Using unsupervised learning techniques to assess interactions among complex traits in soybeans

Soybean yield components and agronomic traits are connected through physiological pathways that impose tradeoffs through genetic and environmental constraints. Our primary aim is to assess the interdependence of soybean traits by using unsupervised machine learning techniques to divide phenotypic associations into environmental and genetic associations. This study was performed on large scale, jointly analyzing 14 quantitative traits in a large multi-parental population designed for genetic studies. We collected phenotypes from 2012 to 2015 from a soybean nested association panel with 40 families of approximately 140 individuals each. Pearson and Spearman correlations measured phenotypic associations. A multivariate mixed linear model provided genotypic and environmental correlations. To evaluate relationships among traits, the study used principal component and undirected graphical models from phenotypic, genotypic, and environmental correlation matrices. Results indicate that high phenotypic correlation occurs when traits display both genetic and environmental correlations. In genetic terms, length of reproductive period, node number, and canopy coverage play important roles in determining yield potential. Optimal grain yield production occurs when the growing environment favors faster canopy closure and extended reproductive length. Environmental associations found among yield components give insight into the nature of yield component compensation. The use of unsupervised learning methods provides a good framework for investigating interactions among various quantitative traits and defining target traits for breeding.     

Improved genetic parameter estimations in zoysiagrass by implementing post hoc blocking

Randomized complete block (RCB) design is the most widely used experimental design in biological sciences. As number of treatments increases, the block size become larger and it looses the capacity to control the variance within block, which is its original purpose. A method known as post hoc blocking could be used in these cases to improve the genetic parameter estimation and thus obtain an unbiased assessment of the performance of a given treatment. In trufgrass breeding, as other breeding program, this is a common challenge. The goal of this study was to test the capacity of different post hoc blocking designs to improve the genetic parameter estimation of zoysiagrass (Zoysia spp.). We evaluated two post hoc blocking designs; row–column (R–C) and incomplete block (IB) designs on five genotype trials located in Florida. The results showed that post hoc R–C design had superior model fitting than both the original RCB and the post hoc IB designs when studied at the single measurement level and at the site level. The narrow-sense heritability (0.24–0.40) and the genotype-by-measurement correlation (0.57–0.99) did not change significantly when R–C was compared to the original RCB design. The ranking of the top performing genotypes changed considerably when comparing RCB to R–C design, but the degree depended on the location analyzed. We conclude that the change in the ranking of the top (potentially select individuals) is coming from the better control of intra-block environmental variation, and this could potentially have a significant impact on the breeding selection process.     

Reaction of <Emphasis Type="Italic">Phaseolus</Emphasis> spp. genotypes to ashy stem blight caused by <Emphasis Type="Italic">Macrophomina phaseolina</Emphasis>

Ashy stem blight (ASB) caused by Macrophomina phaseolina (Tassi) Goidanich (Mp) is a devastating seed-transmitted disease in common bean in the tropics. The identification of resistant cultivars throughout the cropping season contributes to disease management. Resistance is found in the primary and tertiary gene pools. Our objectives were to determine (1) the reaction of Phaseolus spp. genotypes to two Mp isolates at vegetative and reproductive stages, (2) the area under disease progress curve (AUDPC), and (3) resistant plants per genotype at harvest. Twenty-three genotypes from different origins were planted in the greenhouse in 2016 and 2017. One less-aggressive Mp (PRJD16) and one more-aggressive (PRI16) isolate were inoculated one and three times, respectively, by the cut-stem method. ‘Beníquez’, ‘Othello’, and ‘Verano’ were highly susceptible (mean scores >8.0, and AUDPC values from 264.6 to 300.8) to both isolates. BAT 477 and NY6020-4 were intermediate (5.6 and 6.2; AUDPC: 161.6 and 187.1) to PRJD16 and susceptible (7.4 and 8.2; AUDPC: 209.4 and 235.1) to PRI16. Resistant genotypes (mean scores ≤3) were not identified in this study. However, A 195, ‘Badillo’, and ‘PC 50’ possessed lower mean scores (4.3–5.4) and AUDPC values (126.4–150.9) to both isolates. Furthermore, A 195 had the highest percentage of resistant plants (55.6%) followed by PC 50, I9365-31, and PI 321637 (27.8%) to PRJD16 at harvest. Thus, the identification of resistant parents across Phaseolus species is necessary to increase the levels of ASB resistance in common bean cultivars throughout the entire cropping season.     

Tall fescue forage mass in a grass-legume mixture: predicted efficiency of indirect selection

High fertilizer prices and improved environmental stewardship have increased interest in grass-legume mixed pastures. It has been hypothesized, but not validated, that the ecological combining ability between grasses and legumes can be improved by breeding specifically for mixture performance. This experiment examined the predicted efficiency of selection in a grass monoculture environment to indirectly improve tall fescue (Lolium arundinaceum (Schreb.) Darbysh.) forage mass in a grass-legume mixture. Heritability, genetic and rank correlations, and selection efficiencies were estimated for forage mass in a tall fescue half-sib population grown as spaced-plants overseeded with either turf-type tall fescue (monoculture) or alfalfa (mixture). Heritability for tall fescue forage mass in monoculture ranged from 0.32 to 0.70 and were always similar or greater than those in mixture (range 0.27–0.55) for four successive harvests and annual total. Genetic correlations between monoculture and mixture tall fescue forage mass varied with values of 0.48, 0.92, ?0.31, 0.70, and 0.25 in June, July, August, October, and annual total, respectively. Indirect selection efficiencies exceeded or approached direct selection for mixtures only in July and October (1.29, and 0.73, respectively). Whereas, indirect selection efficiencies were low in June, August, and annual forage mass (0.58, ?0.31, and 0.28, respectively). Moreover, low Spearman’s rank correlations (?0.03 to 0.35) indicated differing half-sib family performance between the monoculture and mixture environments. Results indicate that direct selection should be used to improve tall fescue forage mass in a grass-legume mixture, and support the hypothesis of increasing ecological combining ability by breeding for mixtures per se.     

Molecular markers for improving control of soil-borne pathogen <Emphasis Type="Italic">Fusarium oxysporum</Emphasis> in sugar beet

Fusarium spp. cause severe damage in many agricultural crops, including sugar beet, with Fusarium oxysporum historically being considered as the most damaging of all species. Sugar beet needs to be protected from this class of soil-borne pathogens in order to ensure an optimal sugar yield in the field. Genetic control of the disease is crucial in managing these pathogens. Identification of single nucleotide polymorphism (SNP) markers linked to resistance can be a powerful tool for the introgression of valuable genes needed to develop Fusarium-resistant varieties. A candidate gene approach was carried out to identify SNP markers linked to putative Fusarium resistance sources in sugar beet. Five resistant analogue genes (RGAs) were screened by means of high resolution melting (HRM) analysis in a set of sugar beet lines, considered as resistant and susceptible to Fusarium oxysporum. HRM polymorphisms were observed in 80% of amplicons. Two HRM polymorphisms were significantly associated with Fusarium resistance (P < 0.05). The amplicons that showed association were sequenced and two SNPs were identified. The association was further validated on 96 susceptible and 96 resistant plants using competitive allele-specific PCR (KASPar) technology. The selected SNPs could be used for marker-assisted breeding of Fusarium resistance in sugar beet.     

Identification of quantitative trait locus and prediction of candidate genes for grain mineral concentration in maize across multiple environments

Trace metal elements are essential in daily diets for human health and normal growth. Maize is staple food for people in many countries. However, maize has low mineral concentration which makes it difficult to meet human requirements for micronutrients. The objective of this study was to identify quantitative trait locus (QTL) and predict candidate genes associated with mineral concentration in maize grain. Here, a recombinant inbred line population was used to test phenotype of zinc (Zn), iron (Fe), copper (Cu) and manganese (Mn) concentrations in six environments and then a QTL analysis was conducted using single environment analysis along with multiple environment trial (MET) analysis. These two strategies detected a total of 64 and 67 QTLs for target traits, respectively. Single environment analysis revealed 13 QTL bins distributed on seven chromosomes. We found that five candidate genes associated with mineral concentration were located in the same intervals identified by Comparative mapping of meta-QTLs in our previous study. The genetic and phenotypic correlation coefficients were depended on the nutrient traits and they were significant between Fe and Zn, Fe and Cu, Fe and Mn in all environments. The results of this study illustrated the genetic correlation between maize grain mineral concentrations, and identified some promising genomic regions and candidate genes for further studies on the biofortification of mineral concentration in maize grain.     

Determining resistance to <Emphasis Type="Italic">Fusarium verticillioides</Emphasis> and fumonisin accumulation in African maize inbred lines resistant to <Emphasis Type="Italic">Aspergillus flavus</Emphasis> and aflatoxins

Fusarium verticillioides and Aspergillus flavus cause Fusarium ear rot (FER) and Aspergillus ear rot (AER) of maize, respectively. Both pathogens are of concern to producers as they reduce grain yield and affect quality. F. verticillioides and A. flavus also contaminate maize grain with the mycotoxins fumonisins and aflatoxins, respectively, which has been associated with mycotoxicosis in humans and animals. The occurrence of common resistance mechanisms to FER and AER has been reported. Hence, ten Kenyan inbred lines resistant to AER and aflatoxin accumulation were evaluated for resistance to FER, F. verticillioides colonisation and fumonisin accumulation; and compared to nine South African lines resistant to FER and fumonisin accumulation. Field trials were conducted at three localities in South Africa and two localities in Kenya. FER severity was determined by visual assessment, while F. verticillioides colonisation and fumonisin content were quantified by real-time PCR and liquid chromatography tandem mass spectrometry, respectively. Significant genotype x environment interactions was determined at each location (P ≤ 0.05). Kenyan inbred CML495 was most resistant to FER and F. verticillioides colonisation, and accumulated the lowest concentration of fumonisins across localities. It was, however, not significantly more resistant than Kenyan lines CML264 and CKL05015, and the South African line RO549 W, which also exhibited low FER severity (≤5%), fungal target DNA (≤0.025 ng μL^?1) and fumonisin levels (≤2.5 mg kg^?1). Inbred lines resistant to AER and aflatoxin accumulation appear to be promising sources of resistance to F. verticillioides and fumonisin contamination.     

Resistance screening of breeding lines and commercial tomato cultivars for <Emphasis Type="Italic">Meloidogyne incognita</Emphasis> and <Emphasis Type="Italic">M. javanica</Emphasis> populations (Nematoda) from Ethiopia

Soil and root samples were collected from major tomato growing areas of Ethiopia during the 2012/2013 growing season to identify root-knot nematode problems. DNA-based and isozyme techniques revealed that Meloidogyne incognita and M. javanica were the predominant Meloidogyne species across the sampled areas. The aggressiveness of different populations of these species was assessed on tomato cultivars Marmande and Moneymaker. The two most aggressive populations of each species were selected and further tested on 33 tomato genotypes. The resistance screening and mechanism of resistance was performed after inoculation with 100 freshly hatched (<24 h) second-stage juveniles (J2). Eight weeks after inoculation the number of egg masses produced on each cultivar was assessed. For the resistance mechanism study, J2 penetration and their subsequent development inside the tomato roots were examined at 1, 2, 4 and 6 weeks after inoculation. On both cultivars Marmande and Moneymaker all M. incognita and M. javanica populations formed a high number of egg masses indicating highly aggressive behaviour. Populations from ‘Jittu’ and ‘Babile’ for M. incognita and ‘Jittu’ and ‘Koka’ for M. javanica were selected as most aggressive. None of the 33 tomato genotypes were immune for these M. incognita and M. javanica populations. However, several tomato genotypes were found to have a significant effect on the number of egg masses produced indicating possible resistance. For M. javanica populations there were more plants from cultivars or breeding lines on which no egg masses were found compared to M. incognita populations. The lowest number of egg masses for both populations of M. incognita was produced on cultivars Bridget40, Galilea, and Irma while for M. javanica it was on Assila, Eden, Galilea, Tisey, CLN-2366A, CLN-2366B and CLN-2366C. Tomato genotypes, time (weeks after inoculation) and their interaction were significant sources of variation for J2 penetration and their subsequent development inside the tomato roots. Differential penetration was found in breeding lines such as CLN-2366A, CLN-2366B and CLN-2366C, but many of the selected tomato genotypes resistance for the tested M. incognita and M. javanica populations were expressed by delayed nematode development. Therefore, developing a simple screening technique to be used by local farmers or extension workers is crucial to facilitate selection of a suitable cultivar.     

Genetic relationships among resynthesized,semi-resynthesized and natural <Emphasis Type="Italic">Brassica napus</Emphasis> L. genotypes

The allopolyploidization event that created cultivated oilseed rape Brassica napus L, followed by intense breeding, reduced its genetic diversity. Resynthesized (RS) B. napus L. obtained by interspecific hybridization between genotypes of B. rapa L. and B. oleracea L. can be a valuable source for broadening genetic diversity in cultivated oilseed rape. In this study, we determined the extent of DNA polymorphism among natural accessions of oilseed rape, resynthesized B. napus, their parental species and double-low quality semi-RS lines carrying the Rfo gene. Using 10 selected primer combinations, 522 polymorphic AFLP markers were scored in the complete set of 100 Brassica sp. To detect relationships between these genotypes, a cluster analysis was performed using the Jaccard’s distance. Resynthesized allopolyploids clustered directly between their diploid parents. Cultivated accessions of oilseed rape created a compact group away from resynthesized allopolyploids as well as semi-RS lines. The natural oilseed rape group, which consists of 49 cultivars and breeding lines of oilseed rape, is characterized by lower genetic diversity than the group of 33 accessions of resynthesized oilseed rape, and the analysis showed that the double-low quality semi-RS lines represent a specific genetic variation of B. napus. The de novo resynthesized B. napus lines and the semi-RS lines of double-low quality generated from them, provide a significant opportunity for enrichment the gene pool of oilseed rape.