a genebank of grass and legume perennial species for forage and turf usages in the research unit P3F located in Lusignan (region Nouvelle-Aquitaine). This genebank is a component of the plant pillar (BRC4Plants) of the French National Research Infrastructure RARe. Collections comprise natural populations collected across France and other countries, landraces, cultivars removed from registration lists and some scientific materials. 967 accessions are currently available for distribution, among which 852 are available in the frame of the Multilateral System of Access and Benefit-sharing of the UN Food and Agriculture Organization (FAO). The genebank policy has been to set up core collections of the main forage and turf species diversity and to apply a high standard of conservation and regeneration to collections of relatively small size. During past decades, genetic resources held by the genebank highly contributed to the breeding of forage and turf cultivars in France and to various scientific studies. Recently, high-throughput genotyping of accessions has provided unprecedented means to discover phylogeographic patterns and genomic adaptation in natural populations of perennial ryegrass (Lolium perenne L.) and to understand the breeding history of lucerne (Medicago sativa L.). Such approaches open promising prospects for future genetic adaptation of forage and turf species to changing environmental conditions and new usages.

Forage production, persistence, and associated ecosystem services provided by the major forage legume, alfalfa (Medicago sativa), may be affected by disease susceptibility. Resistance to anthracnose, caused by Colletotrichum trifolii, has been described as an oligogenic trait, but the precise location of resistance genes on the alfalfa genome is not known. Therefore, we phenotyped a set of 417 alfalfa accessions for anthracnose resistance as the frequency of resistant plants. With available genotyping by sequencing data for 380 accessions from this collection, we performed quantitative trait locus (QTL) detection by genome-wide association study (GWAS) and genomic prediction using a validation set of 97 accessions randomly selected. A wide range of variation for anthracnose resistance was observed, with newer varieties and breeding materials exhibiting greater resistance than old varieties and landraces. Accessions from America showed the highest resistance, although some European accessions also displayed notable resistance. Six QTLs, controlling 58% of the variation, were identified by GWAS. Two major QTLs were found on chromosome 8, within a region already identified in an alfalfa mapping population. Four other QTLs, each controlling less than 5% of the variation, were also found, including one near a major QTL on chromosome 4 in the model species M. truncatula. The predictive ability of our set of accessions was surprisingly high: 85%. These results are promising and highlight the potential of molecular markers and genomic prediction to improve anthracnose resistance in alfalfa breeding programs.

The introduction of ley in cropping systems can provide multiple services, including high quality forage production and C and N inputs into arable soils. Little is known about the dynamics of these services in degrading grassland and the extent to which trade-offs regarding the use of fixed N for either forage production or as a source of N in the rotation can be found. A six-year field experiment compared the performance of an alfalfa-tall fescue (A-FE) mixture versus the single species alone (A, FE) in terms of forage production, forage quality and root biomass available as residues for the following crop. N mineralisation potential was then simulated for different grassland destruction scenarios using the STICS model. The forage production, proportion of legumes, and forage quality all declined in the A-FE mixture after year 3 to reach less than 7 T.ha−1 and 20

The agro-ecological transition aims at reducing the anthropogenic impacts of crop production on the environment, for instance by decreasing drastically the applications of pesticides, among which herbicides are the most prevalent. In this review, we focus on management of arable weeds in agro-ecological systems, considering a perspective of steady reduction of synthetic herbicides by fostering the breeding of varieties adapted to non-chemical weed management. Diverse strategies of non-chemical weed management are discussed, taking into account agronomic levers and identifying breeding targets. Weed suppression by enhancing crop competition from cash or cover crops, grown in pure stands or as intercrops, is a key strategy that could be considered together with dense canopies and optimal nitrogen management, also in addition to growing varieties that are tolerant to weed competition and/or characterized by low nitrogen requirements. Then, escaping weed competition could be achieved by shifting sowing dates and/or diversifying crop rotations, particularly by targeting varieties of different maturity groups, more productive spring-sown crops and integrating more frequently minor crops in the rotation. Weeds can be also suppressed by mechanical control that requires varieties tolerant to mechanical weeding. Allelopathy is a less applied strategy that deserves further studies e.g. the screening of allochemical composition among varieties of cash and cover crops. For each crop-related agronomic lever contributing to integrated weed management, we identify the functional crop traits to target, i.e. the set of morpho-physiological traits associated with an effective weed management, to be screened within the commercial variety panels or to be integrated in a genetic improvement scheme. For all the functional traits and according to the crop species, the potential availability of genetic resources, as well as the ability of varieties to meet the required genetic variability have been explored while, where relevant, the development of appropriate phenotyping methods and trait assessment procedures have been considered. Finally, we propose a set of non-chemical weed management strategies, functional effect traits and agronomic practices associated, as well as their synergies and antagonisms with the other cropping practices for cash and cover crops. We conclude that, to better combine a set of agronomic levers with crop varieties or reinforcing the efficacy of these levers, there is a need to complete classical agronomy and weed science approaches by plant genetics and breeding when designing and evaluating non-chemical weed management strategies.

ContextThe reduction of chemical inputs in agriculture is a current challenge. Perennial forage legumes such as lucerne produce protein-rich forage without synthetic nitrogen fertilizers but the crops may be invaded by weeds. Association of legumes with grasses are recognized to lower weed pressure but may alter forage yield and quality.ObjectivePure lucerne was compared to lucerne-grass mixtures, with the test of grass species differing for morphological, phenological and quality traits. The grass species were compared in mixture with lucerne for their performance, which was evaluated through weed occurrence, forage yield and quality.MethodsSeven perennial grass species were evaluated in association with lucerne and compared to pure lucerne and pure grass (one species only), in field plot experiments in two locations, without nitrogen fertilisation. Forage yield, botanical composition (lucerne, grass, weeds), protein and acid detergent fibre were measured in four cuts per year during three years. Treatments were compared, and mixture model effects such as direct and associated effects of the grass on dry matter yield of the species in the associations were calculated.ResultsLucerne-grass association reduced weed development compared to pure lucerne. The protein content was slightly lower in the associations than in pure lucerne but the association generally produced more protein per hectare than expected if the two species were grown in separated plots (i.e. protein overyielding). Depending on the grass species, the weed control, the forage quality, the proportion of lucerne and the dry matter yield of the association were differently affected. Species, with a small direct effect (on grass production) and a large associated effect (on lucerne production), such as timothy and meadow fescue, favoured lucerne proportion in the associations.ConclusionsAssociation of grass species with lucerne is a way to limit weed occurrence while maintaining protein content and forage yield in the association. These traits, as well as the lucerne proportion in the association, varied depending on the grass species.ImplicationsAt a practical level, the application of herbicides on lucerne crops could be significantly lowered, and this with a limited impact on forage yield and quality. From a scientific point of view, the calculation of mixture model effects is of interest to analyse the outcome of species associations.

China's and Europe's dependence on imported protein is a threat to the food self-sufficiency of these regions. It could be solved by growing more legumes, including alfalfa that is the highest protein producer under temperate climate. To create productive and high-value varieties, the use of large genetic diversity combined with genomic evaluation could improve current breeding programs. To study alfalfa diversity, we have used a set of 395 alfalfa accessions (i.e. populations), mainly from Europe, North and South America and China, with fall dormancy ranging from 3 to 7 on a scale of 11. Five breeders provided materials (617 accessions) that were compared to the 400 accessions. All accessions were genotyped using Genotyping-by-Sequencing (GBS) to obtain SNP allele frequency. These genomic data were used to describe genetic diversity and identify genetic groups. The accessions were phenotyped for phenology traits (fall dormancy and flowering date) at two locations (Lusignan in France, Novi Sad in Serbia) from 2018 to 2021. The QTL were detected by a Multi-Locus Mixed Model (mlmm). Subsequently, the quality of the genomic prediction for each trait was assessed. Cross-validation was used to assess the quality of prediction by testing GBLUP, Bayesian Ridge Regression (BRR), and Bayesian Lasso methods. A genetic structure with seven groups was found. Most of these groups were related to the geographical origin of the accessions and showed that European and American material is genetically distinct from Chinese material. Several QTL associated with fall dormancy were found and most of these were linked to genes. In our study, the infinitesimal methods showed a higher prediction quality than the Bayesian Lasso, and the genomic prediction achieved high (>0.75) predicting abilities in some cases. Our results are encouraging for alfalfa breeding by showing that it is possible to achieve high genomic prediction quality.

The majority of forage grass species are obligate outbreeders. Their breeding classically consists of an initial selection on spaced plants for highly heritable traits such as disease resistances and heading date, followed by familial selection on swards for forage yield and quality traits. The high level of diversity and heterozygosity, and associated decay of linkage disequilibrium (LD) over very short genomic distances, has hampered the implementation of genomic selection (GS) in these species. However, next generation sequencing technologies in combination with the development of genomic resources have recently facilitated implementation of GS in forage grass species such as perennial ryegrass (Lolium perenne L.), switchgrass (Panicum virgatum L.), and timothy (Phleum pratense L.). Experimental work and simulations have shown that GS can increase significantly the genetic gain per unit of time for traits with different levels of heritability. The main reasons are (1) the possibility to select single plants based on their genomic estimated breeding values (GEBV) for traits measured at sward level, (2) a reduction in the duration of selection cycles, and less importantly (3) an increase in the selection intensity associated with an increase in the genetic variance used for selection. Nevertheless, several factors should be taken into account for the successful implementation of GS in forage grasses. For example, it has been shown that the level of relatedness between the training and the selection population is particularly critical when working with highly structured meta-populations consisting of several genetic groups. A sufficient number of markers should be used to estimate properly the kinship between individuals and to reflect the variability of major QTLs. It is also important that the prediction models are trained for relevant environments when dealing with traits with high genotype × environment interaction (G × E). Finally, in these outbreeding species, measures to reduce inbreeding should be used to counterbalance the high selection intensity that can be achieved in GS.

ABSTRACTHere we present SMAP, a software package that implements a suite of computational tools to extract multi-allelic haplotypes using read-backed haplotyping. SMAP tools first perform accurate read processing and analyze read mapping distributions across sample sets. Then, two complementary modules can be invoked for haplotype calling: SMAP haplotype-sites combines known Single Nucleotide Polymorphisms (SNPs) and/or read mapping position polymorphisms (SMAPs) to reconstruct compressed, read-reference-encoded haplotype strings. In contrast, SMAP haplotype-window works independent of prior knowledge of polymorphisms, groups reads by locus, defines a window enclosed between two custom border sequences, and retains the entire corresponding DNA sequence as haplotype. Haplotype-window is, among many applications, especially useful for high-throughput CRISPR/Cas mutation screens. Either way, SMAP creates a single integrated haplotype call table across all loci and samples. SMAP haplotyping is extremely versatile and can be applied to highly multiplex amplicon sequencing (HiPlex), Shotgun (e.g. whole genome shotgun (WGS) sequencing, probe capture and RNA-Seq), or Genotyping-by-Sequencing (GBS) data; and to Illumina short reads, PacBio and MinION long reads. SMAP creates discrete genotype calls for individuals of any ploidy or quantitative haplotype frequency spectra for Pool-Seq data, and can scale from tens to thousands of loci and/or samples. SMAP, including the source code written in Python is available at https://gitlab.com/truttink/smap, and a detailed user manual and guidelines for accurate read processing is available at https://ngs-smap.readthedocs.io/, under the GNU Affero General Public License v3.0.

Both from the environmental and economical perspective, reducing the use of mineral nitrogen and herbicides is one of the future challenges in cereal production. Growing winter cereals on perennial legume living mulch such as white clover (Trifolium repens L.) or lucerne (Medicago sativa L.) is one of several options to reduce the need for mineral nitrogen fertilizer and herbicides in winter cereal production. Given the importance of winter cereals in the world, adopting this technique could greatly improve the sustainability of crop production. Through competition with the crop however, the living mulch can negatively affect cereal yield. Here, we (i) review how living mulch can be introduced in the system, (ii) synthetize potential advantages and disadvantages of that system, and (iii) explore different strategies to control the competition between the crop and living mulch. The major findings are that (i) competition between cereals and mulch can lead to significant yield reductions if not controlled properly and (ii) perennial legume varieties used as living mulch so far are varieties bred for forage production. We hypothesize that a dedicated breeding program might lead to living mulch varieties with a smaller impact on cereal yield compared to forage varieties, allowing to grow cereals with reduced nitrogen and herbicide inputs. We propose the main characteristics of an ideotype for such a perennial legume variety.

Various adaptive mechanisms can ensure that seedlings are established at the most favourable time and place. These mechanisms include seed dormancy i.e., incapacity to germinate in any environment without a specific environmental trigger and inhibition i.e., incapacity to germinate in an unfavourable environment (water availability, temperature: thermoinhibition and light). The objective of this research was to study in the temperate range for germination of forage and turf grass species perennial ryegrass, if the thermal requirements for germination are under genetic controlled and could be selectively bred. Two divergent selections of three cycles were realized on a natural population: one to select for the capacity to germinate at 10°C vs. the impossibility to germinate at 10°C, and one to select for the capacity to germinate at 32°C vs. the impossibility to germinate at 32°C. Seeds of all the lots obtained from the two divergent selections were then germinated at constant temperatures from 5 to 35°C to evaluate their germination ability. Concerning the positive selection, the first cycle of positive selection at 10°C was highly efficient with a very strong increase in the germination percentage. However, afterward no selection effect was observed during the next two cycles of positive selection. By contrast, the positive selection at 32°C was efficient during all cycles with a linear increase of the percentage of germination at 32°C. Concerning the negative selection, we observed only a large positive effect of the first cycle of selection at 10°C. These findings demonstrate that seed thermoinhibition at 10 and 32°C observed in a natural population of perennial ryegrass has a genetic basis and a single recessive gene seems to be involved at 10°C.

The persistence of perennial herbaceous species is threatened by increasing aridity. However, summer dormancy is a strategy conferring superior survival to grasses adapted to hot and dry summers. The role of temperature on the induction of summer dormancy was investigated in the perennial grass Dactylis glomerata to analyse the potential expression of this strategy under warmer climates. We tested seven populations of D. glomerata originating from Morocco to Norway across the same latitudinal gradient in a five‐site experiment. One population of the highly summer‐dormant grass Poa bulbosa was used as a reference. Plants were grown from autumn in pots under full irrigation for 1 year mostly under open‐air shelters. Heading date (ear emergence preceding flowering) was recorded and foliage senescence was assessed from end of spring until autumn. The maximum plant senescence under summer irrigation indicated the level of dormancy expression. Summer dormancy onset, release, expression and duration were modelled as a function of climatic variables. From north to south, the duration of summer dormancy of the Mediterranean populations of D. glomerata and P. bulbosa ranged from 0 to 122 days, and 79 to 200 days, respectively. P. bulbosa was always completely dormant, while dormancy expression of D. glomerata was positively correlated with the sum of temperatures from winter onset (R2 = 0.57) and with the mean of minimum temperatures in summer (R2 = 0.73). Dormancy onset, release and duration were also positively correlated with thermal time from winter onset, while the duration of summer dormancy was longer as maximum temperatures increased. Mapping the European regions with climates allowing the expression of summer dormancy in D. glomerata showed that the potentially inductive areas for this strategy may expand in parallel with increasing summer aridity under a future climate warming scenario. Synthesis. The large phenotypic variability of the expression of summer dormancy in D. glomerata was driven by temperature, suggesting that this strategy may have a greater role in higher latitudes to increase plant survival over the predicted hotter and drier summers. Leveraging this strategy for the choice and selection of suitable populations could enhance future adaptation of major perennial grasses to climate change.

Perennial ryegrass is an important forage crop in dairy farming, either for grazing or haying purposes. To further optimise the forage use, this study focused on understanding forage digestibility in the two most important cuts of perennial ryegrass, the spring cut at heading and the autumn cut. In a highly diverse collection of 592 Lolium perenne genotypes, the organic matter digestibility (OMD) and underlying traits such as cell wall digestibility (NDFD) and cell wall components (cellulose, hemicellulose, and lignin) were investigated for 2 years. A high genotype × season interaction was found for OMD and NDFD, indicating differences in genetic control of these forage quality traits in spring versus autumn. OMD could be explained by both the quantity of cell wall content (NDF) and the quality of the cell wall content (NDFD). The variability in NDFD in spring was mainly explained by differences in hemicellulose. A 1% increase of the hemicellulose content in the cell wall (HC.NDF) resulted in an increase of 0.81% of NDFD. In autumn, it was mainly explained by the lignin content in the cell wall (ADL.NDF). A 0.1% decrease of ADL.NDF resulted in an increase of 0.41% of NDFD. The seasonal traits were highly heritable and showed a higher variation in autumn versus spring, indicating the potential to select for forage quality in the autumn cut. In a candidate gene association mapping approach, in which 503 genes involved in cell wall biogenesis, plant architecture, and phytohormone biosynthesis and signalling, identified significant quantitative trait loci (QTLs) which could explain from 29 to 52% of the phenotypic variance in the forage quality traits OMD and NDFD, with small effects of each marker taken individually (ranging from 1 to 7%). No identical QTLs were identified between seasons, but within a season, some QTLs were in common between digestibility traits and cell wall composition traits confirming the importance of hemicellulose concentration for spring digestibility and lignin concentration in NDF for autumn digestibility.

Trees could help to reduce livestock production vulnerability to climate change by providing a fodder resource during periods of drought. Fodder trees are commonly used in tropical and Mediterranean areas, but they remain poorly studied in temperate regions. Previous studies highlighted that leaves of some tree fodder species have nutritive values close to those of herbaceous forages in summer (e.g. Mahieu et al. 2021). However, little is known about the variation in nutritive value of fodder trees throughout their growing period, from early summer to autumn. This study focused on 16 tree species sampled in June, August and October, from 2014 to 2017, in 22 French locations ( Acer pseudoplatanus , Alnus cordata , Castanea sativa , Corylus avellana , Fagus sylvatica , Fraxinus americana , Fraxinus excelsior , Gleditsia triacanthos , Juglans x intermedia, Morus alba , Paulownia tomentosa , Prunus avium , Robinia pseudoacacia , Sorbus domestica , Ulmus minor , Ulmus ‘Nanguen’; n = 292). Leaf samples

Selection of grazing-tolerant lucerne (Medicago sativa) germplasm with low autumn dormancy and erect growth habit, suitable to environments with mild winters, is hindered by genetically-based tradeoffs among these traits. Marker-assisted selection (MAS) may help overcoming the difficulty of phenotypic selection and hasten the breeding process. Over 400 progenies from four bi-parental crosses between contrasting parents from two cultivars were phenotyped for grazing tolerance and six other morpho-physiological characters. In parallel, genotyping-by-sequencing marker data were generated, with the objective to facilitate the multi-trait MAS by identifying quantitative trait loci (QTL) for the traits and assessing the extent of their co-location. In total, 9594 markers, with less than 20% missing data, segregated as simplex markers in at least one progeny family. Twenty-seven QTL were identified for six traits in the different bi-parental crosses. QTL were found to a different extent on all homology groups except homology group 4. No QTL was found for dry-matter yield before grazing imposition. The number of QTL for other traits varied from one (basal plant diameter) to nine (plant growth habit). Only for growth habit were QTL observed in all four crosses. The QTL analysis largely confirmed the association between grazing tolerance, non-erect growth habit and autumn dormancy. However, the absence of complete co-location of QTL highlighted a partly different genetic control of these traits, thereby offering scope for MAS exploitation to select genotypes carrying useful trait combinations.