Edge-relational window-attentional graph neural network for gene expression prediction in spatial transcriptomics analysis

Spatial transcriptomics (ST), containing gene expression with fine-grained (i.e., different windows) spatial location within tissue samples, has become vital in developing innovative treatments. Traditional ST technology, however, rely on costly specialized commercial equipment. Addressing this, our article aims to creates a cost-effective, virtual ST approach using standard tissue images for gene expression prediction, eliminating the need for expensive equipment. Conventional approaches in this field often overlook the long-distance spatial dependencies between different sample windows or need prior gene expression data. To overcome these limitations, we propose the Edge-Relational Window-Attentional Network (ErwaNet), enhancing gene prediction by capturing both local interactions and global structural information from tissue images, without prior gene expression data. ErwaNet innovatively constructs heterogeneous graphs to model local window interactions and incorporates an attention mechanism for global information analysis. This dual framework not only provides a cost-effective solution for gene expression predictions but also obviates the necessity of prior knowledge gene expression information, a significant advantage in the field of cancer research where it enables a more efficient and accessible analytical paradigm. ErwaNet stands out as a prior-free and easy-to-implement Graph Convolution Network (GCN) method for predicting gene expression from tissue images. Evaluation of the two public breast cancer datasets shows that ErwaNet, without additional information, outperforms the state-of-the-art (SOTA) methods. Code is available at https://github.com/biyecc/ErwaNet.

Medical object detector jointly driven by knowledge and data

Most of the existing object detection algorithms are trained on medical datasets and then used for prediction. When the features of an object are not obvious in an image, these models are prone to mislocalize and misclassify it. In this paper, we propose a medical Object Detection algorithm jointly driven by Knowledge and Data (ODKD). It enables medical semantic knowledge provided by specialized physicians to be effective and helpful when traditional models have difficulty in correctly detecting objects relying on features alone. Our model consists of a base object detector together with a fusion module: the base object detector is trained based on medical datasets to obtain data-driven results; then we use a graph to represent external semantic knowledge and map the data-driven results to the nodes embedding of this graph structure. In the fusion module, a graph convolution network is used to fuse the data-driven results with the external semantic knowledge to output category adjustment coefficients. Finally, the adjustment coefficients are used to adjust the data-driven results to obtain results jointly driven by knowledge and data. Experiments show that professional medical semantic knowledge can effectively correct the erroneous results of the base detector, and the effect of our model outperforms Faster Rcnn, YOLOv5, YOLOv7, etc. on three medical datasets, Camus, Synapse, and AMOS.

GCN-Based Risk Prediction for Necrosis Slide of Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is one of the most common cancers in the world which ranks fourth in cancer deaths. Primary pathological necrosis is an effective prognostic indicator for hepatocellular carcinoma. We propose a GCN-based approach that mimics the pathologist's perspective for global assessment of necrosis tissue distribution to analyze patient survival. Specifically, we introduced a graph convolutional neural network to construct a spatial map with necrotic tissue and tumor tissue as graph nodes, aiming to mine the contextual information between necrotic tissue in pathological sections. We used 1381 slides from 303 patients from the First Affiliated Hospital of Zhejiang University School to train the model and used TCGA-LIHC for external validation. The C-index of our method outperforms the baseline by about 4.45%, which proves that the information about the spatial distribution of necrosis learned by GCN is meaningful for guiding patient prognosis.

Comprehensive Assessment and Early Prediction of Gross Motor Performance in Toddlers With Graph Convolutional Networks-Based Deep Learning: Development and Validation Study

BACKGROUND

Accurate and timely assessment of children's developmental status is crucial for early diagnosis and intervention. More accurate and automated developmental assessments are essential due to the lack of trained health care providers and imprecise parental reporting. In various areas of development, gross motor development in toddlers is known to be predictive of subsequent childhood developments.

OBJECTIVE

The purpose of this study was to develop a model to assess gross motor behavior and integrate the results to determine the overall gross motor status of toddlers. This study also aimed to identify behaviors that are important in the assessment of overall gross motor skills and detect critical moments and important body parts for the assessment of each behavior.

METHODS

We used behavioral videos of toddlers aged 18-35 months. To assess gross motor development, we selected 4 behaviors (climb up the stairs, go down the stairs, throw the ball, and stand on 1 foot) that have been validated with the Korean Developmental Screening Test for Infants and Children. In the child behavior videos, we estimated each child's position as a bounding box and extracted human keypoints within the box. In the first stage, the videos with the extracted human keypoints of each behavior were evaluated separately using a graph convolutional networks (GCN)-based algorithm. The probability values obtained for each label in the first-stage model were used as input for the second-stage model, the extreme gradient boosting (XGBoost) algorithm, to predict the overall gross motor status. For interpretability, we used gradient-weighted class activation mapping (Grad-CAM) to identify important moments and relevant body parts during the movements. The Shapley additive explanations method was used for the assessment of variable importance, to determine the movements that contributed the most to the overall developmental assessment.

RESULTS

Behavioral videos of 4 gross motor skills were collected from 147 children, resulting in a total of 2395 videos. The stage-1 GCN model to evaluate each behavior had an area under the receiver operating characteristic curve (AUROC) of 0.79 to 0.90. Keypoint-mapping Grad-CAM visualization identified important moments in each behavior and differences in important body parts. The stage-2 XGBoost model to assess the overall gross motor status had an AUROC of 0.90. Among the 4 behaviors, "go down the stairs" contributed the most to the overall developmental assessment.

CONCLUSIONS

Using movement videos of toddlers aged 18-35 months, we developed objective and automated models to evaluate each behavior and assess each child's overall gross motor performance. We identified the important behaviors for assessing gross motor performance and developed methods to recognize important moments and body parts while evaluating gross motor performance.

In silico prediction of ocular toxicity of compounds using explainable machine learning and deep learning approaches

The accurate identification of chemicals with ocular toxicity is of paramount importance in health hazard assessment. In contemporary chemical toxicology, there is a growing emphasis on refining, reducing, and replacing animal testing in safety evaluations. Therefore, the development of robust computational tools is crucial for regulatory applications. The performance of predictive models is heavily reliant on the quality and quantity of data. In this investigation, we amalgamated the most extensive dataset (4901 compounds) sourced from governmental GHS-compliant databases and literature to develop binary classification models of chemical ocular toxicity. We employed 12 molecular representations in conjunction with six machine learning algorithms and two deep learning algorithms to create a series of binary classification models. The findings indicated that the deep learning method GCN outperformed the machine learning models in cross-validation, achieving an impressive AUC of 0.915. However, the top-performing machine learning model (RF-Descriptor) demonstrated excellent performance with an AUC of 0.869 on the test set and was therefore selected as the best model. To enhance model interpretability, we conducted the SHAP method and attention weights analysis. The two approaches offered visual depictions of the relevance of key descriptors and substructures in predicting ocular toxicity of chemicals. Thus, we successfully struck a delicate balance between data quality and model interpretability, rendering our model valuable for predicting and comprehending potential ocular-toxic compounds in the early stages of drug discovery.

A Novel Approach for Automatic Detection of Driver Fatigue Using EEG Signals Based on Graph Convolutional Networks

Nowadays, the automatic detection of driver fatigue has become one of the important measures to prevent traffic accidents. For this purpose, a lot of research has been conducted in this field in recent years. However, the diagnosis of fatigue in recent research is binary and has no operational capability. This research presents a multi-class driver fatigue detection system based on electroencephalography (EEG) signals using deep learning networks. In the proposed system, a standard driving simulator has been designed, and a database has been collected based on the recording of EEG signals from 20 participants in five different classes of fatigue. In addition to self-report questionnaires, changes in physiological patterns are used to confirm the various stages of weariness in the suggested model. To pre-process and process the signal, a combination of generative adversarial networks (GAN) and graph convolutional networks (GCN) has been used. The proposed deep model includes five convolutional graph layers, one dense layer, and one fully connected layer. The accuracy obtained for the proposed model is 99%, 97%, 96%, and 91%, respectively, for the four different considered practical cases. The proposed model is compared to one developed through recent methods and research and has a promising performance.

In silico prediction of hERG blockers using machine learning and deep learning approaches

The human ether-à-go-go-related gene (hERG) is associated with drug cardiotoxicity. If the hERG channel is blocked, it will lead to prolonged QT interval and cause sudden death in severe cases. Therefore, it is important to evaluate the hERG-blocking property of compounds in early drug discovery. In this study, a dataset containing 4556 compounds with IC50 values determined by patch clamp techniques on mammalian lineage cells was collected, and hERG blockers and non-blockers were distinguished according to three single thresholds and two binary thresholds. Four machine learning (ML) algorithms combining four molecular fingerprints and molecular descriptors as well as graph convolutional neural networks (GCNs) were used to construct a series of binary classification models. The results showed that the best models varied for different thresholds. The ML models implemented by support vector machine and random forest performed well based on Morgan fingerprints and molecular descriptors, with AUCs ranging from 0.884 to 0.950. GCN showed superior prediction performance with AUCs above 0.952, which might be related to its direct extraction of molecular features from the original input. Meanwhile, the classification of binary threshold was better than that of single threshold, which could provide us with a more accurate prediction of hERG blockers. At last, the applicability domain for the model was defined, and seven structural alerts that might generate hERG blockage were identified by information gain and substructure frequency analysis. Our work would be beneficial for identifying hERG blockers in chemicals.

Dual-GCN-based deep clustering with triplet contrast for ScRNA-seq data analysis

Single-cell RNA sequencing (ScRNA-seq) technology reveals gene expression information at the cellular level. The critical tasks in ScRNA-seq data analysis are clustering and dimensionality reduction. Recent deep clustering algorithms are used to optimize the two tasks jointly, and their variations, graph-based deep clustering algorithms, are used to capture and preserve topological information in the process. However, the existing graph-based deep clustering algorithms ignore the distribution information of nodes when constructing cell graphs which leads to incomplete information in the embedding representation; and graph convolutional networks (GCN), which are most commonly used, often suffer from over-smoothing that leads to high sample similarity in the embedding representation and then poor clustering performance. Here, the dual-GCN-based deep clustering with Triplet contrast (scDGDC) is proposed for dimensionality reduction and clustering of scRNA-seq data. Two critical components are dual-GCN-based encoder for capturing more comprehensive topological information and triplet contrast for reducing GCN over-smoothing. The two components improve the dimensionality reduction and clustering performance of scDGDC in terms of information acquisition and model optimization, respectively. The experiments on eight real ScRNA-seq datasets showed that scDGDC achieves excellent performance for both clustering and dimensionality reduction tasks and is high robustness to parameters.

PLSNet: Position-aware GCN-based autism spectrum disorder diagnosis via FC learning and ROIs sifting

Brain function connectivity, derived from functional magnetic resonance imaging (fMRI), has enjoyed high popularity in the studies of Autism Spectrum Disorder (ASD) diagnosis. Albeit rapid progress has been made, most studies still suffer from several knotty issues: (1) the hardship of modeling the sophisticated brain neuronal connectivity; (2) the mismatch of identically graph node setup to the variations of different brain regions; (3) the dimensionality explosion resulted from excessive voxels in each fMRI sample; (4) the poor interpretability giving rise to unpersuasive diagnosis. To ameliorate these issues, we propose a position-aware graph-convolution-network-based model, namely PLSNet, with superior accuracy and compelling built-in interpretability for ASD diagnosis. Specifically, a time-series encoder is designed for context-rich feature extraction, followed by a function connectivity generator to model the correlation with long range dependencies. In addition, to discriminate the brain nodes with different locations, the position embedding technique is adopted, giving a unique identity to each graph region. We then embed a rarefying method to sift the salient nodes during message diffusion, which would also benefit the reduction of the dimensionality complexity. Extensive experiments conducted on Autism Brain Imaging Data Exchange demonstrate that our PLSNet achieves state-of-the-art performance. Notably, on CC200 atlas, PLSNet reaches an accuracy of 76.4% and a specificity of 78.6%, overwhelming the previous state-of-the-art with 2.5% and 6.5% under five-fold cross-validation policy. Moreover, the most salient brain regions predicted by PLSNet are closely consistent with the theoretical knowledge in the medical domain, providing potential biomarkers for ASD clinical diagnosis. Our code is available at https://github.com/CodeGoat24/PLSNet.

Chinese text classification by combining Chinese-BERTology-wwm and GCN

Text classification is an important and classic application in natural language processing (NLP). Recent studies have shown that graph neural networks (GNNs) are effective in tasks with rich structural relationships and serve as effective transductive learning approaches. Text representation learning methods based on large-scale pretraining can learn implicit but rich semantic information from text. However, few studies have comprehensively utilized the contextual semantic and structural information for Chinese text classification. Moreover, the existing GNN methods for text classification did not consider the applicability of their graph construction methods to long or short texts. In this work, we propose Chinese-BERTology-wwm-GCN, a framework that combines Chinese bidirectional encoder representations from transformers (BERT) series models with whole word masking (Chinese-BERTology-wwm) and the graph convolutional network (GCN) for Chinese text classification. When building text graph, we use documents and words as nodes to construct a heterogeneous graph for the entire corpus. Specifically, we use the term frequency-inverse document frequency (TF-IDF) to construct the word-document edge weights. For long text corpora, we propose an improved pointwise mutual information (PMI*) measure for words according to their word co-occurrence distances to represent the weights of word-word edges. For short text corpora, the co-occurrence information between words is often limited. Therefore, we utilize cosine similarity to represent the word-word edge weights. During the training stage, we effectively combine the cross-entropy and hinge losses and use them to jointly train Chinese-BERTology-wwm and GCN. Experiments show that our proposed framework significantly outperforms the baselines on three Chinese benchmark datasets and achieves good performance even with few labeled training sets.

Irregular Scene Text Detection Based on a Graph Convolutional Network

Detecting irregular or arbitrary shape text in natural scene images is a challenging task that has recently attracted considerable attention from research communities. However, limited by the CNN receptive field, these methods cannot directly capture relations between distant component regions by local convolutional operators. In this paper, we propose a novel method that can effectively and robustly detect irregular text in natural scene images. First, we employ a fully convolutional network architecture based on VGG16_BN to generate text components via the estimated character center points, which can ensure a high text component detection recall rate and fewer noncharacter text components. Second, text line grouping is treated as a problem of inferring the adjacency relations of text components with a graph convolution network (GCN). Finally, to evaluate our algorithm, we compare it with other existing algorithms by performing experiments on three public datasets: ICDAR2013, CTW-1500 and MSRA-TD500. The results show that the proposed method handles irregular scene text well and that it achieves promising results on these three public datasets.

Ordering taxa in image convolution networks improves microbiome-based machine learning accuracy

The human gut microbiome is associated with a large number of disease etiologies. As such, it is a natural candidate for machine-learning-based biomarker development for multiple diseases and conditions. The microbiome is often analyzed using 16S rRNA gene sequencing or shotgun metagenomics. However, several properties of microbial sequence-based studies hinder machine learning (ML), including non-uniform representation, a small number of samples compared with the dimension of each sample, and sparsity of the data, with the majority of taxa present in a small subset of samples. We show here using a graph representation that the cladogram structure is as informative as the taxa frequency. We then suggest a novel method to combine information from different taxa and improve data representation for ML using microbial taxonomy. iMic (image microbiome) translates the microbiome to images through an iterative ordering scheme, and applies convolutional neural networks to the resulting image. We show that iMic has a higher precision in static microbiome gene sequence-based ML than state-of-the-art methods. iMic also facilitates the interpretation of the classifiers through an explainable artificial intelligence (AI) algorithm to iMic to detect taxa relevant to each condition. iMic is then extended to dynamic microbiome samples by translating them to movies.

A survey of field programmable gate array (FPGA)-based graph convolutional neural network accelerators: challenges and opportunities

Graph convolutional networks (GCNs) based on convolutional operations have been developed recently to extract high-level representations from graph data. They have shown advantages in many critical applications, such as recommendation system, natural language processing, and prediction of chemical reactivity. The problem for the GCN is that its target applications generally pose stringent constraints on latency and energy efficiency. Several studies have demonstrated that field programmable gate array (FPGA)-based GCNs accelerators, which balance high performance and low power consumption, can continue to achieve orders-of-magnitude improvements in the inference of GCNs models. However, there still are many challenges in customizing FPGA-based accelerators for GCNs. It is necessary to sort out the current solutions to these challenges for further research. For this purpose, we first summarize the four challenges in FPGA-based GCNs accelerators. Then we introduce the process of the typical GNN algorithm and several examples of representative GCNs. Next, we review the FPGA-based GCNs accelerators in recent years and introduce their design details according to different challenges. Moreover, we compare the key metrics of these accelerators, including resource utilization, performance, and power consumption. Finally, we anticipate the future challenges and directions for FPGA-based GCNs accelerators: algorithm and hardware co-design, efficient task scheduling, higher generality, and faster development.

Cocrystal Prediction of Bexarotene by Graph Convolution Network and Bioavailability Improvement

Bexarotene (BEX) was approved by the FDA in 1999 for the treatment of cutaneous T-cell lymphoma (CTCL). The poor aqueous solubility causes the low bioavailability of the drug and thereby limits the clinical application. In this study, we developed a GCN-based deep learning model (CocrystalGCN) for in-silico screening of the cocrystals of BEX. The results show that our model obtained high performance relative to baseline models. The top 30 of 109 coformer candidates were scored by CocrystalGCN and then validated experimentally. Finally, cocrystals of BEX-pyrazine, BEX-2,5-dimethylpyrazine, BEX-methyl isonicotinate, and BEX-ethyl isonicotinate were successfully obtained. The crystal structures were determined by single-crystal X-ray diffraction. Powder X-ray diffraction, differential scanning calorimetry, and thermogravimetric analysis were utilized to characterize these multi-component forms. All cocrystals present superior solubility and dissolution over the parent drug. The pharmacokinetic studies show that the plasma exposures (AUC_0−8h) of BEX-pyrazine and BEX-2,5-dimethylpyrazine are 1.7 and 1.8 times that of the commercially available BEX powder, respectively. This work sets a good example for integrating virtual prediction and experimental screening to discover the new cocrystals of water-insoluble drugs.

Applying Multiple Functional Connectivity Features in GCN for EEG-Based Human Identification

EEG-based human identification has gained a wide range of attention due to the further increase in demand for security. How to improve the accuracy of the human identification system is an issue worthy of attention. Using more features in the human identification system is a potential solution. However, too many features may cause overfitting, resulting in the decline of system accuracy. In this work, the graph convolutional neural network (GCN) was adopted for classification. Multiple features were combined and utilized as the structure matrix of the GCN. Because of the constant signal matrix, the training parameters would not increase as the structure matrix grows. We evaluated the classification accuracy on a classic public dataset. The results showed that utilizing multiple features of functional connectivity (FC) can improve the accuracy of the identity authentication system, the best results of which are at 98.56%. In addition, our methods showed less sensitivity to channel reduction. The method proposed in this paper combines different FCs and reaches high classification accuracy for unpreprocessed data, which inspires reducing the system cost in the actual human identification system.

Constant community identification in million-scale networks

The inherently stochastic nature of community detection in real-world complex networks poses an important challenge in assessing the accuracy of the results. In order to eliminate the algorithmic and implementation artifacts, it is necessary to identify the groups of vertices that are always clustered together, independent of the community detection algorithm used. Such groups of vertices are called constant communities. Current approaches for finding constant communities are very expensive and do not scale to large networks. In this paper, we use binary edge classification to find constant communities. The key idea is to classify edges based on whether they form a constant community or not. We present two methods for edge classification. The first is a GCN-based semi-supervised approach that we term Line-GCN. The second is an unsupervised approach based on image thresholding methods. Neither of these methods requires explicit detection of communities and can thus scale to very large networks of the order of millions of vertices. Both of our semi-supervised and unsupervised results on real-world graphs demonstrate that the constant communities obtained by our method have higher F1-scores and comparable or higher NMI scores than other state-of-the-art baseline methods for constant community detection. While the training step of Line-GCN can be expensive, the unsupervised algorithm is 10 times faster than the baseline methods. For larger networks, the baseline methods cannot complete, whereas all of our algorithms can find constant communities in a reasonable amount of time. Finally, we also demonstrate that our methods are robust under noisy conditions. We use three different, well-studied noise models to add noise to the networks and show that our results are mostly stable.

Degradation of mixed cationic dye pollutant by metal free melem derivatives and graphitic carbon nitride

Graphitic carbon nitride (GCN), a polymeric metal free catalyst is widely used to degrade the toxic organic dye from the aqueous pollution. However, its catalytic efficiency and effective simultaneous reduction of mixed dye is still a challenge. Here, we have tuned the physiochemical properties of the GCN and melem derivatives by facilely tuning the degree of polycondensation and examined their catalytic activity towards the removal of cationic dye individually and together in solution. Catalysts were synthesized by thermal treatment of low-cost melamine and characterized by XRD, FTIR, RAMAN, FE-SEM, EDX, UV-DRS, and FL spectroscopy to confirm materials' structure, phase, morphology and optical properties. A suitable phase of the catalyst (M-450) exhibited superior removal capacity with a high-rate constant compared to others. The results demonstrate that M-450 has a maximum loading efficacy of 2.13 and 1.12 mg g^-1 for methylene blue (MB) and Rhodamine B (RhB) dyes respectively in a single dye system. Attractively, when MB and RhB co-exist in the solution, the efficacy increased by 14% (2.44 mg g^-1) and 27% (1.43 mg g^-1) for MB and RhB respectively. The adsorption kinetics, stability, effect of pH and reusability of M-450 catalyst was testified. Further, radical scavenger experiments and terephthalic acid tests were carried out to explain the reaction mechanism involved in the degradation of textile dyes. Moreover, electron paramagnetic resonance (EPR) analysis validated the availability of hydroxyl radicals in the photocatalytic reaction. Excellent stability and reusability were attained even after five successive cycles, demonstrating a suitable photocatalyst for the efficient degradation of mixed dye.

Decoding Color Visual Working Memory from EEG Signals Using Graph Convolutional Neural Networks

Color has an important role in object recognition and visual working memory (VWM). Decoding color VWM in the human brain is helpful to understand the mechanism of visual cognitive process and evaluate memory ability. Recently, several studies showed that color could be decoded from scalp electroencephalogram (EEG) signals during the encoding stage of VWM, which process visible information with strong neural coding. Whether color could be decoded from other VWM processing stages, especially the maintaining stage which processes invisible information, is still unknown. Here, we constructed an EEG color graph convolutional network model (ECo-GCN) to decode colors during different VWM stages. Based on graph convolutional networks, ECo-GCN considers the graph structure of EEG signals and may be more efficient in color decoding. We found that (1) decoding accuracies for colors during the encoding, early, and late maintaining stages were 81.58%, 79.36%, and 77.06%, respectively, exceeding those during the pre-stimuli stage (67.34%), and (2) the decoding accuracy during maintaining stage could predict participants' memory performance. The results suggest that EEG signals during the maintaining stage may be more sensitive than behavioral measurement to predict the VWM performance of human, and ECo-GCN provides an effective approach to explore human cognitive function.

HPCGCN: A Predictive Framework on High Performance Computing Cluster Log Data Using Graph Convolutional Networks

This paper presents a novel use case of Graph Convolutional Network (GCN) learning representations for predictive data mining, specifically from user/task data in the domain of high-performance computing (HPC). It outlines an approach based on a coalesced data set: logs from the Slurm workload manager, joined with user experience survey data from computational cluster users. We introduce a new method of constructing a heterogeneous unweighted HPC graph consisting of multiple typed nodes after revealing the manifold relations between the nodes. The GCN structure used here supports two tasks: i) determining whether a job will complete or fail and ii) predicting memory and CPU requirements by training the GCN semi-supervised classification model and regression models on the generated graph. The graph is partitioned into partitions using graph clustering. We conducted classification and regression experiments using the proposed framework on our HPC log dataset and evaluated predictions by our trained models against baselines using test_score, F1-score, precision, recall for classification, and R1 score for regression, showing that our framework achieves significant improvements.

Interaction-Temporal GCN: A Hybrid Deep Framework For Covid-19 Pandemic Analysis

The Covid-19 pandemic is still spreading around the world and seriously imperils humankind's health. This swift spread has caused the public to panic and look to scientists for answers. Fortunately, these scientists already have a wealth of data—the Covid-19 reports that each country releases, reports with valuable spatial-temporal properties. These data point toward some key actions that humans can take in their fight against Covid-19. Technically, the Covid-19 records can be described as sequences, which represent spatial-temporal linkages among the data elements with graph structure. Therefore, we propose a novel framework, the Interaction-Temporal Graph Convolution Network (IT-GCN), to analyze pandemic data. Specifically, IT-GCN introduces ARIMA into GCN to model the data which originate on nodes in a graph, indicating the severity of the pandemic in different cities. Instead of regular spatial topology, we construct the graph nodes with the vectors via ARIMA parameterization to find out the interaction topology underlying in the pandemic data. Experimental results show that IT-GCN is able to capture the comprehensive interaction-temporal topology and achieve well-performed short-term prediction of the Covid-19 daily infected cases in the United States. Our framework outperforms state-of-art baselines in terms of MAE, RMSE and MAPE. We believe that IT-GCN is a valid and reasonable method to forecast the Covid-19 daily infected cases and other related time-series. Moreover, the prediction can assist in improving containment policies.

Exploiting dynamic spatio-temporal graph convolutional neural networks for citywide traffic flows prediction

The prediction of crowd flows is an important urban computing issue whose purpose is to predict the future number of incoming and outgoing people in regions. Measuring the complicated spatial-temporal dependencies with external factors, such as weather conditions and surrounding point-of-interest (POI) distribution is the most difficult aspect of predicting crowd flows movement. To overcome the above issue, this paper advises a unified dynamic deep spatio-temporal neural network model based on convolutional neural networks and long short-term memory, termed as (DHSTNet) to simultaneously predict crowd flows in every region of a city. The DHSTNet model is made up of four separate components: a recent, daily, weekly, and an external branch component. Our proposed approach simultaneously assigns various weights to different branches and integrates the four properties' outputs to generate final predictions. Moreover, to verify the generalization and scalability of the proposed model, we apply a Graph Convolutional Network (GCN) based on Long Short Term Memory (LSTM) with the previously published model, termed as GCN-DHSTNet; to capture the spatial patterns and short-term temporal features; and to illustrate its exceptional accomplishment in predicting the traffic crowd flows. The GCN-DHSTNet model not only depicts the spatio-temporal dependencies but also reveals the influence of different time granularity, which are recent, daily, weekly periodicity and external properties, respectively. Finally, a fully connected neural network is utilized to fuse the spatio-temporal features and external properties together. Using two different real-world traffic datasets, our evaluation suggests that the proposed GCN-DHSTNet method is approximately 7.9%-27.2% and 11.2%-11.9% better than the AAtt-DHSTNet method in terms of RMSE and MAPE metrics, respectively. Furthermore, AAtt-DHSTNet outperforms other state-of-the-art methods.

Motor Intention Decoding from the Upper Limb by Graph Convolutional Network Based on Functional Connectivity

Decoding brain intention from noninvasively measured neural signals has recently been a hot topic in brain-computer interface (BCI). The motor commands about the movements of fine parts can increase the degrees of freedom under control and be applied to external equipment without stimulus. In the decoding process, the classifier is one of the key factors, and the graph information of the EEG was ignored by most researchers. In this paper, a graph convolutional network (GCN) based on functional connectivity was proposed to decode the motor intention of four fine parts movements (shoulder, elbow, wrist, hand). First, event-related desynchronization was analyzed to reveal the differences between the four classes. Second, functional connectivity was constructed by using synchronization likelihood (SL), phase-locking value (PLV), H index (H), mutual information (MI), and weighted phase-lag index (WPLI) to acquire the electrode pairs with a difference. Subsequently, a GCN and convolutional neural networks (CNN) were performed based on functional topological structures and time points, respectively. The results demonstrated that the proposed method achieved a decoding accuracy of up to 92.81% in the four-class task. Besides, the combination of GCN and functional connectivity can promote the development of BCI.

Multi-label classification of fundus images based on graph convolutional network

Background

Diabetic Retinopathy (DR) is the most common and serious microvascular complication in the diabetic population. Using computer-aided diagnosis from the fundus images has become a method of detecting retinal diseases, but the detection of multiple lesions is still a difficult point in current research.

Methods

This study proposed a multi-label classification method based on the graph convolutional network (GCN), so as to detect 8 types of fundus lesions in color fundus images. We collected 7459 fundus images (1887 left eyes, 1966 right eyes) from 2282 patients (1283 women, 999 men), and labeled 8 types of lesions, laser scars, drusen, cup disc ratio (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$C/D>0.6$$\end{document}C/D>0.6), hemorrhages, retinal arteriosclerosis, microaneurysms, hard exudates and soft exudates. We constructed a specialized corpus of the related fundus lesions. A multi-label classification algorithm for fundus images was proposed based on the corpus, and the collected data were trained.

Results

The average overall F1 Score (OF1) and the average per-class F1 Score (CF1) of the model were 0.808 and 0.792 respectively. The area under the ROC curve (AUC) of our proposed model reached 0.986, 0.954, 0.946, 0.957, 0.952, 0.889, 0.937 and 0.926 for detecting laser scars, drusen, cup disc ratio, hemorrhages, retinal arteriosclerosis, microaneurysms, hard exudates and soft exudates, respectively.

Conclusions

Our results demonstrated that our proposed model can detect a variety of lesions in the color images of the fundus, which lays a foundation for assisting doctors in diagnosis and makes it possible to carry out rapid and efficient large-scale screening of fundus lesions.

Relation Classification for Bleeding Events From Electronic Health Records Using Deep Learning Systems: An Empirical Study

BACKGROUND

Accurate detection of bleeding events from electronic health records (EHRs) is crucial for identifying and characterizing different common and serious medical problems. To extract such information from EHRs, it is essential to identify the relations between bleeding events and related clinical entities (eg, bleeding anatomic sites and lab tests). With the advent of natural language processing (NLP) and deep learning (DL)-based techniques, many studies have focused on their applicability for various clinical applications. However, no prior work has utilized DL to extract relations between bleeding events and relevant entities.

OBJECTIVE

In this study, we aimed to evaluate multiple DL systems on a novel EHR data set for bleeding event-related relation classification.

METHODS

We first expert annotated a new data set of 1046 deidentified EHR notes for bleeding events and their attributes. On this data set, we evaluated three state-of-the-art DL architectures for the bleeding event relation classification task, namely, convolutional neural network (CNN), attention-guided graph convolutional network (AGGCN), and Bidirectional Encoder Representations from Transformers (BERT). We used three BERT-based models, namely, BERT pretrained on biomedical data (BioBERT), BioBERT pretrained on clinical text (Bio+Clinical BERT), and BioBERT pretrained on EHR notes (EhrBERT).

RESULTS

Our experiments showed that the BERT-based models significantly outperformed the CNN and AGGCN models. Specifically, BioBERT achieved a macro F1 score of 0.842, outperforming both the AGGCN (macro F1 score, 0.828) and CNN models (macro F1 score, 0.763) by 1.4% (P<.001) and 7.9% (P<.001), respectively.

CONCLUSIONS

In this comprehensive study, we explored and compared different DL systems to classify relations between bleeding events and other medical concepts. On our corpus, BERT-based models outperformed other DL models for identifying the relations of bleeding-related entities. In addition to pretrained contextualized word representation, BERT-based models benefited from the use of target entity representation over traditional sequence representation.

Pan-Cancer Metastasis Prediction Based on Graph Deep Learning Method

Tumor metastasis is the major cause of mortality from cancer. From this perspective, detecting cancer gene expression and transcriptome changes is important for exploring tumor metastasis molecular mechanisms and cellular events. Precisely estimating a patient’s cancer state and prognosis is the key challenge to develop a patient’s therapeutic schedule. In the recent years, a variety of machine learning techniques widely contributed to analyzing real-world gene expression data and predicting tumor outcomes. In this area, data mining and machine learning techniques have widely contributed to gene expression data analysis by supplying computational models to support decision-making on real-world data. Nevertheless, limitation of real-world data extremely restricted model predictive performance, and the complexity of data makes it difficult to extract vital features. Besides these, the efficacy of standard machine learning pipelines is far from being satisfactory despite the fact that diverse feature selection strategy had been applied. To address these problems, we developed directed relation-graph convolutional network to provide an advanced feature extraction strategy. We first constructed gene regulation network and extracted gene expression features based on relational graph convolutional network method. The high-dimensional features of each sample were regarded as an image pixel, and convolutional neural network was implemented to predict the risk of metastasis for each patient. Ten cross-validations on 1,779 cases from The Cancer Genome Atlas show that our model’s performance (area under the curve, AUC = 0.837; area under precision recall curve, AUPRC = 0.717) outstands that of an existing network-based method (AUC = 0.707, AUPRC = 0.555).

An Attention-Enhanced Multi-Scale and Dual Sign Language Recognition Network Based on a Graph Convolution Network

Sign language is the most important way of communication for hearing-impaired people. Research on sign language recognition can help normal people understand sign language. We reviewed the classic methods of sign language recognition, and the recognition accuracy is not high enough because of redundant information, human finger occlusion, motion blurring, the diversified signing styles of different people, and so on. To overcome these shortcomings, we propose a multi-scale and dual sign language recognition Network (SLR-Net) based on a graph convolutional network (GCN). The original input data was RGB videos. We first extracted the skeleton data from them and then used the skeleton data for sign language recognition. SLR-Net is mainly composed of three sub-modules: multi-scale attention network (MSA), multi-scale spatiotemporal attention network (MSSTA) and attention enhanced temporal convolution network (ATCN). MSA allows the GCN to learn the dependencies between long-distance vertices; MSSTA can directly learn the spatiotemporal features; ATCN allows the GCN network to better learn the long temporal dependencies. The three different attention mechanisms, multi-scale attention mechanism, spatiotemporal attention mechanism, and temporal attention mechanism, are proposed to further improve the robustness and accuracy. Besides, a keyframe extraction algorithm is proposed, which can greatly improve efficiency by sacrificing a little accuracy. Experimental results showed that our method can reach 98.08% accuracy rate in the CSL-500 dataset with a 500-word vocabulary. Even on the challenging dataset DEVISIGN-L with a 2000-word vocabulary, it also reached a 64.57% accuracy rate, outperforming other state-of-the-art sign language recognition methods.

DS-GCNs: Connectome Classification using Dynamic Spectral Graph Convolution Networks with Assistant Task Training

Functional connectivity (FC) matrices measure the regional interactions in the brain and have been widely used in neurological brain disease classification. A brain network, also named as connectome, could form a graph structure naturally, the nodes of which are brain regions and the edges are interregional connectivity. Thus, in this study, we proposed novel graph convolutional networks (GCNs) to extract efficient disease-related features from FC matrices. Considering the time-dependent nature of brain activity, we computed dynamic FC matrices with sliding windows and implemented a graph convolution-based LSTM (long short-term memory) layer to process dynamic graphs. Moreover, the demographics of patients were also used as additional outputs to guide the classification. In this paper, we proposed to utilize the demographic information as extra outputs and to share parameters among three networks predicting subject status, gender, and age, which serve as assistant tasks. We tested the performance of the proposed architecture in ADNI II dataset to classify Alzheimer's disease patients from normal controls. The classification accuracy, sensitivity, and specificity reach 90.0%, 91.7%, and 88.6%, respectively, on ADNI II dataset.

Automatic Localization of Landmarks in Craniomaxillofacial CBCT Images Using a Local Attention-Based Graph Convolution Network

Landmark localization is an important step in quantifying craniomaxillofacial (CMF) deformities and designing treatment plans of reconstructive surgery. However, due to the severity of deformities and defects (partially missing anatomy), it is difficult to automatically and accurately localize a large set of landmarks simultaneously. In this work, we propose two cascaded networks for digitizing 60 anatomical CMF landmarks in cone-beam computed tomography (CBCT) images. The first network is a U-Net that outputs heatmaps for landmark locations and landmark features extracted with a local attention mechanism. The second network is a graph convolution network that takes the features extracted by the first network as input and determines whether each landmark exists via binary classification. We evaluated our approach on 50 sets of CBCT scans of patients with CMF deformities and compared them with state-of-the-art methods. The results indicate that our approach can achieve an average detection error of 1.47mm with a false positive rate of 19%, outperforming related methods.

Towards Robust and Accurate Detection of Abnormalities in Musculoskeletal Radiographs with a Multi-Network Model

This study proposes a novel multi-network architecture consisting of a multi-scale convolution neural network (MSCNN) with fully connected graph convolution network (GCN), named MSCNN-GCN, for the detection of musculoskeletal abnormalities via musculoskeletal radiographs. To obtain both detailed and contextual information for a better description of the characteristics of the radiographs, the designed MSCNN contains three subnetwork sequences (three different scales). It maintains high resolution in each sub-network, while fusing features with different resolutions. A GCN structure was employed to demonstrate global structure information of the images. Furthermore, both the outputs of MSCNN and GCN were fused through the concat of the two feature vectors from them, thus making the novel framework more discriminative. The effectiveness of this model was verified by comparing the performance of radiologists and three popular CNN models (DenseNet169, CapsNet, and MSCNN) with three evaluation metrics (Accuracy, F1 score, and Kappa score) using the MURA dataset (a large dataset of bone X-rays). Experimental results showed that the proposed framework not only reached the highest accuracy, but also demonstrated top scores on both F1 metric and kappa metric. This indicates that the proposed model achieves high accuracy and strong robustness in musculoskeletal radiographs, which presents strong potential for a feasible scheme with intelligent medical cases.

Transcriptional Regulatory Network of GA Floral Induction Pathway in LA Hybrid Lily

Background: The LA hybrid lily ‘Aladdin’ has both excellent traits of Longiflorum hybrids and Asiatic hybrids—such as big and vivid flower, strong stem, high self-propagation coefficient, and shorter low temperature time required to release bulb dormancy in contrast to Oriental hybrids. A genome-wide transcriptional analysis using transcriptome RNA-Seq was performed in order to explore whether there is a gibberellin floral induction pathway in the LA hybrid lily. Subsequently, gene co-expression network analysis was used to analyze the possible interactions of key candidate genes screened from transcriptome data. At the same time, a series of physiological, biochemical, and cultivation tests were carried out. Results: The content of five endogenous hormones changed sharply in the shoot apex during the treatment of 200 mg/L exogenous gibberellin and the ratio of ABA/GA₃ dropped and stayed at a lower level after 4 hours’ treatment from the higher levels initially, reaching a dynamic balance. In addition, the metabolism of carbohydrates in the bulbs increase during exogenous gibberellin treatment. A total of 124,041 unigenes were obtained by RNA-seq. With the transcriptome analysis, 48,927 unigenes and 48,725 unigenes respectively aligned to the NR database and the Uniprot database. 114,138 unigenes, 25,369 unigenes, and 19,704 unigenes respectively aligned to the COG, GO, and KEGG databases. 2148 differentially expression genes (DEGs) were selected with the indicators RPKM ≥ 0, FDR ≤ 0.05 and |log2(ratio)| ≥ 2. The number of the upregulated unigenes was significantly more than the number of the downregulated unigenes. Some MADS-box genes related to flowering transformation—such as AGL20, SOC1, and CO—were found to be upregulated. A large number of gibberellin biosynthesis related genes such as GA2ox, GA3ox, GA20ox, Cytochrome P450, CYP81, and gibberellin signal transduction genes such as DELLA, GASA, and GID1 were significantly differentially expressed. The plant hormones related genes such as NCED3 and sugar metabolism related genes such as α-amylase, sucrose synthase hexokinase, and so on were also found expressing differentially. In addition, stress resistance related genes such as LEA1, LEA2, LEA4, serine/threonine protein kinase, LRR receptor-like serine/threonine protein kinase, P34 kinase, histidine kinase 3 and epigenetic related genes in DNA methylation, histone methylation, acetylation, ubiquitination of ribose were also found. Particularly, a large number of transcription factors responsive to the exogenous gibberellin signal including WRKY40, WRKY33, WRKY27, WRKY21, WRKY7, MYB, AP2/EREBP, bHLH, NAC1, NAC2, and NAC11 were found to be specially expressing. 30 gene sequences were selected from a large number of differentially expressed candidate genes for qRT-PCR expression verification (0, 2, 4, 8, and 16 h) and compared with the transcriptome expression levels. Conclusions: 200mg/L exogenous GA₃ can successfully break the bulb’s dormancy of the LA hybrid lily and significantly accelerated the flowering process, indicating that gibberellin floral induction pathway is present in the LA lily ‘Aladdin’. With the GCNs analysis, two second messenger G protein-coupled receptor related genes that respond to gibberellin signals in the cell were discovered. The downstream transport proteins such as AMT, calcium transport ATPase, and plasma membrane ATPase were also discovered participating in GA signal transduction. Transcription factors including WRKY7, NAC2, NAC11, and CBF specially regulated phosphorylation and glycosylation during the ubiquitination degradation process of DELLA proteins. These transcription factors also activated in abscisic acid metabolism. A large number of transcription factors such as WRKY21, WRKY22, NAC1, AP2, EREB1, P450, and CYP81 that both regulate gibberellin signaling and low-temperature signals have also been found. Finally, the molecular mechanism of GA floral induction pathway in the LA hybrid lily ‘Aladdin’ was constructed.

Identifying condition specific key genes from basal-like breast cancer gene expression data

Mining patterns of co-expressed genes across the subset of conditions help to narrow down the search space for the analysis of gene expression data. Identifying conditions specific key genes from the large-scale gene expression data is a challenging task. The conditions specific key gene signifies functional behavior of a group of co-expressed genes across the subset of conditions and can be act as biomarkers of the diseases. In this paper, we have propose a novel approach for identification of conditions specific key genes from Basal-Like Breast Cancer (BLBC) disease using biclustering algorithm and Gene Co-expression Network (GCN). The proposed approach is a two-stage approach. In the first stage, significant biclusters have been extracted with the help of 'runibic' biclustering algorithm. The second stage identifies conditions specific key genes from the extracted significant biclusters with the help of GCN. By using difference matrix and gene correlation matrix, we have constructed biologically meaningful and statistically strong GCN. Also, presented the proposed approach with the help of a process diagram and demonstrated the procedure with an example of bicluster number 93 (Bic93). From the experimental results, we observed that 95% and 85% of the extracted biclusters are found to be biologically significant at the p-values less than 0.05 and 0.01 respectively. We have compared proposed approach with the Weighted Gene Co-expression Network Analysis (WGCNA) based approach. From the comparison, our approach has performed effectively and extracted biologically significant biclusters. Also, identified conditions specific key genes which cannot be extracted using the WGCNA based approach. Some of the important identified known key genes are PIK3CA, SHC3, ERBB2, SHC4, PTOV1, STAG1, ZNF215 etc. These key genes can be used as a diagnostic and prognostic biomarker for the BLBC disease after the rigorous analysis. The identified conditions specific key genes can be helpful to reduce the analysis time and increase the accuracy of further research such as biomarker identification, drug target discovery etc.

IDO1 metabolites activate β-catenin signaling to promote cancer cell proliferation and colon tumorigenesis in mice

BACKGROUND & AIMS

Indoleamine 2,3 dioxygenase-1 (IDO1) catabolizes tryptophan along the kynurenine pathway. Although IDO1 is expressed in inflamed and neoplastic epithelial cells of the colon, its role in colon tumorigenesis is not well understood. We used genetic and pharmacologic approaches to manipulate IDO1 activity in mice with colitis-associated cancer and human colon cancer cell lines.

METHODS

C57Bl6 wild-type (control), IDO1-/-, Rag1-/-, and Rag1/IDO1 double-knockout mice were exposed to azoxymethane and dextran sodium sulfate to induce colitis and tumorigenesis. Colitis severity was assessed by measurements of disease activity, cytokine levels, and histologic analysis. In vitro experiments were conducted using HCT 116 and HT-29 human colon cancer cells. 1-methyl tryptophan and small interfering RNA were used to inhibit IDO1. Kynurenine pathway metabolites were used to simulate IDO1 activity.

RESULTS

C57Bl6 mice given pharmacologic inhibitors of IDO1 and IDO1-/- mice had lower tumor burdens and reduced proliferation in the neoplastic epithelium after administration of dextran sodium sulfate and azoxymethane than control mice. These reductions also were observed in Rag1/IDO1 double-knockout mice compared with Rag1-/- mice (which lack mature adaptive immunity). In human colon cancer cells, blockade of IDO1 activity reduced nuclear and activated β-catenin, transcription of its target genes (cyclin D1 and Axin2), and, ultimately, proliferation. Exogenous administration of IDO1 pathway metabolites kynurenine and quinolinic acid led to activation of β-catenin and proliferation of human colon cancer cells, and increased tumor growth in mice.

CONCLUSIONS

IDO1, which catabolizes tryptophan, promotes colitis-associated tumorigenesis in mice, independent of its ability to limit T-cell-mediated immune surveillance. The epithelial cell-autonomous survival advantage provided by IDO1 to colon epithelial cells indicate its potential as a therapeutic target.