Machine learning methods in near infrared spectroscopy for predicting sensory traits in sweetpotatoes.

It has been established that near infrared (NIR) spectroscopy has the potential of estimating sensory traits given the direct spectral responses that these properties have in the NIR region. In sweetpotato, sensory and texture traits are key for improving acceptability of the crop for food security and nutrition. Studies have statistically modelled the levels of NIR spectroscopy sensory characteristics using partial least squares (PLS) regression methods. To improve prediction accuracy, there are many advanced techniques, which could enhance modelling of fresh (wet and un-processed) samples or nonlinear dependence relationships. Performance of different quantitative prediction models for sensory traits developed using different machine learning methods were compared. Overall, results show that linear methods; linear support vector machine (L-SVM), principal component regression (PCR) and PLS exhibited higher mean R2 values than other statistical methods. For all the 27 sensory traits, calibration models using L-SVM and PCR has slightly higher overall R2 (x¯ = 0.33) compared to PLS (x¯ = 0.32) and radial-based SVM (NL-SVM; x¯= 0.30). The levels of orange color intensity were the best predicted by all the calibration models (R2 = 0.87 - 0.89). The elastic net linear regression (ENR) and tree-based methods; extreme gradient boost (XGBoost) and random forest (RF) performed worse than would be expected but could possibly be improved with increased sample size. Lower average R2 values were observed for calibration models of ENR (x¯ = 0.26), XGBoost (x¯ = 0.26) and RF (x¯ = 0.22). The overall RMSE in calibration models was lower in PCR models (X = 0.82) compared to L-SVM (x¯ = 0.86) and PLS (x¯ = 0.90). ENR, XGBoost and RF also had higher RMSE (x¯ = 0.90 - 0.92). Effective wavelengths selection using the interval partial least-squares regression (iPLS), improved the performance of the models but did not perform as good as the PLS. SNV pre-treatment was useful in improving model performance.

Using machine learning for image-based analysis of sweetpotato root sensory attributes.

The sweetpotato breeding process involves assessing different phenotypic traits, such as the sensory attributes, to decide which varieties to progress to the next stage during the breeding cycle. Sensory attributes like appearance, taste, colour and mealiness are important for consumer acceptability and adoption of new varieties. Therefore, measuring these sensory attributes is critical to inform the selection of varieties during breeding. Current methods using a trained human panel enable screening of different sweetpotato sensory attributes. Despite this, such methods are costly and time-consuming, leading to low throughput, which remains the biggest challenge for breeders. In this paper, we describe an approach to apply machine learning techniques with image-based analysis to predict flesh-colour and mealiness sweetpotato sensory attributes. The developed models can be used as high-throughput methods to augment existing approaches for the evaluation of flesh-colour and mealiness for different sweetpotato varieties. The work involved capturing images of boiled sweetpotato cross-sections using the DigiEye imaging system, data pre-processing for background elimination and feature extraction to develop machine learning models to predict the flesh-colour and mealiness sensory attributes of different sweetpotato varieties. For flesh-colour the trained Linear Regression and Random Forest Regression models attained R2 values of 0.92 and 0.87, respectively, against the ground truth values given by a human sensory panel. In contrast, the Random Forest Regressor and Gradient Boosting model attained R2 values of 0.85 and 0.80, respectively, for the prediction of mealiness. The performance of the models matched the desirable R2 threshold of 0.80 for acceptable comparability to the human sensory panel showing that this approach can be used for the prediction of these attributes with high accuracy. The machine learning models were deployed and tested by the sweetpotato breeding team at the International Potato Center in Uganda. This solution can automate and increase throughput for analysing flesh-colour and mealiness sweetpotato sensory attributes. Using machine learning tools for analysis can inform and quicken the selection of promising varieties that can be progressed for participatory evaluation during breeding cycles and potentially lead to increased chances of adoption of the varieties by consumers.

End-user preferences to enhance prospects for varietal acceptance and adoption in potato breeding in Uganda.

BACKGROUND: Potato varieties have diverse biophysical characteristics, so it is important for breeders to have the capacity to choose those that meet the preferences of end users, such as mealiness, firmness, and taste, among others. Combining user preferences with descriptive information regarding the sensory characteristics of boiled potatoes can contribute to the improvement of consumer-driven varieties. This study aimed to factor in the preferences of end users to improve the prospects for varietal acceptance, adoption, and discrimination among genotypes in potato breeding. RESULTS: The priority quality traits (traits that play the most significant roles in acceptance and adoption) of the boiled potatoes were determined by evaluating gender and livelihood using the G+ tool. The G+ tool is designed to assess gender impact on roots, tubers and bananas (RTB) traits by serving as a validation check to reflect on important gender-based issues in agricultural food systems in order to reduce harm and promote positive impact. Potato genotypes were differentiated by penetration (textural parameters as measured by standard texture probe) and the procedure was repeatable, as there was no significant difference between the cooking replicates at 40 min of cooking. Instrument-based texture parameters, such as penetration peak force (hardness/firmness) and area (area under the curve, which represents energy needed to penetrate) of boiled potato tubers were significantly associated with sensory attributes such as fracturability and hardness in the mouth. An attempt to differentiate genotypes using near-infrared spectroscopy (NIRS) revealed that the average results observed for the calibration for yellow color (r2 = 0.70), homogeneity of color (r2 = 0.48), moisture in mass (r2 = 0.40), and uniformity of texture (r2 = 0.56) suggested that these parameters could be used for initial breeding screening purposes. CONCLUSIONS: The preferred traits of the boiled potato can be integrated into the potato-breeding program/product profile. Near-infrared spectroscopy shows strong potential to predict potato color and the ability of NIRS models to predict some texture attributes is also promising. © 2023 The Authors. Journal of The Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Genetic diversity and population structure of Vernonia amygdalina Del. in Uganda based on genome wide markers.

Determining the extent and distribution of genetic diversity is an essential component of plant breeding. In the present study, we explored the genetic diversity and population structure of Vernonia amygdalina, a fodder, vegetable and medicinal species of Africa and some parts of Yemen. Most empirical studies demonstrate that populations that are separated by geographic or ecological factors may experience genetic differentiation resulting from restricted gene flow between populations. A total of 238 individuals were sampled from two populations: i) Lake Victoria crescent (LVC) and ii) Southern and Eastern Lake Kyoga basin (SEK) agroecological zones of Uganda and genotyped using DArT platform. Of the two populations, the overall mean observed heterozygosity (Ho) was low to medium (Ho = 0.07[silicoDArTs] and 0.2[SNPs]). Inbreeding levels were also very low (-0.04 to -0.08) suggesting the presence of random mating. Partitioning of genetic structure in the two populations indicated that SEK exhibited a higher genetic diversity than LVC. The principal coordinates analysis (PCA) showed no geographical structuring, consistent with the low genetic differentiation (Fst = 0.00) and the low Euclidean genetic distance (1.38-1.39) between the LVC and SEK populations. However, STRUCTURE analysis with admixture models revealed weak possible genetic clusters with very small genetic distance among them. Overall, the results suggest low genetic diversity and weak genetic differentiation between the two populations. One possible explanation of the results could be the presence of human assisted gene flow over long distances.

What should the standard be for passing and mastery on the Critical Thinking about Health Test? A consensus study.

OBJECTIVE: Most health literacy measures rely on subjective self-assessment. The Critical Thinking about Health Test is an objective measure that includes two multiple-choice questions (MCQs) for each of the nine Informed Health Choices Key Concepts included in the educational resources for secondary schools. The objective of this study was to determine cut-off scores for passing (the border between having and not having a basic understanding and the ability to apply the nine concepts) and mastery (the border between having mastered and not having mastered them). DESIGN: Using a combination of two widely used methods: Angoff's and Nedelsky's, a panel judged the likelihood that an individual on the border of passing and another on the border of having mastered the concepts would answer each MCQ correctly. The cut-off scores were determined by summing up the probability of answering each MCQ correctly. Their independent assessments were summarised and discussed. A nominal group technique was used to reach a consensus. SETTING: The study was conducted in secondary schools in East Africa. PARTICIPANTS: The panel included eight individuals with 5 or more years' experience in the following areas: evaluation of critical thinking interventions, curriculum development, teaching of lower secondary school and evidence-informed decision-making. RESULTS: The panel agreed that for a passing score, students had to answer 9 of the 18 questions and for a mastery score, 14 out of 18 questions correctly. CONCLUSION: There was wide variation in the judgements made by individual panel members for many of the questions, but they quickly reached a consensus on the cut-off scores after discussions.

Genomic selection for resistance to mammalian bark stripping and associated chemical compounds in radiata pine.

The integration of genomic data into genetic evaluations can facilitate the rapid selection of superior genotypes and accelerate the breeding cycle in trees. In this study, 390 trees from 74 control-pollinated families were genotyped using a 36K Axiom SNP array. A total of 15,624 high-quality SNPs were used to develop genomic prediction models for mammalian bark stripping, tree height, and selected primary and secondary chemical compounds in the bark. Genetic parameters from different genomic prediction methods-single-trait best linear unbiased prediction based on a marker-based relationship matrix (genomic best linear unbiased prediction), multitrait single-step genomic best linear unbiased prediction, which integrated the marker-based and pedigree-based relationship matrices (single-step genomic best linear unbiased prediction) and the single-trait generalized ridge regression-were compared to equivalent single- or multitrait pedigree-based approaches (ABLUP). The influence of the statistical distribution of data on the genetic parameters was assessed. Results indicated that the heritability estimates were increased nearly 2-fold with genomic models compared to the equivalent pedigree-based models. Predictive accuracy of the single-step genomic best linear unbiased prediction was higher than the ABLUP for most traits. Allowing for heterogeneity in marker effects through the use of generalized ridge regression did not markedly improve predictive ability over genomic best linear unbiased prediction, arguing that most of the chemical traits are modulated by many genes with small effects. Overall, the traits with low pedigree-based heritability benefited more from genomic models compared to the traits with high pedigree-based heritability. There was no evidence that data skewness or the presence of outliers affected the genomic or pedigree-based genetic estimates.

SilicoDArT and SNP markers for genetic diversity and population structure analysis of Trema orientalis; a fodder species.

Establishing the genetic diversity and population structure of a species can guide the selection of appropriate conservation and sustainable utilization strategies. Next-generation sequencing (NGS) approaches are increasingly being used to generate multi-locus data for genetic structure determination. This study presents the genetic structure of a fodder species -Trema orientalis based on two genome-wide high-throughput diversity array technology (DArT) markers; silicoDArT and single nucleotide polymorphisms (SNPs). Genotyping of 119 individuals generated 40,650 silicoDArT and 4767 SNP markers. Both marker types had a high average scoring reproducibility (>99%). Genetic relationships explored by principal coordinates analysis (PCoA) showed that the first principal coordinate axis explained most of the variation in both the SilicoDArT (34.2%) and SNP (89.6%) marker data. The average polymorphic information content did not highly differ between silicoDArT (0.22) and SNPs (0.17) suggesting minimal differences in informativeness in the two groups of markers. The, mean observed (Ho) and expected (He) heterozygosity were low and differed between the silicoDArT and SNPs respectively, estimated at Ho = 0.08 and He = 0.05 for silicoDArT and Ho = 0.23 and He = 0.19 for SNPs. The population of T. orientalis was moderately differentiated (FST = 0.20-0.53) and formed 2 distinct clusters based on maximum likelihood and principal coordinates analysis. Analysis of molecular variance revealed that clusters contributed more to the variation (46.3-60.8%) than individuals (32.9-31.2%). Overall, the results suggest a high relatedness of the individuals sampled and a threatened genetic potential of T. orientalis in the wild. Therefore, genetic management activities such as ex-situ germplasm management are required for the sustainability of the species. Ex-situ conservation efforts should involve core collection of individuals from different populations to capture efficient diversity. This study demonstrates the importance of silicoDArT and SNP makers in population structure and genetic diversity analysis of Trema orientalis, useful for future genome wide studies in the species.

Nutritional prospects of jackfruit and its potential for improving dietary diversity in Uganda.

OBJECTIVE: A sustainable way of providing essential nutrients from crops for the poor and undernourished is biofortification, through plant breeding. This study characterised the intraspecific variation of selected nutritional elements in the flakes and seeds of Ugandan jackfruit (Artocarpus heterophyllus) plus the phytochemical composition of leaves. The overall aim was to explore possibilities of selecting for varieties that are higher in selected essential nutrients. Selecting for nutrient dense crops has been mostly done for annual agricultural crops, and rarely for perennial fruit trees. RESULTS: Uganda's Jackfruit has high macronutrients, especially magnesium and calcium. This study revealed that the amounts of these macronutrients were higher than those found in commonly consumed fruits, giving jackfruit a nutritional advantage with respect to these nutrients. The varieties sampled also differed significantly (p < 0.01) for some nutrients such as vitamin C, crude fat, crude fibre, total soluble solids and juice yield, highlighting the potential for selection for targeted nutritional gains. The seeds however, had less amounts of most of the quantified nutrients that also differed among the varieties. Significant intraspecific variation of the leaf total phenolics was also observed. With regard to the quantified nutritional elements in the flakes, the ethno-varieties were separated in space along PC1 (p < 0.001), PC2 (p < 0.001) and PC3 (p < 0.01) indicating their distinctness.

Chemical Traits that Predict Susceptibility of Pinus radiata to Marsupial Bark Stripping.

Bark stripping by mammals is a major problem in managed conifer forests worldwide. In Australia, bark stripping in the exotic plantations of Pinus radiata is mainly caused by native marsupials and results in reduced survival, growth, and in extreme cases death of trees. Herbivory is influenced by a balance between primary metabolites that are sources of nutrition and secondary metabolites that act as defences. Identifying the compounds that influence herbivory may be a useful tool in the management of forest systems. This study aimed to detect and identify both constitutive and induced compounds that are associated with genetic differences in susceptibility of two-year-old P. radiata trees to bark stripping by marsupials. An untargeted profiling of 83 primary and secondary compounds of the needles and bark samples from 21 susceptible and 21 resistant families was undertaken. These were among the most and least damaged families, respectively, screened in a trial of 74 families that were exposed to natural field bark stripping by marsupials. Experimental plants were in the same field trial but protected from bark stripping and a subset were subjected to artificial bark stripping to examine induced and constitutive chemistry differences between resistant and susceptible families. Machine learning (random forest), partial least squares plus discriminant analysis (PLS-DA), and principal components analysis with discriminant analysis (PCA-DA), as well as univariate methods were used to identify the most important totals by compound group and individual compounds differentiating the resistant and susceptible families. In the bark, the constitutive amount of two sesquiterpenoids - bicyclogermacrene and an unknown sesquiterpenoid alcohol -were shown to be of higher levels in the resistant families, whereas the constitutive sugars, fructose, and glucose, as well individual phenolics, were higher in the more susceptible families. The chemistry of the needles was not useful in differentiating the resistant and susceptible families to marsupial bark stripping. After artificial bark stripping, the terpenes, sugars, and phenolics responded in both the resistant and susceptible families by increasing or reducing amounts, which leveled the differences in the amounts of the compounds between the different resistant and susceptible classes observed at the constitutive level. Overall, based on the families with extreme values for less and more susceptibility, differences in the amounts of secondary compounds were subtle and susceptibility due to sugars may outweigh defence as the cause of the genetic variation in bark stripping observed in this non-native tree herbivory system.

Additive genetic variation in Pinus radiata bark chemistry and the chemical traits associated with variation in mammalian bark stripping.

Secondary metabolites are suggested as a major mechanism explaining genetic variation in herbivory levels in Pinus radiata. The potential to incorporate these chemical traits into breeding/deployment programmes partly depends on the presence of additive genetic variation for the relevant chemical traits. In this study, near-infrared spectroscopy was used to quantify the constitutive and induced levels of 54 compounds in the bark of trees from 74 P. radiata full-sib families. The trees sampled for chemistry were protected from browsing and induced levels were obtained by subjecting half of the trees to artificial bark stripping. The treatment effect on bark chemistry was assessed along with narrow-sense heritability, the significance of non-additive genetic effects and the additive genetic correlations of compounds with bark stripping by mammalian herbivores that was observed in unprotected replicates of the field trial. The results indicated: (i) significant additive genetic variation, with low-moderate narrow-sense heritability estimates for most compounds; (ii) while significant induced effects were detected for some chemicals, no significant genetic variation in inducibility was detected; and (iii) sugars, fatty acids and a diterpenoid positively genetically correlated while a sesquiterpenoid negatively genetically correlated with bark stripping by the mammalian herbivore, the Bennett's wallaby (Macropus rufogriseus). At the onset of browsing, a trade-off with height was detected for selecting higher amounts of this sesquiterpenoid. However, overall, results showed potential to incorporate chemical traits into breeding/deployment programmes. The quantitative genetic analyses of the near infrared predicted chemical traits produced associations with mammalian bark stripping that mostly conform with those obtained using standard wet chemistry.

Variability of phenolic and alkaloid content in different plant parts of Carissa edulis Vahl and Zanthoxylum chalybeum Engl.

OBJECTIVE: The objective of the study was to investigate the relative abundance and effect of post-harvest treatment on total phenolics (TP) and total alkaloids in the leaves and bark of Carissa edulis and Zanthoxylum chalybeum, which would give an indication of the suitability of leaves as alternative sources of medicine in these plant species. RESULTS: Results indicated higher levels of total phenolics than total alkaloids in both of the species under both freezing and air drying conditions. While more alkaloids were found in leaves compared to bark, there was no difference in abundance of phenols between the plant parts of both species. Air drying preserved more TPs than freezing and the opposite was true for alkaloids. For sustainability, leaves are recommended as an alternative source of medicine instead of the preferred root or stem bark. However, the choice of whether to dry or freeze will depend on the specific compound of interest. Assessment of spatial variability of medicinal properties is highly recommended.

Detection of Self Incompatibility Genotypes in Prunus africana: Characterization, Evolution and Spatial Analysis.

In flowering plants, self-incompatibility is an effective genetic mechanism that prevents self-fertilization. Most Prunus tree species exhibit a homomorphic gametophytic self-incompatibility (GSI) system, in which the pollen phenotype is encoded by its own haploid genome. To date, no identification of S-alleles had been done in Prunus africana, the only member of the genus in Africa. To identify S-RNase alleles and hence determine S-genotypes in African cherry (Prunus africana) from Mabira Forest Reserve, Uganda, primers flanking the first and second intron were designed and these amplified two bands in most individuals. PCR bands on agarose indicated 26 and 8 different S-alleles for second and first intron respectively. Partial or full sequences were obtained for all these fragments. Comparison with published S-RNase data indicated that the amplified products were S-RNase alleles with very high interspecies homology despite the high intraspecific variation. Against expectations for a locus under balancing selection, frequency and spatial distribution of the alleles in a study plot was not random. Implications of the results to breeding efforts in the species are discussed, and mating experiments are strongly suggested to finally prove the functionality of SI in P. africana.