Biologically meaningful genome interpretation models to address data underdetermination for the leaf and seed ionome prediction in Arabidopsis thaliana.

Genome interpretation (GI) encompasses the computational attempts to model the relationship between genotype and phenotype with the goal of understanding how the first leads to the second. While traditional approaches have focused on sub-problems such as predicting the effect of single nucleotide variants or finding genetic associations, recent advances in neural networks (NNs) have made it possible to develop end-to-end GI models that take genomic data as input and predict phenotypes as output. However, technical and modeling issues still need to be fixed for these models to be effective, including the widespread underdetermination of genomic datasets, making them unsuitable for training large, overfitting-prone, NNs. Here we propose novel GI models to address this issue, exploring the use of two types of transfer learning approaches and proposing a novel Biologically Meaningful Sparse NN layer specifically designed for end-to-end GI. Our models predict the leaf and seed ionome in A.thaliana, obtaining comparable results to our previous over-parameterized model while reducing the number of parameters by 8.8 folds. We also investigate how the effect of population stratification influences the evaluation of the performances, highlighting how it leads to (1) an instance of the Simpson's Paradox, and (2) model generalization limitations.

The Effect of Genome Parametrization and SNP Marker Subsetting on Genomic Selection in Autotetraploid Alfalfa.

BACKGROUND: Alfalfa, the most economically important forage legume worldwide, features modest genetic progress due to long selection cycles and the extent of the non-additive genetic variance associated with its autotetraploid genome. METHODS: To improve the efficiency of genomic selection in alfalfa, we explored the effects of genome parametrization (as tetraploid and diploid dosages, plus allele ratios) and SNP marker subsetting (all available SNPs, only genic regions, and only non-genic regions) on genomic regressions, together with various levels of filtering on reading depth and missing rates. We used genotyping by sequencing-generated data and focused on traits of different genetic complexity, i.e., dry biomass yield in moisture-favorable (FE) and drought stress (SE) environments, leaf size, and the onset of flowering, which were assessed in 143 genotyped plants from a genetically broad European reference population and their phenotyped half-sib progenies. RESULTS: On average, the allele ratio improved the predictive ability compared with other genome parametrizations (+7.9% vs. tetraploid dosage, +12.6% vs. diploid dosage), while using all the SNPs offered an advantage compared with any specific SNP subsetting (+3.7% vs. genic regions, +7.6% vs. non-genic regions). However, when focusing on specific traits, different combinations of genome parametrization and subsetting achieved better performances. We also released Legpipe2, an SNP calling pipeline tailored for reduced representation (GBS, RAD) in medium-sized genotyping experiments.

High-speed cold centrifugation of milk modifies the microbiota, the ripening process and the sensory characteristics of raw-milk hard cheeses.

The microbial population of raw milk plays a crucial role in the development of distinctive traits of raw-milk cheeses particularly appreciated by consumers. It was previously demonstrated that the microbial population of raw milk is modified by a high-speed centrifugation (also called bactofugation) conducted at 39 °C. The aim of the present study was to evaluate the effects of this process, performed once or twice, on the microbial, compositional, biochemical, and sensory characteristics of the derived hard cheeses. Experimental and control cheesemaking were conducted in parallel at a cheese factory during a 13-month period. Cheeses were analysed after 9, 15 and 20 months of ripening for microbial count, composition, proteolysis extent, volatile compounds, and sensory profile. Results evidenced that experimental cheeses were characterized by lower numbers of viable lactobacilli respect to control. Experimental cheeses also showed differences in the progress of primary and secondary proteolysis which, in turn, caused different patterns of free amino acids at all ripening times. Experimental cheeses had significantly lower content of esters and were differentiated from control for some traits by assessors. In conclusion, use of high-speed centrifugation of milk shall be discouraged if characteristic traits of raw-milk cheeses, particularly PDO cheeses, want to be retained.

Assessment of residual plant DNA in bulk milk for Grana Padano PDO production by a metabarcoding approach.

The aim of this study was to evaluate the ability of DNA metabarcoding, by rbcl as barcode marker, to identify and classify the small traces of plant DNA isolated from raw milk used to produce Grana Padano (GP) cheese. GP is one of the most popular Italian PDO (Protected Designation of Origin) produced in Italy in accordance with the GP PDO specification rules that define which forage can be used for feeding cows. A total of 42 GP bulk tank milk samples were collected from 14 dairies located in the Grana Padano production area. For the taxonomic classification, a local database with the rbcL sequences available in NCBI on September 2020/March 2021 for the Italian flora was generated. A total of 8,399,591 reads were produced with an average of 204,868 per sample (range 37,002-408,724) resulting in 16, 31 and 28 dominant OTUs at family, genus and species level, respectively. The taxonomic analysis of plant species in milk samples identified 7 families, 14 genera and 14 species, the statistical analysis conducted using alpha and beta diversity approaches, did not highlight differences among the investigated samples. However, the milk samples are featured by a high plant variability and the lack of differences at multiple taxonomic levels could be due to the standardisation of the feed rationing, as requested by the GP rules. The results suggest that DNA metabarcoding is a valuable resource to explore plant DNA traces in a complex matrix such as milk.

Genetic Diversity and Population Structure of European Soybean Germplasm Revealed by Single Nucleotide Polymorphism.

Soybean is the most grown high-protein crop in the world. Despite the rapid increase of acreage and production volume, European soybean production accounts for only 34% of its consumption in Europe. This study aims to support the optimal exploitation of genetic resources by European breeding programs by investigating the genetic diversity and the genetic structure of 207 European cultivars or American introductions registered in Europe, which were genotyped by the SoySNP50K array. The expected heterozygosity (He) was 0.34 for the entire collection and ranged among countries from 0.24 for Swiss cultivars to 0.32 for American cultivars (partly reflecting differences in sample size between countries). Cluster analysis grouped all genotypes into two main clusters with eight subgroups that corresponded to the country of origin of cultivars and their maturity group. Pairwise Fst values between countries of origin showed the highest differentiation of Swiss cultivars from the rest of the European gene pool, while the lowest mean differentiation was found between American introductions and all other European countries. On the other hand, Fst values between maturity groups were much lower compared to those observed between countries. In analysis of molecular variance, the total genetic variation was partitioned either by country of origin or by maturity group, explaining 9.1% and 3.5% of the total genetic variance, respectively. On the whole, our results suggest that the European soybean gene pool still has sufficient diversity due to the different historical breeding practices in western and eastern countries and the relatively short period of breeding in Europe.

White Lupin Adaptation to Moderately Calcareous Soils: Phenotypic Variation and Genome-Enabled Prediction.

White lupin is a promising high-protein crop, the cultivation of which is limited by a lack of adaptation to soils that are even just mildly calcareous. This study aimed to assess the phenotypic variation, the trait architecture based on a GWAS, and the predictive ability of genome-enabled models for grain yield and contributing traits of a genetically-broad population of 140 lines grown in an autumn-sown environment of Greece (Larissa) and a spring-sown environment of the Netherlands (Ens) that featured moderately calcareous and alkaline soils. We found large genotype × environment interaction and modest or nil genetic correlation for line responses across locations for grain yield, a lime susceptibility score, and other traits, with the exception of individual seed weight and plant height. The GWAS identified significant SNP markers associated with various traits that were markedly inconsistent across locations, while providing direct or indirect evidence for widespread polygenic trait control. Genomic selection proved to be a feasible strategy, owing to a moderate predictive ability for yield and lime susceptibility in Larissa (the site featuring greater lime soil stress). Other supporting results for breeding programs where the identification of a candidate gene for lime tolerance and the high reliability of genome-enabled predictions for individual seed weight.

White Lupin Drought Tolerance: Genetic Variation, Trait Genetic Architecture, and Genome-Enabled Prediction.

White lupin is a high-protein crop requiring drought tolerance improvement. This study focused on a genetically-broad population of 138 lines to investigate the phenotypic variation and genotype × environment interaction (GEI) for grain yield and other traits across drought-prone and moisture-favourable managed environments, the trait genetic architecture and relevant genomic regions by a GWAS using 9828 mapped SNP markers, and the predictive ability of genomic selection (GS) models. Water treatments across two late cropping months implied max. available soil water content of 60-80% for favourable conditions and from wilting point to 15% for severe drought. Line yield responses across environments featured a genetic correlation of 0.84. Relatively better line yield under drought was associated with an increased harvest index. Two significant QTLs emerged for yield in each condition that differed across conditions. Line yield under stress displayed an inverse linear relationship with the onset of flowering, confirmed genomically by a common major QTL. An adjusted grain yield computed as deviation from phenology-predicted yield acted as an indicator of intrinsic drought tolerance. On the whole, the yield in both conditions and the adjusted yield were polygenic, heritable, and exploitable by GS with a high predictive ability (0.62-0.78). Our results can support selection for climatically different drought-prone regions.

Genomic prediction and allele mining of agronomic and morphological traits in pea (Pisum sativum) germplasm collections.

Well-performing genomic prediction (GP) models for polygenic traits and molecular marker sets for oligogenic traits could be useful for identifying promising genetic resources in germplasm collections, setting core collections, and establishing molecular variety distinction. This study aimed at (i) defining GP models and key marker sets for predicting 15 agronomic or morphological traits in germplasm collections, (ii) verifying the GP model usefulness also for selection in breeding programs, (iii) investigating the consistency between molecular and phenotypic diversity patterns, and (iv) identifying genomic regions associated with to the target traits. The study was based on phenotyping data and over 41,000 genotyping-by-sequencing-generated SNP markers of 220 landraces or old cultivars belonging to a world germplasm collection and 11 modern cultivars. Non-metric multi-dimensional scaling (NMDS) and an analysis of population genetic structure indicated a high level of genetic differentiation of material from Western Asia, a major West-East diversity gradient, and quite limited genetic diversity of the improved germplasm. Mantel's test revealed a low correlation (r = 0.12) between phenotypic and molecular diversity, which increased (r = 0.45) when considering only the molecular diversity relative to significant SNPs from genome-wide association analyses. These analyses identified, inter alia, several areas of chromosome 6 involved in a largely pleiotropic control of vegetative or reproductive organ pigmentation. We found various significant SNPs for grain and straw yield under severe drought and onset of flowering, and one SNP on chromosome 5 for grain protein content. GP models displayed moderately high predictive ability (0.43 to 0.61) for protein content, grain and straw yield, and onset of flowering, and high predictive ability (0.76) for individual seed weight, based on intra-population, intra-environment cross-validations. The inter-population, inter-environment assessment of the models trained on the germplasm collection for breeding material of three recombinant inbred line (RIL) populations, which was challenged by much narrower diversity of the material, over eight-fold less available markers and quite different test environments, led to an overall loss of predictive ability of about 40% for seed weight, 50% for protein content and straw yield, and 60% for onset of flowering, and no prediction for grain yield. Within-RIL population predictive ability differed among populations.

Stacked kinship CNN vs. GBLUP for genomic predictions of additive and complex continuous phenotypes.

Deep learning is impacting many fields of data science with often spectacular results. However, its application to whole-genome predictions in plant and animal science or in human biology has been rather limited, with mostly underwhelming results. While most works focus on exploring alternative network architectures, in this study we propose an innovative representation of marker genotype data and tested it against the GBLUP (Genomic BLUP) benchmark with linear and nonlinear phenotypes. From publicly available cattle SNP genotype data, different types of genomic kinship matrices are stacked together in a 3D pile from where 2D grayscale slices are extracted and fed to a deep convolutional neural network (DNN). We simulated nine phenotype scenarios with combinations of additivity, dominance and epistasis, and compared the DNN to GBLUP-A (computed using only the additive kinship matrix) and GBLUP-optim (additive, dominance, and epistasis kinship matrices, as needed). Results varied depending on the accuracy metric employed, with DNN performing better in terms of root mean squared error (1-12% lower than GBLUP-A; 1-9% lower than GBLUP-optim) but worse in terms of Pearson's correlation (0.505 for DNN compared to 0.672 and 0.669 of GBLUP-A and GBLUP-optim for fully additive case; 0.274 for DNN, 0.279 for GBLUP-A, and 0.477 for GBLUP-optim for fully dominant case). The proposed approach offers a basis to explore further the application of DNN to tabular data in whole-genome predictions.

Alfalfa genomic selection for different stress-prone growing regions.

Alfalfa (Medicago sativa L.) selection for stress-prone regions has high priority for sustainable crop-livestock systems. This study assessed the genomic selection (GS) ability to predict alfalfa breeding values for drought-prone agricultural sites of Algeria, Morocco, and Argentina; managed-stress (MS) environments of Italy featuring moderate or intense drought; and one Tunisian site irrigated with moderately saline water. Additional aims were to investigate genotype × environment interaction (GEI) patterns and the effect on GS predictions of three single-nucleotide polymorphism (SNP) calling procedures, 12 statistical models that exclude or incorporate GEI, and allele dosage information. Our study included 127 genotypes from a Mediterranean reference population originated from three geographically contrasting populations, genotyped via genotyping-by-sequencing and phenotyped based on multi-year biomass dry matter yield of their dense-planted half-sib progenies. The GEI was very large, as shown by 27-fold greater additive genetic variance × environment interaction relative to the additive genetic variance and low genetic correlation for progeny yield responses across environments. The predictive ability of GS (using at least 37,969 SNP markers) exceeded 0.20 for moderate MS (representing Italian stress-prone sites) and the sites of Algeria and Argentina while being quite low for the Tunisian site and intense MS. Predictions of GS were complicated by rapid linkage disequilibrium decay. The weighted GBLUP model, GEI incorporation into GS models, and SNP calling based on a mock reference genome exhibited a predictive ability advantage for some environments. Our results support the specific breeding for each target region and suggest a positive role for GS in most regions when considering the challenges associated with phenotypic selection.

Genome-wide association study reveals white lupin candidate gene involved in anthracnose resistance.

KEY MESSAGE: GWAS identifies candidate gene controlling resistance to anthracnose disease in white lupin. White lupin (Lupinus albus L.) is a promising grain legume to meet the growing demand for plant-based protein. Its cultivation, however, is severely threatened by anthracnose disease caused by the fungal pathogen Colletotrichum lupini. To dissect the genetic architecture for anthracnose resistance, genotyping by sequencing was performed on white lupin accessions collected from the center of domestication and traditional cultivation regions. GBS resulted in 4611 high-quality single-nucleotide polymorphisms (SNPs) for 181 accessions, which were combined with resistance data observed under controlled conditions to perform a genome-wide association study (GWAS). Obtained disease phenotypes were shown to highly correlate with overall three-year disease assessments under Swiss field conditions (r > 0.8). GWAS results identified two significant SNPs associated with anthracnose resistance on gene Lalb_Chr05_g0216161 encoding a RING zinc-finger E3 ubiquitin ligase which is potentially involved in plant immunity. Population analysis showed a remarkably fast linkage disequilibrium decay, weak population structure and grouping of commercial varieties with landraces, corresponding to the slow domestication history and scarcity of modern breeding efforts in white lupin. Together with 15 highly resistant accessions identified in the resistance assay, our findings show promise for further crop improvement. This study provides the basis for marker-assisted selection, genomic prediction and studies aimed at understanding anthracnose resistance mechanisms in white lupin and contributes to improving breeding programs worldwide.

Merging genotyping-by-sequencing data from two ex situ collections provides insights on the pea evolutionary history.

Pea (Pisum sativum L. subsp. sativum) is one of the oldest domesticated species and a widely cultivated legume. In this study, we combined next generation sequencing (NGS) data referring to two genotyping-by-sequencing (GBS) libraries, each one prepared from a different Pisum germplasm collection. The selection of single nucleotide polymorphism (SNP) loci called in both germplasm collections caused some loss of information; however, this did not prevent the obtainment of one of the largest datasets ever used to explore pea biodiversity, consisting of 652 accessions and 22 127 markers. The analysis of population structure reflected genetic variation based on geographic patterns and allowed the definition of a model for the expansion of pea cultivation from the domestication centre to other regions of the world. In genetically distinct populations, the average decay of linkage disequilibrium (LD) ranged from a few bases to hundreds of kilobases, thus indicating different evolutionary histories leading to their diversification. Genome-wide scans resulted in the identification of putative selective sweeps associated with domestication and breeding, including genes known to regulate shoot branching, cotyledon colour and resistance to lodging, and the correct mapping of two Mendelian genes. In addition to providing information of major interest for fundamental and applied research on pea, our work describes the first successful example of integration of different GBS datasets generated from ex situ collections - a process of potential interest for a variety of purposes, including conservation genetics, genome-wide association studies, and breeding.

The colors of our brain: an integrated approach for dimensionality reduction and explainability in fMRI through color coding (i-ECO).

Several systematic reviews have highlighted the role of multiple sources in the investigation of psychiatric illness. For what concerns fMRI, the focus of recent literature preferentially lies on three lines of research, namely: functional connectivity, network analysis and spectral analysis. Data was gathered from the UCLA Consortium for Neuropsychiatric Phenomics. The sample was composed by 130 neurotypicals, 50 participants diagnosed with Schizophrenia, 49 with Bipolar disorder and 43 with ADHD. Single fMRI scans were reduced in their dimensionality by a novel method (i-ECO) averaging results per Region of Interest and through an additive color method (RGB): local connectivity values (Regional Homogeneity), network centrality measures (Eigenvector Centrality), spectral dimensions (fractional Amplitude of Low-Frequency Fluctuations). Average images per diagnostic group were plotted and described. The discriminative power of this novel method for visualizing and analyzing fMRI results in an integrative manner was explored through the usage of convolutional neural networks. The new methodology of i-ECO showed between-groups differences that could be easily appreciated by the human eye. The precision-recall Area Under the Curve (PR-AUC) of our models was > 84.5% for each diagnostic group as evaluated on the test-set - 80/20 split. In conclusion, this study provides evidence for an integrative and easy-to-understand approach in the analysis and visualization of fMRI results. A high discriminative power for psychiatric conditions was reached. This proof-of-work study may serve to investigate further developments over more extensive datasets covering a wider range of psychiatric diagnoses.

Pea Breeding for Intercropping With Cereals: Variation for Competitive Ability and Associated Traits, and Assessment of Phenotypic and Genomic Selection Strategies.

Mixed stand (MS) cropping of pea with small-grain cereals can produce more productive and environment-friendly grain crops relative to pure stand (PS) crops but may require selection to alleviate the pea competitive disadvantage. This study aimed to assess the pea variation for competitive ability and its associated traits and the efficiency of four phenotypic or genomic selection strategies. A set of 138 semi-leafless, semi-dwarf pea lines belonging to six recombinant inbred line populations and six parent lines were genotyped using genotyping-by-sequencing and grown in PS and in MS simultaneously with one barley and one bread wheat cultivar in two autumn-sown trials in Northern Italy. Cereal companions were selected in a preliminary study that highlighted the paucity of cultivars with sufficient earliness for association. Pea was severely outcompeted in both years albeit with variation for pea proportion ranging from nearly complete suppression (<3%) to values approaching a balanced mixture. Greater pea proportion in MS was associated with greater total yield of the mixture (r ≥ 0.46). The genetic correlation for pea yield across MS and PS conditions slightly exceeded 0.40 in both years. Later onset of flowering and taller plant height at flowering onset displayed a definite correlation with pea yield in MS (r ≥ 0.46) but not in PS, whereas tolerance to ascochyta blight exhibited the opposite pattern. Comparisons of phenotypic selection strategies within or across populations based on predicted or actual yield gains for independent years indicated an efficiency of 52-64% for indirect selection based on pea yield in PS relative to pea yield selection in MS. The efficiency of an indirect selection index including onset of flowering, plant height, and grain yield in PS was comparable to that of pea yield selection in MS. A genome-wide association study based on 5,909 SNP markers revealed the substantial diversity of genomic areas associated with pea yield in MS and PS. Genomic selection for pea yield in MS displayed an efficiency close to that of phenotypic selection for pea yield in MS, and nearly two-fold greater efficiency when also taking into account its shorter selection cycle and smaller evaluation cost.

Bacterial Community of Grana Padano PDO Cheese and Generical Hard Cheeses: DNA Metabarcoding and DNA Metafingerprinting Analysis to Assess Similarities and Differences.

The microbiota of Protected Designation of Origin (PDO) cheeses plays an essential role in defining their quality and typicity and could be applied to protect these products from counterfeiting. To study the possible role of cheese microbiota in distinguishing Grana Padano (GP) cheese from generical hard cheeses (HC), the microbial structure of 119 GP cheese samples was studied by DNA metabarcoding and DNA metafingerprinting and compared with 49 samples of generical hard cheeses taken from retail. DNA metabarcoding highlighted the presence, as dominant taxa, of Lacticaseibacillus rhamnosus, Lactobacillus helveticus, Streptococcus thermophilus, Limosilactobacillus fermentum, Lactobacillus delbrueckii, Lactobacillus spp., and Lactococcus spp. in both GP cheese and HC. Differential multivariate statistical analysis of metataxonomic and metafingerprinting data highlighted significant differences in the Shannon index, bacterial composition, and species abundance within both dominant and subdominant taxa between the two cheese groups. A supervised Neural Network (NN) classification tool, trained by metagenotypic data, was implemented, allowing to correctly classify GP cheese and HC samples. Further implementation and validation to increase the robustness and improve the predictive capacity of the NN classifier will be needed. Nonetheless, the proposed tool opens interesting perspectives in helping protection and valorization of GP and other PDO cheeses.

A High-Throughput Phenotyping Tool to Identify Field-Relevant Anthracnose Resistance in White Lupin.

The seed- and air-borne pathogen Colletotrichum lupini, the causal agent of lupin anthracnose, is the most important disease in white lupin (Lupinus albus) worldwide and can cause total yield loss. The aims of this study were to establish a reliable high-throughput phenotyping tool to identify anthracnose resistance in white lupin germplasm and to evaluate a genomic prediction model, accounting for previously reported resistance quantitative trait loci, on a set of independent lupin genotypes. Phenotyping under controlled conditions, performing stem inoculation on seedlings, showed to be applicable for high throughput, and its disease score strongly correlated with field plot disease assessments (r = 0.95, P < 0.0001) and yield (r = -0.64, P = 0.035). Traditional one-row field disease phenotyping showed no significant correlation with field plot disease assessments (r = 0.31, P = 0.34) and yield (r = -0.45, P = 0.17). Genomically predicted resistance values showed no correlation with values observed under controlled or field conditions, and the parental lines of the recombinant inbred line population used for constructing the prediction model exhibited a resistance pattern opposite to that displayed in the original (Australian) environment used for model construction. Differing environmental conditions, inoculation procedures, or population structure may account for this result. Phenotyping a diverse set of 40 white lupin accessions under controlled conditions revealed eight accessions with improved resistance to anthracnose. The standardized area under the disease progress curves (sAUDPC) ranged from 2.1 to 2.8, compared with the susceptible reference accession with a sAUDPC of 3.85. These accessions can be incorporated into white lupin breeding programs. In conclusion, our data support stem inoculation-based disease phenotyping under controlled conditions as a time-effective approach to identify field-relevant resistance, which can now be applied to further identify sources of resistance and their underlying genetics.

Evaluation of bacterial communities of Grana Padano cheese by DNA metabarcoding and DNA fingerprinting analysis.

The composition of the bacterial community of Grana Padano (GP) cheese was evaluated by an amplicon-based metagenomic approach (DNA metabarcoding) and RAPD-PCR fingerprinting. One hundred eighteen cheeses, which included 118 dairies located in the production area of GP, were collected. Two hundred fifty-four OTUs were detected, of which 82 were further discriminated between dominant (32 OTUs; > 1% total reads) and subdominant (50 OTUs; between 0.1% and 1% total reads) taxa. Lactobacillus (L.) delbrueckii, Lacticaseibacillus (Lact.) rhamnosus, Lact. casei, Limosilactobacillus fermentum, Lactococcus (Lc.) raffinolactis, L. helveticus, Streptococcus thermophilus, and Lc. lactis were the major dominant taxa ('core microbiota'). The origin of samples significantly impacted on both richness, evenness, and the relative abundance of bacterial species, with peculiar pattern distribution among the five GP production regions. A differential analysis allowed to find bacterial species significantly associated with specific region pairings. The analysis of pattern similarity among RAPD-PCR profiles highlighted the presence of a 'core' community banding pattern present in all the GP samples, which was strictly associated with the core microbiota highlighted by DNA metabarcoding. A trend to group samples according to the five production regions was also observed. This study widened our knowledge on the bacterial composition and ecology of Grana Padano cheese.

Pea Grain Protein Content Across Italian Environments: Genetic Relationship With Grain Yield, and Opportunities for Genome-Enabled Selection for Protein Yield.

Wider pea (Pisum sativum L.) cultivation has great interest for European agriculture, owing to its favorable environmental impact and provision of high-protein feedstuff. This work aimed to investigate the extent of genotype × environment interaction (GEI), genetically based trade-offs and polygenic control for crude protein content and grain yield of pea targeted to Italian environments, and to assess the efficiency of genomic selection (GS) as an alternative to phenotypic selection (PS) to increase protein yield per unit area. Some 306 genotypes belonging to three connected recombinant inbred line (RIL) populations derived from paired crosses between elite cultivars were genotyped through genotyping-by-sequencing and phenotyped for grain yield and protein content on a dry matter basis in three autumn-sown environments of northern or central Italy. Line variation for mean protein content ranged from 21.7 to 26.6%. Purely genetic effects, compared with GEI effects, were over two-fold larger for protein content, and over 2-fold smaller for grain and protein yield per unit area. Grain yield and protein content exhibited no inverse genetic correlation. A genome-wide association study revealed a definite polygenic control not only for grain yield but also for protein content, with small amounts of trait variation accounted for by individual loci. On average, the GS predictive ability for individual RIL populations based on the rrBLUP model (which was selected out of four tested models) using by turns two environments for selection and one for validation was moderately high for protein content (0.53) and moderate for grain yield (0.40) and protein yield (0.41). These values were about halved for inter-environment, inter-population predictions using one RIL population for model construction to predict data of the other populations. The comparison between GS and PS for protein yield based on predicted gains per unit time and similar evaluation costs indicated an advantage of GS for model construction including the target RIL population and, in case of multi-year PS, even for model training based on data of a non-target population. In conclusion, protein content is less challenging than grain yield for phenotypic or genome-enabled improvement, and GS is promising for the simultaneous improvement of both traits.

Development of PCR-based markers and whole-genome selection model for anthracnose resistance in white lupin (Lupinus albus L.).

White lupin (Lupinus albus L.) is a high-protein grain legume crop, grown since ancient Greece and Rome. Despite long domestication history, its cultivation remains limited, partly because of susceptibility to anthracnose. Only some late-flowering, bitter, low-yielding landraces from Ethiopian mountains displayed resistance to this devastating disease. The resistance is controlled by various genes, thereby complicating the breeding efforts. The objective of this study was developing tools for molecular tracking of Ethiopian resistance genes based on genotyping-by-sequencing (GBS) data, envisaging linkage mapping and genomic selection approaches. Twenty GBS markers from two major quantitative trait loci (QTLs), antr04_1/antr05_1 and antr04_2/antr05_2, were converted to PCR-based markers using assigned transcriptome sequences. Newly developed markers improved mapping resolution around both anthracnose resistance loci, providing more precise QTL estimation. PCR-based screening of diversified domesticated and primitive germplasm revealed the high specificity of two markers for the antr04_1/antr05_1 locus (TP222136 and TP47110) and one for the antr04_2/antr05_2 locus (TP338761), highlighted by simple matching coefficients of 0.96 and 0.89, respectively. Moreover, a genomic selection approach based on GBS data of a recombinant inbred line mapping population was assessed, providing an average predictive ability of 0.56. These tools can be used for preselection of candidate white lupin germplasm for anthracnose resistance assays.

Development and Proof-of-Concept Application of Genome-Enabled Selection for Pea Grain Yield under Severe Terminal Drought.

Terminal drought is the main stress limiting pea (Pisum sativum L.) grain yield in Mediterranean environments. This study aimed to investigate genotype × environment (GE) interaction patterns, define a genomic selection (GS) model for yield under severe drought based on single nucleotide polymorphism (SNP) markers from genotyping-by-sequencing, and compare GS with phenotypic selection (PS) and marker-assisted selection (MAS). Some 288 lines belonging to three connected RIL populations were evaluated in a managed-stress (MS) environment of Northern Italy, Marchouch (Morocco), and Alger (Algeria). Intra-environment, cross-environment, and cross-population predictive ability were assessed by Ridge Regression best linear unbiased prediction (rrBLUP) and Bayesian Lasso models. GE interaction was particularly large across moderate-stress and severe-stress environments. In proof-of-concept experiments performed in a MS environment, GS models constructed from MS environment and Marchouch data applied to independent material separated top-performing lines from mid- and bottom-performing ones, and produced actual yield gains similar to PS. The latter result would imply somewhat greater GS efficiency when considering same selection costs, in partial agreement with predicted efficiency results. GS, which exploited drought escape and intrinsic drought tolerance, exhibited 18% greater selection efficiency than MAS (albeit with non-significant difference between selections) and moderate to high cross-population predictive ability. GS can be cost-efficient to raise yields under severe drought.