DC2Net: An Asian Soybean Rust Detection Model Based on Hyperspectral Imaging and Deep Learning.

Asian soybean rust (ASR) is one of the major diseases that causes serious yield loss worldwide, even up to 80%. Early and accurate detection of ASR is critical to reduce economic losses. Hyperspectral imaging, combined with deep learning, has already been proved as a powerful tool to detect crop diseases. However, current deep learning models are limited to extract both spatial and spectral features in hyperspectral images due to the use of fixed geometric structure of the convolutional kernels, leading to the fact that the detection accuracy of current models remains further improvement. In this study, we proposed a deformable convolution and dilated convolution neural network (DC2Net) for the ASR detection. The deformable convolution module was used to extract the spatial features, while the dilated convolution module was applied to extract features from the spectral dimension. We also adopted the Shapley value and the channel attention methods to evaluate the importance of each wavelength during decision-making, thereby identifying the most contributing ones. The proposed DC2Net can realize early asymptomatic detection of ASR even when visual symptoms have not appeared. The results of the experiment showed that the detection performance of DC2Net dominated state-of-the-art methods, reaching an overall accuracy at 96.73%. Meanwhile, the experimental result suggested that the Shapley Additive exPlanations method was able to extract feature wavelengths correctly, thereby helping DC2Net achieve reasonable performance with less input data. The research result of this study could provide early warning of ASR outbreak in advance, even at the asymptomatic period.

An online monitoring device for measuring the concentration of four types of in-situ microorganisms by using the near-infrared band.

Using optical density at 600 nm (OD600) to measure the microbial concentration is a popular approach due to its advantages like quick response and non-destructive. However, the OD600 measurement might be affected by the metabolic pigment, and it would become invalid when the solution dilution is insufficient. To overcome these issues, we proposed to adopt a more robust wavelength at 890 nm to quantify the attenuation of transmission light. After selecting this light source, we designed the light path and the circuit of the online monitoring device. Meanwhile, the random forest algorithm was introduced for temperature compensation and improving the stability of the device. This device was verified by monitoring the microbial concentration of four strains (Yeast, Bacillus, Arthrobacter, and Escherichia coli). The experimental result suggested that the mean absolute percentage error reached 4.11 %, 4.28 %, 4.49 %, and 4.53 % respectively, which is helpful to improve the accuracy of microbial concentration measurement.

Identification of Increased Grain Length 1 (IGL1), a novel gene encoded by a major QTL for modulating grain length in rice.

KEY MESSAGE: A novel quantitative trait locus qIGL1, which performed a positive function in regulating grain length in rice, was cloned by the map-based cloning approach; further studies revealed that it corresponded to LOC_Os03g30530, and the IGL1 appeared to contribute to lengthening and widening of the cells on the surface of grain hulls. Grain length is a prominent determinant for grain weight and appearance quality of rice. In this study, we conducted quantitative trait locus mapping to determine a genomic interval responsible for a long-grain phenotype observed in a japonica cultivar HD385. This led to the identification of a novel QTL for grain length on chromosome 3, named qIGL1 (for Increased Grain Length 1); the HD385 (Handao 385)-derived allele showed enhancement effects on grain length, and such an allele as well as NIP (Nipponbare)-derived allele was designated qigl1 HD385 and qIGL1NIP, respectively. Genetic analysis revealed that the qigl1HD385 allele displayed semidominant effects on grain length. Fine mapping further narrowed down the qIGL1 to an ~ 70.8-kb region containing 9 open reading frames (ORFs). A comprehensive analysis indicated that LOC_Os03g30530, which corresponded to ORF6 and carried base substitutions and deletions in HD385 relative to NIP, thereby causing changes or losses of amino-acid residues, was the true gene for qIGL1. Comparison of grain traits between a pair of near-isogenic lines (NILs), termed NIL-igl1HD385 and NIL-IGL1NIP, discovered that introduction of the igl1HD385 into the NIP background significantly resulted in the elevations of grain length and 1000-grain weight. Closer inspection of grain surfaces revealed that the cell length and width in the longitudinal direction were significantly longer and greater, respectively, in NIL-igl1HD385 line compared with in NIL-IGL1NIP line. Hence, our studies identified a new semidominant natural allele contributing to the increase of grain length and further shed light on the regulatory mechanisms of grain length.

DiSignAtlas: an atlas of human and mouse disease signatures based on bulk and single-cell transcriptomics.

Molecular signatures are usually sets of biomolecules that can serve as diagnostic, prognostic, predictive, or therapeutic markers for a specific disease. Omics data derived from various high-throughput molecular biology technologies offer global, unbiased and appropriately comparable data, which can be used to identify such molecular signatures. To address the need for comprehensive disease signatures, DiSignAtlas (http://www.inbirg.com/disignatlas/) was developed to provide transcriptomics-based signatures for a wide range of diseases. A total of 181 434 transcriptome profiles were manually curated from studies involving 1836 nonredundant disease types in humans and mice. Then, 10 306 comparison datasets comprising both disease and control samples, including 328 single-cell RNA sequencing datasets, were established. Furthermore, a total of 3 775 317 differentially expressed genes in humans and 1 723 674 in mice were identified as disease signatures by analysing transcriptome profiles using commonly used pipelines. In addition to providing multiple methods for the retrieval of disease signatures, DiSignAtlas provides downstream functional enrichment analysis, cell type analysis and signature correlation analysis between diseases or species when available. Moreover, multiple analytical and comparison tools for disease signatures are available. DiSignAtlas is expected to become a valuable resource for both bioscientists and bioinformaticians engaged in translational research.

Integrative multi-omics and drug-response characterization of patient-derived prostate cancer primary cells.

Prostate cancer (PCa) is the second most prevalent malignancy in males across the world. A greater knowledge of the relationship between protein abundance and drug responses would benefit precision treatment for PCa. Herein, we establish 35 Chinese PCa primary cell models to capture specific characteristics among PCa patients, including gene mutations, mRNA/protein/surface protein distributions, and pharmaceutical responses. The multi-omics analyses identify Anterior Gradient 2 (AGR2) as a pre-operative prognostic biomarker in PCa. Through the drug library screening, we describe crizotinib as a selective compound for malignant PCa primary cells. We further perform the pharmacoproteome analysis and identify 14,372 significant protein-drug correlations. Surprisingly, the diminished AGR2 enhances the inhibition activity of crizotinib via ALK/c-MET-AKT axis activation which is validated by PC3 and xenograft model. Our integrated multi-omics approach yields a comprehensive understanding of PCa biomarkers and pharmacological responses, allowing for more precise diagnosis and therapies.

Research of intelligent reasoning system of Arabidopsis thaliana phenotype based on automated multi-task machine learning.

Traditional machine learning in plant phenotyping research requires the assistance of professional data scientists and domain experts to adjust the structure and hy-perparameters tuning of neural network models with much human intervention, making the model training and deployment ineffective. In this paper, the automated machine learning method is researched to construct a multi-task learning model for Arabidopsis thaliana genotype classification, leaf number, and leaf area regression tasks. The experimental results show that the genotype classification task's accuracy and recall achieved 98.78%, precision reached 98.83%, and classification F 1 value reached 98.79%, as well as the R 2 of leaf number regression task and leaf area regression task reached 0.9925 and 0.9997 respectively. The experimental results demonstrated that the multi-task automated machine learning model can combine the benefits of multi-task learning and automated machine learning, which achieved more bias information from related tasks and improved the overall classification and prediction effect. Additionally, the model can be created automatically and has a high degree of generalization for better phenotype reasoning. In addition, the trained model and system can be deployed on cloud platforms for convenient application.

OrganoidDB: a comprehensive organoid database for the multi-perspective exploration of bulk and single-cell transcriptomic profiles of organoids.

Organoids, three-dimensional in vitro tissue cultures derived from pluripotent (embryonic or induced) or adult stem cells, are promising models for the study of human processes and structures, disease onset and preclinical drug development. An increasing amount of omics data has been generated for organoid studies. Here, we introduce OrganoidDB (http://www.inbirg.com/organoid_db/), a comprehensive resource for the multi-perspective exploration of the transcriptomes of organoids. The current release of OrganoidDB includes curated bulk and single-cell transcriptome profiles of 16 218 organoid samples from both human and mouse. Other types of samples, such as primary tissue and cell line samples, are also integrated to enable comparisons with organoids. OrganoidDB enables queries of gene expression under different modes, e.g. across different organoid types, between different organoids from different sources or protocols, between organoids and other sample types, across different development stages, and via correlation analysis. Datasets and organoid samples can also be browsed for detailed information, including organoid information, differentially expressed genes, enriched pathways and single-cell clustering. OrganoidDB will facilitate a better understanding of organoids and help improve organoid culture protocols to yield organoids that are highly similar to living organs in terms of composition, architecture and function.

PertOrg 1.0: a comprehensive resource of multilevel alterations induced in model organisms by in vivo genetic perturbation.

Genetically modified organisms (GMOs) can be generated to model human genetic disease or plant disease resistance, and they have contributed to the exploration and understanding of gene function, physiology, disease onset and drug target discovery. Here, PertOrg (http://www.inbirg.com/pertorg/) was introduced to provide multilevel alterations in GMOs. Raw data of 58 707 transcriptome profiles and associated information, such as phenotypic alterations, were collected and curated from studies involving in vivo genetic perturbation (e.g. knockdown, knockout and overexpression) in eight model organisms, including mouse, rat and zebrafish. The transcriptome profiles from before and after perturbation were organized into 10 116 comparison datasets, including 122 single-cell RNA-seq datasets. The raw data were checked and analysed using widely accepted and standardized pipelines to identify differentially expressed genes (DEGs) in perturbed organisms. As a result, 8 644 148 DEGs were identified and deposited as signatures of gene perturbations. Downstream functional enrichment analysis, cell type analysis and phenotypic alterations were also provided when available. Multiple search methods and analytical tools were created and implemented. Furthermore, case studies were presented to demonstrate how users can utilize the database. PertOrg 1.0 will be a valuable resource aiding in the exploration of gene functions, biological processes and disease models.

Detecting Asymptomatic Infections of Rice Bacterial Leaf Blight Using Hyperspectral Imaging and 3-Dimensional Convolutional Neural Network With Spectral Dilated Convolution.

Rice is one of the most important food crops for human beings. Its total production ranks third in the grain crop output. Bacterial Leaf Blight (BLB), as one of the three major diseases of rice, occurs every year, posing a huge threat to rice production and safety. There is an asymptomatic period between the infection and the onset periods, and BLB will spread rapidly and widely under suitable conditions. Therefore, accurate detection of early asymptomatic BLB is very necessary. The purpose of this study was to test the feasibility of detecting early asymptomatic infection of the rice BLB disease based on hyperspectral imaging and Spectral Dilated Convolution 3-Dimensional Convolutional Neural Network (SDC-3DCNN). First, hyperspectral images were obtained from rice leaves infected with the BLB disease at the tillering stage. The spectrum was smoothed by the Savitzky-Golay (SG) method, and the wavelength between 450 and 950 nm was intercepted for analysis. Then Principal Component Analysis (PCA) and Random Forest (RF) were used to extract the feature information from the original spectra as inputs. The overall performance of the SDC-3DCNN model with different numbers of input features and different spectral dilated ratios was evaluated. Lastly, the saliency map visualization was used to explain the sensitivity of individual wavelengths. The results showed that the performance of the SDC-3DCNN model reached an accuracy of 95.4427% when the number of inputs is 50 characteristic wavelengths (extracted by RF) and the dilated ratio is set at 5. The saliency-sensitive wavelengths were identified in the range from 530 to 570 nm, which overlaps with the important wavelengths extracted by RF. According to our findings, combining hyperspectral imaging and deep learning can be a reliable approach for identifying early asymptomatic infection of the rice BLB disease, providing sufficient support for early warning and rice disease prevention.

Applying spectral fractal dimension index to predict the SPAD value of rice leaves under bacterial blight disease stress.

BACKGROUND: The chlorophyll content is a vital indicator for reflecting the photosynthesis ability of plants and it plays a significant role in monitoring the general health of plants. Since the chlorophyll content and the soil-plant analysis development (SPAD) value are positively correlated, it is feasible to predict the SPAD value by calculating the vegetation indices (VIs) through hyperspectral images, thereby evaluating the severity of plant diseases. However, current indices simply adopt few wavelengths of the hyperspectral information, which may decrease the prediction accuracy. Besides, few researches explored the applicability of VIs over rice under the bacterial blight disease stress. METHODS: In this study, the SPAD value was predicted by calculating the spectral fractal dimension index (SFDI) from a hyperspectral curve (420 to 950 nm). The correlation between the SPAD value and hyperspectral information was further analyzed for determining the sensitive bands that correspond to different disease levels. In addition, a SPAD prediction model was built upon the combination of selected indices and four machine learning methods. RESULTS: The results suggested that the SPAD value of rice leaves under different disease levels are sensitive to different wavelengths. Compared with current VIs, a stronger positive correlation was detected between the SPAD value and the SFDI, reaching an average correlation coefficient of 0.8263. For the prediction model, the one built with support vector regression and SFDI achieved the best performance, reaching R2, RMSE, and RE at 0.8752, 3.7715, and 7.8614%, respectively. CONCLUSIONS: This work provides an in-depth insight for accurately and robustly predicting the SPAD value of rice leaves under the bacterial blight disease stress, and the SFDI is of great significance for monitoring the chlorophyll content in large-scale fields non-destructively.

DDPD 1.0: a manually curated and standardized database of digital properties of approved drugs for drug-likeness evaluation and drug development.

Drug-likeness is a vital consideration when selecting compounds in the early stage of drug discovery. A series of drug-like properties are needed to predict the drug-likeness of a given compound and provide useful guidelines to increase the likelihood of converting lead compounds into drugs. Experimental physicochemical properties, pharmacokinetic/toxicokinetic properties and maximum dosages of approved small-molecule drugs from multiple text-based unstructured data resources have been manually assembled, curated, further digitized and processed into structured data, which are deposited in the Database of Digital Properties of approved Drugs (DDPD). DDPD 1.0 contains 30 212 drug property entries, including 2250 approved drugs and 32 properties, in a standardized value/unit format. Moreover, two analysis tools are provided to examine the drug-likeness features of given molecules based on the collected property data of approved drugs. Additionally, three case studies are presented to demonstrate how users can utilize the database. We believe that this database will be a valuable resource for the drug discovery and development field. Database URL:  http://www.inbirg.com/ddpd.

Roles of MEM1 in safeguarding Arabidopsis genome against DNA damage, inhibiting ATM/SOG1-mediated DNA damage response, and antagonizing global DNA hypermethylation.

Arabidopsis methylation elevated mutant 1 (mem1) mutants have elevated levels of global DNA methylation. In this study, such mutant alleles showed increased sensitivity to methyl methanesulfonate (MMS). In mem1 mutants, an assortment of genes engaged in DNA damage response (DDR), especially DNA-repair-associated genes, were largely upregulated without MMS treatment, suggestive of activation of the DDR pathway in them. Following MMS treatment, expression levels of multiple DNA-repair-associated genes in mem1 mutants were generally lower than in Col-0 plants, which accounted for the MMS-sensitive phenotype of the mem1 mutants. A group of DNA methylation pathway genes were upregulated in mem1 mutants under non-MMS-treated conditions, causing elevated global DNA methylation, especially in RNA-directed DNA methylation (RdDM)-targeted regions. Moreover, MEM1 seemed to help ATAXIA-TELANGIECTASIA MUTATED (ATM) and/or SUPPRESSOR OF GAMMA RESPONSE 1 (SOG1) to fully activate/suppress transcription of a subset of genes regulated simultaneously by MEM1 and ATM and/or SOG1, because expression of such genes decreased/increased consistently in mem1 and atm and/or sog1 mutants, but the decreases/increases in the mem1 mutants were not as dramatic as in the atm and/or sog1 mutants. Thus, our studies reveals roles of MEM1 in safeguarding genome, and interrelationships among DNA damage, activation of DDR, DNA methylation/demethylation, and DNA repair.

An Associated Representation Method for Defining Agricultural Cases in a Case-Based Reasoning System for Fast Case Retrieval.

As an artificial intelligence technique, case-based reasoning has considerable potential to build intelligent systems for smart agriculture, providing farmers with advice about farming operation management. A proper case representation method plays a crucial role in case-based reasoning systems. Some methods like textual, attribute-value pair, and ontological representations have been well explored by researchers. However, these methods may lead to inefficient case retrieval when a large volume of data is stored in the case base. Thus, an associated representation method is proposed in this paper for fast case retrieval. Each case is interconnected with several similar and dissimilar ones. Once a new case is reported, its features are compared with historical data by similarity measurements for identifying a relative similar past case. The similarity of associated cases is measured preferentially, instead of comparing all the cases in the case base. Experiments on case retrieval were performed between the associated case representation and traditional methods, following two criteria: the number of visited cases and retrieval accuracy. The result demonstrates that our proposal enables fast case retrieval with promising accuracy by visiting fewer past cases. In conclusion, the associated case representation method outperforms traditional methods in the aspect of retrieval efficiency.

A Triangular Similarity Measure for Case Retrieval in CBR and Its Application to an Agricultural Decision Support System.

Case-based reasoning has been a widely-used approach to assist humans in making decisions through four steps: retrieve, reuse, revise, and retain. Among these steps, case retrieval plays a significant role because the rest of processes cannot proceed without successfully identifying the most similar past case beforehand. Some popular methods such as angle-based and distance-based similarity measures have been well explored for case retrieval. However, these methods may match inaccurate cases under certain extreme circumstances. Thus, a triangular similarity measure is proposed to identify commonalities between cases, overcoming the drawbacks of angle-based and distance-based measures. For verifying the effectiveness and performance of the proposed measure, case-based reasoning was applied to an agricultural decision support system for pest management and 300 new cases were used for testing purposes. Once a new pest problem is reported, its attributes are compared with historical data by the proposed triangular similarity measure. Farmers can obtain quick decision support on managing pest problems by learning from the retrieved solution of the most similar past case. The experimental result shows that the proposed measure can retrieve the most similar case with an average accuracy of 91.99% and it outperforms the other measures in the aspects of accuracy and robustness.

A Rule-Based Reasoner for Underwater Robots Using OWL and SWRL.

Web Ontology Language (OWL) is designed to represent varied knowledge about things and the relationships of things. It is widely used to express complex models and address information heterogeneity of specific domains, such as underwater environments and robots. With the help of OWL, heterogeneous underwater robots are able to cooperate with each other by exchanging information with the same meaning and robot operators can organize the coordination easier. However, OWL has expressivity limitations on representing general rules, especially the statement "If … Then … Else …". Fortunately, the Semantic Web Rule Language (SWRL) has strong rule representation capabilities. In this paper, we propose a rule-based reasoner for inferring and providing query services based on OWL and SWRL. SWRL rules are directly inserted into the ontologies by several steps of model transformations instead of using a specific editor. In the verification experiments, the SWRL rules were successfully and efficiently inserted into the OWL-based ontologies, obtaining completely correct query results. This rule-based reasoner is a promising approach to increase the inference capability of ontology-based models and it achieves significant contributions when semantic queries are done.

A Mission Planning Approach for Precision Farming Systems Based on Multi-Objective Optimization.

As the demand for food grows continuously, intelligent agriculture has drawn much attention due to its capability of producing great quantities of food efficiently. The main purpose of intelligent agriculture is to plan agricultural missions properly and use limited resources reasonably with minor human intervention. This paper proposes a Precision Farming System (PFS) as a Multi-Agent System (MAS). Components of PFS are treated as agents with different functionalities. These agents could form several coalitions to complete the complex agricultural missions cooperatively. In PFS, mission planning should consider several criteria, like expected benefit, energy consumption or equipment loss. Hence, mission planning could be treated as a Multi-objective Optimization Problem (MOP). In order to solve MOP, an improved algorithm, MP-PSOGA, is proposed, taking advantages of the Genetic Algorithms and Particle Swarm Optimization. A simulation, called precise pesticide spraying mission, is performed to verify the feasibility of the proposed approach. Simulation results illustrate that the proposed approach works properly. This approach enables the PFS to plan missions and allocate scarce resources efficiently. The theoretical analysis and simulation is a good foundation for the future study. Once the proposed approach is applied to a real scenario, it is expected to bring significant economic improvement.