US of thyroid nodules: can AI-assisted diagnostic system compete with fine needle aspiration?

OBJECTIVES: Artificial intelligence (AI) systems can diagnose thyroid nodules with similar or better performance than radiologists. Little is known about how this performance compares with that achieved through fine needle aspiration (FNA). This study aims to compare the diagnostic yields of FNA cytopathology alone and combined with BRAFV600E mutation analysis and an AI diagnostic system. METHODS: The ultrasound images of 637 thyroid nodules were collected in three hospitals. The diagnostic efficacies of an AI diagnostic system, FNA-based cytopathology, and BRAFV600E mutation analysis were evaluated in terms of sensitivity, specificity, accuracy, and the κ coefficient with respect to the gold standard, defined by postsurgical pathology and consistent benign outcomes from two combined FNA and mutation analysis examinations performed with a half-year interval. RESULTS: The malignancy threshold for the AI system was selected according to the Youden index from a retrospective cohort of 346 nodules and then applied to a prospective cohort of 291 nodules. The combination of FNA cytopathology according to the Bethesda criteria and BRAFV600E mutation analysis showed no significant difference from the AI system in terms of accuracy for either cohort in our multicenter study. In addition, for 45 included indeterminate Bethesda category III and IV nodules, the accuracy, sensitivity, and specificity of the AI system were 84.44%, 95.45%, and 73.91%, respectively. CONCLUSIONS: The AI diagnostic system showed similar diagnostic performance to FNA cytopathology combined with BRAFV600E mutation analysis. Given its advantages in terms of operability, time efficiency, non-invasiveness, and the wide availability of ultrasonography, it provides a new alternative for thyroid nodule diagnosis. CLINICAL RELEVANCE STATEMENT: Thyroid ultrasonic artificial intelligence shows statistically equivalent performance for thyroid nodule diagnosis to FNA cytopathology combined with BRAFV600E mutation analysis. It can be widely applied in hospitals and clinics to assist radiologists in thyroid nodule screening and is expected to reduce the need for relatively invasive FNA biopsies. KEY POINTS: • In a retrospective cohort of 346 nodules, the evaluated artificial intelligence (AI) system did not significantly differ from fine needle aspiration (FNA) cytopathology alone and combined with gene mutation analysis in accuracy. • In a prospective multicenter cohort of 291 nodules, the accuracy of the AI diagnostic system was not significantly different from that of FNA cytopathology either alone or combined with gene mutation analysis. • For 45 indeterminate Bethesda category III and IV nodules, the AI system did not perform significantly differently from BRAFV600E mutation analysis.

Gene-allele system of shade tolerance in southern China soybean germplasm revealed by genome-wide association study using gene-allele sequence as markers.

KEY MESSAGE: Fifty-three shade tolerance genes with 281 alleles in the SCSGP were identified directly using gene-allele sequence as markers in RTM GWAS, from which optimized crosses, evolutionary motivators, and gene-allele networks were explored. Shade tolerance is a key for optimal cultivation of soybean inter/relay-cropped with corn. To explore the shade tolerance gene-allele system in the southern China soybean germplasm, we proposed using gene-allele sequence markers (GASMs) in a restricted two-stage multi-locus model genome-wide association study (GASM-RTM-GWAS). A representative sample with 394 accessions was tested for their shade tolerance index (STI), in Nanning, China. Through whole-genome re-sequencing, 47,586 GASMs were assembled. From GASM-RTM-GWAS, 53 main-effect STI genes with 281 alleles (2-13 alleles/gene) (totally 63 genes with 308 alleles, including 38 G × E genes with 191 alleles) were identified and then organized into a gene-allele matrix composed of eight submatrices corresponding to geo-seasonal subpopulations. The population featured mild STI changes (1.69 → 1.56-1.82) and mild gene-allele changes (92.5% alleles inherited, 0% alleles excluded, 7.5% alleles emerged) from the primitive (SAIII) to the derived seven subpopulations, but large transgressive recombination potentials and optimal crosses were predicted. The 63 STI genes were annotated into six biological categories (metabolic process, catalytic activity, response to stresses, transcription and translation, signal transduction and transport and unknown functions), interacted as gene networks. From the STI gene-allele system, 38 important alleles of 22 genes were nominated for further in-depth study. GASM-RTM-GWAS performed powerful and efficient in germplasm population genetic study comparing to other procedures through facilitating direct and thorough identification of its gene-allele system, from which genome-wide breeding by design could be achieved, and evolutionary motivators and gene-allele networks could be explored.

Genome-Wide Detection of Quantitative Trait Loci and Prediction of Candidate Genes for Seed Sugar Composition in Early Mature Soybean.

Seed sugar composition, mainly including fructose, glucose, sucrose, raffinose, and stachyose, is an important indicator of soybean [Glycine max (L.) Merr.] seed quality. However, research on soybean sugar composition is limited. To better understand the genetic architecture underlying the sugar composition in soybean seeds, we conducted a genome-wide association study (GWAS) using a population of 323 soybean germplasm accessions which were grown and evaluated under three different environments. A total of 31,245 single-nucleotide polymorphisms (SNPs) with minor allele frequencies (MAFs) ≥ 5% and missing data ≤ 10% were selected and used in the GWAS. The analysis identified 72 quantitative trait loci (QTLs) associated with individual sugars and 14 with total sugar. Ten candidate genes within the 100 Kb flanking regions of the lead SNPs across six chromosomes were significantly associated with sugar contents. According to GO and KEGG classification, eight genes were involved in the sugar metabolism in soybean and showed similar functions in Arabidopsis. The other two, located in known QTL regions associated with sugar composition, may play a role in sugar metabolism in soybean. This study advances our understanding of the genetic basis of soybean sugar composition and facilitates the identification of genes controlling this trait. The identified candidate genes will help improve seed sugar composition in soybean.

Genetic variation and marker-trait association affect the genomic selection prediction accuracy of soybean protein and oil content.

Introduction: Genomic selection (GS) is a potential breeding approach for soybean improvement. Methods: In this study, GS was performed on soybean protein and oil content using the Ridge Regression Best Linear Unbiased Predictor (RR-BLUP) based on 1,007 soybean accessions. The SoySNP50K SNP dataset of the accessions was obtained from the USDA-ARS, Beltsville, MD lab, and the protein and oil content of the accessions were obtained from GRIN. Results: Our results showed that the prediction accuracy of oil content was higher than that of protein content. When the training population size was 100, the prediction accuracies for protein content and oil content were 0.60 and 0.79, respectively. The prediction accuracy increased with the size of the training population. Training populations with similar phenotype or with close genetic relationships to the prediction population exhibited better prediction accuracy. A greatest prediction accuracy for both protein and oil content was observed when approximately 3,000 markers with -log10(P) greater than 1 were included. Discussion: This information will help improve GS efficiency and facilitate the application of GS.

Deep learning-based phenotyping for genome wide association studies of sudden death syndrome in soybean.

Using a reliable and accurate method to phenotype disease incidence and severity is essential to unravel the complex genetic architecture of disease resistance in plants, and to develop disease resistant cultivars. Genome-wide association studies (GWAS) involve phenotyping large numbers of accessions, and have been used for a myriad of traits. In field studies, genetic accessions are phenotyped across multiple environments and replications, which takes a significant amount of labor and resources. Deep Learning (DL) techniques can be effective for analyzing image-based tasks; thus DL methods are becoming more routine for phenotyping traits to save time and effort. This research aims to conduct GWAS on sudden death syndrome (SDS) of soybean [Glycine max L. (Merr.)] using disease severity from both visual field ratings and DL-based (using images) severity ratings collected from 473 accessions. Images were processed through a DL framework that identified soybean leaflets with SDS symptoms, and then quantified the disease severity on those leaflets into a few classes with mean Average Precision of 0.34 on unseen test data. Both visual field ratings and image-based ratings identified significant single nucleotide polymorphism (SNP) markers associated with disease resistance. These significant SNP markers are either in the proximity of previously reported candidate genes for SDS or near potentially novel candidate genes. Four previously reported SDS QTL were identified that contained a significant SNPs, from this study, from both a visual field rating and an image-based rating. The results of this study provide an exciting avenue of using DL to capture complex phenotypic traits from images to get comparable or more insightful results compared to subjective visual field phenotyping of traits for disease symptoms.

Transgressive Potential Prediction and Optimal Cross Design of Seed Protein Content in the Northeast China Soybean Population Based on Full Exploration of the QTL-Allele System.

Northeast China is a major soybean production region in China. A representative sample of the Northeast China soybean germplasm population (NECSGP) composed of 361 accessions was evaluated for their seed protein content (SPC) in Tieling, Northeast China. This SPC varied greatly, with a mean SPC of 40.77%, ranging from 36.60 to 46.07%, but it was lower than that of the Chinese soybean landrace population (43.10%, ranging from 37.51 to 50.46%). The SPC increased slightly from 40.32-40.97% in the old maturity groups (MG, MGIII + II + I) to 40.93-41.58% in the new MGs (MG0 + 00 + 000). The restricted two-stage multi-locus genome-wide association study (RTM-GWAS) with 15,501 SNP linkage-disequilibrium block (SNPLDB) markers identified 73 SPC quantitative trait loci (QTLs) with 273 alleles, explaining 71.70% of the phenotypic variation, wherein 28 QTLs were new ones. The evolutionary changes of QTL-allele structures from old MGs to new MGs were analyzed, and 97.79% of the alleles in new MGs were inherited from the old MGs and 2.21% were new. The small amount of new positive allele emergence and possible recombination between alleles might explain the slight SPC increase in the new MGs. The prediction of recombination potentials in the SPC of all the possible crosses indicated that the mean of SPC overall crosses was 43.29% (+2.52%) and the maximum was 50.00% (+9.23%) in the SPC, and the maximum transgressive potential was 3.93%, suggesting that SPC breeding potentials do exist in the NECSGP. A total of 120 candidate genes were annotated and functionally classified into 13 categories, indicating that SPC is a complex trait conferred by a gene network.

Proximal and remote sensing in plant phenomics: 20 years of progress, challenges, and perspectives.

Plant phenomics (PP) has been recognized as a bottleneck in studying the interactions of genomics and environment on plants, limiting the progress of smart breeding and precise cultivation. High-throughput plant phenotyping is challenging owing to the spatio-temporal dynamics of traits. Proximal and remote sensing (PRS) techniques are increasingly used for plant phenotyping because of their advantages in multi-dimensional data acquisition and analysis. Substantial progress of PRS applications in PP has been observed over the last two decades and is analyzed here from an interdisciplinary perspective based on 2972 publications. This progress covers most aspects of PRS application in PP, including patterns of global spatial distribution and temporal dynamics, specific PRS technologies, phenotypic research fields, working environments, species, and traits. Subsequently, we demonstrate how to link PRS to multi-omics studies, including how to achieve multi-dimensional PRS data acquisition and processing, how to systematically integrate all kinds of phenotypic information and derive phenotypic knowledge with biological significance, and how to link PP to multi-omics association analysis. Finally, we identify three future perspectives for PRS-based PP: (1) strengthening the spatial and temporal consistency of PRS data, (2) exploring novel phenotypic traits, and (3) facilitating multi-omics communication.

Effect of mouse nerve growth factor combined with the exposure of the external branch of superior laryngeal nerve on voice changes after thyroidectomy: a prospective non-randomized controlled trial.

Background: The superior laryngeal nerve (SLN) injury may also affect vocal fold function and voice quality. It is efficient yet simple approach to expose the external branch of the superior laryngeal nerve (EBSLN). Neurotrophic agent mouse nerve growth factor (mNGF) to treat patients after thyroid surgery, and found it had significant efficacy in improving the voice of patients. However, the potential effectiveness and safety of mNGF combined with EBSLN were unclear. Methods: In this study, 96 patients who suffered from hoarseness after thyroidectomy at Hangzhou First People's Hospital between January 2018 and October 2019 were screened and divided into the control group and the observation group by patients' choice. In the control group, the SLN was not exposed. In the observation group, the SLN was exposed. The mNGF treatment was administered for observation group once a day at 20 µg each time for 4 weeks. The data of acoustic voice indicators was analysis by univariate analyses. Patients in both groups were followed up for more than 6 months. The rate of SLN damage was compared between two groups. Results: The baseline clinical characteristics of the two groups showed no statistic difference. The results showed that the fundamental frequency was significantly lower 1 month after surgery than 3 days after surgery in both groups. The fundamental frequency perturbation, shimmer, maximum phonation time, highest fundamental frequency, and dysphonia severity index in 1 month after surgery were significantly higher than they were 3 days after surgery (all P<0.001). There was no significant difference in the postoperative harmonic-to-noise ratio between the 2 groups (P=0.426). Conclusions: MNGF combined with the exposure and protection of the EBSLN effectively may prevent voice damage after thyroid surgery.

Meta-GWAS for quantitative trait loci identification in soybean.

We report a meta-Genome Wide Association Study involving 73 published studies in soybean [Glycine max L. (Merr.)] covering 17,556 unique accessions, with improved statistical power for robust detection of loci associated with a broad range of traits. De novo GWAS and meta-analysis were conducted for composition traits including fatty acid and amino acid composition traits, disease resistance traits, and agronomic traits including seed yield, plant height, stem lodging, seed weight, seed mottling, seed quality, flowering timing, and pod shattering. To examine differences in detectability and test statistical power between single- and multi-environment GWAS, comparison of meta-GWAS results to those from the constituent experiments were performed. Using meta-GWAS analysis and the analysis of individual studies, we report 483 peaks at 393 unique loci. Using stringent criteria to detect significant marker-trait associations, 59 candidate genes were identified, including 17 agronomic traits loci, 19 for seed-related traits, and 33 for disease reaction traits. This study identified potentially valuable candidate genes that affect multiple traits. The success in narrowing down the genomic region for some loci through overlapping mapping results of multiple studies is a promising avenue for community-based studies and plant breeding applications.

Identification and Pleiotropic Effect Analysis of GSE5 on Rice Chalkiness and Grain Shape.

Chalkiness is one of several major restricting factors for the improvement of rice quality. Although many chalkiness-related quantitative trait loci have been mapped, only a small number of genes have been cloned to date. In this study, the candidate gene GSE5 of a major quantitative trait locus (QTL) for rice chalkiness, qDEC5, was identified by map-based cloning. Phenotyping and haplotype analysis of proActin:GSE5 transgenic line, gse5-cr mutant, and 69 rice varieties further confirmed that GSE5 had the pleiotropic effects and regulated both chalkiness and grain shape. Genetic analysis showed GSE5 was a dominant gene for grain length and a semi-dominant gene for grain width and chalkiness. The DNA interval closely linked to GSE5 was introgressed to Zhenshan 97B (ZB) based on molecular marker-assisted selection, and the improved ZB showed lower chalkiness and longer but smaller grains, which showed that GSE5 played an important role in breeding rice varieties with high yield and good quality. Transcriptomics, proteomics, and qRT-PCR analyses showed that thirty-nine genes associated with carbon and protein metabolism are regulated by GSE5 to affect the formation of chalkiness, including some newly discovered genes, such as OsCESA9, OsHSP70, OsTPS8, OsPFK04, OsSTA1, OsERdj3A, etc. The low-chalkiness lines showed higher amino sugar and nucleotide sugar metabolism at 10 days after pollination (DAP), lower carbohydrate metabolism at 15 DAP, and lower protein metabolism at 10 and 15 DAP. With heat shock at 34/30°C, rice chalkiness increased significantly; OsDjC10 and OsSUS3 were upregulated at 6 and 12 DAP, respectively, and OsGSTL2 was downregulated at 12 DAP. Our results identified the function and pleiotropic effects of qDEC5 dissected its genetic characteristics and the expression profiles of the genes affecting the chalkiness formation, and provided a theoretical basis and application value to harmoniously pursue high yield and good quality in rice production.

Deconstructing the genetic architecture of iron deficiency chlorosis in soybean using genome-wide approaches.

BACKGROUND: Iron (Fe) is an essential micronutrient for plant growth and development. Iron deficiency chlorosis (IDC), caused by calcareous soils or high soil pH, can limit iron availability, negatively affecting soybean (Glycine max) yield. This study leverages genome-wide association study (GWAS) and a genome-wide epistatic study (GWES) with previous gene expression studies to identify regions of the soybean genome important in iron deficiency tolerance. RESULTS: A GWAS and a GWES were performed using 460 diverse soybean PI lines from 27 countries, in field and hydroponic iron stress conditions, using more than 36,000 single nucleotide polymorphism (SNP) markers. Combining this approach with available RNA-sequencing data identified significant markers, genomic regions, and novel genes associated with or responding to iron deficiency. Sixty-nine genomic regions associated with IDC tolerance were identified across 19 chromosomes via the GWAS, including the major-effect quantitative trait locus (QTL) on chromosome Gm03. Cluster analysis of significant SNPs in this region deconstructed this historically prominent QTL into four distinct linkage blocks, enabling the identification of multiple candidate genes for iron chlorosis tolerance. The complementary GWES identified SNPs in this region interacting with nine other genomic regions, providing the first evidence of epistatic interactions impacting iron deficiency tolerance. CONCLUSIONS: This study demonstrates that integrating cutting edge genome wide association (GWA), genome wide epistasis (GWE), and gene expression studies is a powerful strategy to identify novel iron tolerance QTL and candidate loci from diverse germplasm. Crops, unlike model species, have undergone selection for thousands of years, constraining and/or enhancing stress responses. Leveraging genomics-enabled approaches to study these adaptations is essential for future crop improvement.

Genetic Control and Geo-Climate Adaptation of Pod Dehiscence Provide Novel Insights into Soybean Domestication.

Loss of pod dehiscence was a key step in soybean [Glycine max (L.) Merr.] domestication. Genome-wide association analysis for soybean shattering identified loci harboring Pdh1, NST1A and SHAT1-5 Pairwise epistatic interactions were observed, and the dehiscent Pdh1 overcomes resistance conferred by NST1A or SHAT1-5 locus. Further candidate gene association analysis identified a nonsense mutation in NST1A associated with pod dehiscence. Geographic analysis showed that in Northeast China (NEC), indehiscence at both Pdh1 and NST1A were required in cultivated soybean, while indehiscent Pdh1 alone is capable of preventing shattering in Huang-Huai-Hai (HHH) valleys. Indehiscent Pdh1 allele was only identified in wild soybean (Glycine soja L.) accession from HHH valleys suggesting that it may have originated in this region. No specific indehiscence was required in Southern China. Geo-climatic investigation revealed strong correlation between relative humidity and frequency of indehiscent Pdh1 across China. This study demonstrates that epistatic interaction between Pdh1 and NST1A fulfills a pivotal role in determining the level of resistance against pod dehiscence, and humidity shapes the distribution of indehiscent alleles. Our results give further evidence to the hypothesis that HHH valleys was at least one of the origin centers of cultivated soybean.

Identification and Genetic Characterization of Soybean Accessions Exhibiting Antibiosis and Antixenosis Resistance to Aphis glycines (Hemiptera: Aphididae).

Cultivation of aphid-resistant soybean varieties can reduce yield losses caused by soybean aphids. However, discovery of aphid biotypes that are virulent on resistant soybean greatly threatens sustained utilization of host plant resistance to control soybean aphids. The objective of this study was to identify and genetically characterize aphid resistant soybean accessions in a diverse collection of 308 plant introductions in maturity groups (MG) I and II. In large-scale screening experiments conducted in the greenhouse, we identified 12 soybean accessions (10 aphid-resistant and 2 moderately resistant), including nine previously not reported for resistance against soybean aphids. Three accessions (PI 578374, PI 612759C, and PI 603546A) and the Rag3 resistant check (PI 567543C) were susceptible when infested with a high initial aphid level but resistant when infested with a low initial aphid level, a phenomenon termed as density-dependent aphid resistance. Six accessions (PI 054854, PI 378663, PI 578374, PI 612759C, PI 540739, and PI 603546A) conferred antibiosis, five (PI 438031, PI 603337A, PI 612711B, PI 437950, and PI 096162) conferred both antibiosis and antixenosis, while one (PI 417513B) had neither when tested in no-choice and pairwise choice experiments. Molecular genotyping of the 12 accessions using single-nucleotide polymorphism (SNP) markers linked to known aphid resistance (Rag) genes revealed that PI 578374 and PI 540739 did not have any tested marker variants and could potentially carry unreported Rag genes. Genome-wide association analyses for MG I accessions identified genomic regions associated with aphid resistance on chromosomes 10 and 12 for each level of initial aphid colonization.

Genome-wide Scan for Seed Composition Provides Insights into Soybean Quality Improvement and the Impacts of Domestication and Breeding.

The complex genetic architecture of quality traits has hindered efforts to modify seed nutrients in soybean. Genome-wide association studies were conducted for seed composition, including protein, oil, fatty acids, and amino acids, using 313 diverse soybean germplasm accessions genotyped with a high-density SNP array. A total of 87 chromosomal regions were identified to be associated with seed composition, explaining 8%-89% of genetic variances. The candidate genes GmSAT1, AK-HSDH, SACPD-C, and FAD3A of known function, and putative MtN21 nodulin, FATB, and steroid-5-α-reductase involved in N2 fixation, amino acid biosynthesis, and fatty acid metabolism were found at the major-effect loci. Further analysis of additional germplasm accessions indicated that these major-effect loci had been subjected to domestication or modern breeding selection, and the allelic variants and distributions were relevant to geographic regions. We also revealed that amino acid concentrations related to seed weight and to total protein had a different genetic basis. This helps uncover the in-depth genetic mechanism of the intricate relationships among the seed compounds. Thus, our study not only provides valuable genes and markers for soybean nutrient improvement, both quantitatively and qualitatively, but also offers insights into the alteration of soybean quality during domestication and breeding.

Genetic Architecture of Charcoal Rot (Macrophomina phaseolina) Resistance in Soybean Revealed Using a Diverse Panel.

Charcoal rot (CR) disease caused by Macrophomina phaseolina is responsible for significant yield losses in soybean production. Among the methods available for controlling this disease, breeding for resistance is the most promising. Progress in breeding efforts has been slow due to the insufficient information available on the genetic mechanisms related to resistance. Genome-wide association studies (GWAS) enable unraveling the genetic architecture of resistance and identification of causal genes. The aims of this study were to identify new sources of resistance to CR in a collection of 459 diverse plant introductions from the USDA Soybean Germplasm Core Collection using field and greenhouse screenings, and to conduct GWAS to identify candidate genes and associated molecular markers. New sources for CR resistance were identified from both field and greenhouse screening from maturity groups I, II, and III. Five significant single nucleotide polymorphism (SNP) and putative candidate genes related to abiotic and biotic stress responses are reported from the field screening; while greenhouse screening revealed eight loci associated with eight candidate gene families, all associated with functions controlling plant defense response. No overlap of markers or genes was observed between field and greenhouse screenings suggesting a complex molecular mechanism underlying resistance to CR in soybean with varied response to different environments; but our findings provide useful information for advancing breeding for CR resistance as well as the genetic mechanism of resistance.

Main and epistatic loci studies in soybean for Sclerotinia sclerotiorum resistance reveal multiple modes of resistance in multi-environments.

Genome-wide association (GWAS) and epistatic (GWES) studies along with expression studies in soybean [Glycine max (L.) Merr.] were leveraged to dissect the genetics of Sclerotinia stem rot (SSR) [caused by Sclerotinia sclerotiorum (Lib.) de Bary], a significant fungal disease causing yield and quality losses. A large association panel of 466 diverse plant introduction accessions were phenotyped in multiple field and controlled environments to: (1) discover sources of resistance, (2) identify SNPs associated with resistance, and (3) determine putative candidate genes to elucidate the mode of resistance. We report 58 significant main effect loci and 24 significant epistatic interactions associated with SSR resistance, with candidate genes involved in a wide range of processes including cell wall structure, hormone signaling, and sugar allocation related to plant immunity, revealing the complex nature of SSR resistance. Putative candidate genes [for example, PHYTOALEXIN DEFFICIENT 4 (PAD4), ETHYLENE-INSENSITIVE 3-LIKE 1 (EIL3), and ETHYLENE RESPONSE FACTOR 1 (ERF1)] clustered into salicylic acid (SA), jasmonic acid (JA), and ethylene (ET) pathways suggest the involvement of a complex hormonal network typically activated by both necrotrophic (ET/JA) and biotrophic (SA) pathogens supporting that S. sclerotiorum is a hemibiotrophic plant pathogen.

Leveraging genomic prediction to scan germplasm collection for crop improvement.

The objective of this study was to explore the potential of genomic prediction (GP) for soybean resistance against Sclerotinia sclerotiorum (Lib.) de Bary, the causal agent of white mold (WM). A diverse panel of 465 soybean plant introduction accessions was phenotyped for WM resistance in replicated field and greenhouse tests. All plant accessions were previously genotyped using the SoySNP50K BeadChip. The predictive ability of six GP models were compared, and the impact of marker density and training population size on the predictive ability was investigated. Cross-prediction among environments was tested to determine the effectiveness of the prediction models. GP models had similar prediction accuracies for all experiments. Predictive ability did not improve significantly by using more than 5k SNPs, or by increasing the training population size (from 50% to 90% of the total of individuals). The GP model effectively predicted WM resistance across field and greenhouse experiments when each was used as either the training or validation population. The GP model was able to identify WM-resistant accessions in the USDA soybean germplasm collection that had previously been reported and were not included in the study panel. This study demonstrated the applicability of GP to identify useful genetic sources of WM resistance for soybean breeding. Further research will confirm the applicability of the proposed approach to other complex disease resistance traits and in other crops.

A real-time phenotyping framework using machine learning for plant stress severity rating in soybean.

BACKGROUND: Phenotyping is a critical component of plant research. Accurate and precise trait collection, when integrated with genetic tools, can greatly accelerate the rate of genetic gain in crop improvement. However, efficient and automatic phenotyping of traits across large populations is a challenge; which is further exacerbated by the necessity of sampling multiple environments and growing replicated trials. A promising approach is to leverage current advances in imaging technology, data analytics and machine learning to enable automated and fast phenotyping and subsequent decision support. In this context, the workflow for phenotyping (image capture â†’ data storage and curation â†’ trait extraction â†’ machine learning/classification â†’ models/apps for decision support) has to be carefully designed and efficiently executed to minimize resource usage and maximize utility. We illustrate such an end-to-end phenotyping workflow for the case of plant stress severity phenotyping in soybean, with a specific focus on the rapid and automatic assessment of iron deficiency chlorosis (IDC) severity on thousands of field plots. We showcase this analytics framework by extracting IDC features from a set of ~4500 unique canopies representing a diverse germplasm base that have different levels of IDC, and subsequently training a variety of classification models to predict plant stress severity. The best classifier is then deployed as a smartphone app for rapid and real time severity rating in the field. RESULTS: We investigated 10 different classification approaches, with the best classifier being a hierarchical classifier with a mean per-class accuracy of ~96%. We construct a phenotypically meaningful 'population canopy graph', connecting the automatically extracted canopy trait features with plant stress severity rating. We incorporated this image capture â†’ image processing â†’ classification workflow into a smartphone app that enables automated real-time evaluation of IDC scores using digital images of the canopy. CONCLUSION: We expect this high-throughput framework to help increase the rate of genetic gain by providing a robust extendable framework for other abiotic and biotic stresses. We further envision this workflow embedded onto a high throughput phenotyping ground vehicle and unmanned aerial system that will allow real-time, automated stress trait detection and quantification for plant research, breeding and stress scouting applications.

Computer vision and machine learning for robust phenotyping in genome-wide studies.

Traditional evaluation of crop biotic and abiotic stresses are time-consuming and labor-intensive limiting the ability to dissect the genetic basis of quantitative traits. A machine learning (ML)-enabled image-phenotyping pipeline for the genetic studies of abiotic stress iron deficiency chlorosis (IDC) of soybean is reported. IDC classification and severity for an association panel of 461 diverse plant-introduction accessions was evaluated using an end-to-end phenotyping workflow. The workflow consisted of a multi-stage procedure including: (1) optimized protocols for consistent image capture across plant canopies, (2) canopy identification and registration from cluttered backgrounds, (3) extraction of domain expert informed features from the processed images to accurately represent IDC expression, and (4) supervised ML-based classifiers that linked the automatically extracted features with expert-rating equivalent IDC scores. ML-generated phenotypic data were subsequently utilized for the genome-wide association study and genomic prediction. The results illustrate the reliability and advantage of ML-enabled image-phenotyping pipeline by identifying previously reported locus and a novel locus harboring a gene homolog involved in iron acquisition. This study demonstrates a promising path for integrating the phenotyping pipeline into genomic prediction, and provides a systematic framework enabling robust and quicker phenotyping through ground-based systems.

Deploying Fourier Coefficients to Unravel Soybean Canopy Diversity.

Soybean canopy outline is an important trait used to understand light interception ability, canopy closure rates, row spacing response, which in turn affects crop growth and yield, and directly impacts weed species germination and emergence. In this manuscript, we utilize a methodology that constructs geometric measures of the soybean canopy outline from digital images of canopies, allowing visualization of the genetic diversity as well as a rigorous quantification of shape parameters. Our choice of data analysis approach is partially dictated by the need to efficiently store and analyze large datasets, especially in the context of planned high-throughput phenotyping experiments to capture time evolution of canopy outline which will produce very large datasets. Using the Elliptical Fourier Transformation (EFT) and Fourier Descriptors (EFD), canopy outlines of 446 soybean plant introduction (PI) lines from 25 different countries exhibiting a wide variety of maturity, seed weight, and stem termination were investigated in a field experiment planted as a randomized complete block design with up to four replications. Canopy outlines were extracted from digital images, and subsequently chain coded, and expanded into a shape spectrum by obtaining the Fourier coefficients/descriptors. These coefficients successfully reconstruct the canopy outline, and were used to measure traditional morphometric traits. Highest phenotypic diversity was observed for roundness, while solidity showed the lowest diversity across all countries. Some PI lines had extraordinary shape diversity in solidity. For interpretation and visualization of the complexity in shape, Principal Component Analysis (PCA) was performed on the EFD. PI lines were grouped in terms of origins, maturity index, seed weight, and stem termination index. No significant pattern or similarity was observed among the groups; although interestingly when genetic marker data was used for the PCA, patterns similar to canopy outline traits was observed for origins, and maturity indexes. These results indicate the usefulness of EFT method for reconstruction and study of canopy morphometric traits, and provides opportunities for data reduction of large images for ease in future use.