Graph transformer networks have received more attention in hyperspectral image (HSI) classification. However, they overlooked the influence of graph connectivity strength in positional encoding and distribution. In order to address the above deficiencies, we proposed the novel graph transformer with structural embedding and training (GTSET) for HSI classification. Specifically, the structural embedding module firstly aimed at extracting effectively local and non-local feature information via patch-based distance encoding and centrality correlation coefficients based on graph connectivity strength, alleviating spectral variability. Secondly, the structural training module aimed at addressing imbalanced structural position distribution of labeled samples by leveraging the topological graph connectivity to determine their structural position distribution and reweighting the influence of labeled samples on the graph transformer training stage, exploring the guiding role of labeled samples in low spatial resolution of HSI. Next, we further refine training weights based on the spectral feature smoothness of labeled samples. Finally, comprehensive experiments on three real-world HSI datasets demonstrate that the GTSET achieves superior performance in HSI classification with limited labeled samples, compared to other popular classification methods. Implementation of GTSET, along with examples, can be found on the GitHub repository: https://github.com/xuchengchao0/GTSET.

Single-cell RNA sequencing (scRNA-seq) enables the reconstruction of cell type-specific gene regulatory networks (GRNs), offering detailed insights into gene regulation at high resolution. While graph neural networks have become widely used for GRN inference, their message-passing mechanisms are often limited by issues such as over-smoothing and over-squashing, which hinder the preservation of essential network structure. To address these challenges, we propose a novel graph transformer-based model, AttentionGRN, which leverages soft encoding to enhance model expressiveness and improve the accuracy of GRN inference from scRNA-seq data. Furthermore, the GRN-oriented message aggregation strategies are designed to capture both the directed network structure information and functional information inherent in GRNs. Specifically, we design directed structure encoding to facilitate the learning of directed network topologies and employ functional gene sampling to capture key functional modules and global network structure. Our extensive experiments, conducted on 88 datasets across two distinct tasks, demonstrate that AttentionGRN consistently outperforms existing methods. Furthermore, AttentionGRN has been successfully applied to reconstruct cell type-specific GRNs for human mature hepatocytes, revealing novel hub genes and previously unidentified transcription factor-target gene regulatory associations.

Reconstructing high-resolution gene regulatory networks (GRNs) based on single-cell RNA sequencing data provides an opportunity to gain insight into disease pathogenesis. At present, there are a large number of GRN reconstruction methods based on graph neural networks, and they can obtain excellent performance in GRN inference by extracting network structure features. However, most of these methods fail to fully exploit the directional characteristics or even ignore them when extracting network structural features. To this end, a novel framework called GAEDGRN is proposed based on gravity-inspired graph autoencoder (GIGAE) to infer potential causal relationships between genes. Among them, GIGAE can help us capture the complex directed network topology in GRN. Additionally, due to the uneven distribution of the latent vectors generated by the graph autoencoder, a random walk-based method is used to regularize the latent vectors learnt by the encoder. Furthermore, considering that some genes in GRN usually have a significant impact on biological functions, GAEDGRN designs a gene importance score calculation method and pays attention to genes with high importance in the process of GRN reconstruction. Experimental results on seven cell types of three GRN types show that GAEDGRN achieves high accuracy and strong robustness. Moreover, a case study on human embryonic stem cells demonstrates that GAEDGRN can help identify important genes.

BACKGROUND: The progression of cancer is driven by the accumulation of mutations in driver genes. Many researches promote to identify cancer driver genes. However, most of them ignore the high-order features in the network. RESULT: In this study, we propose a novel method MLGCN-Driver based on multi-layer graph convolutional neural networks (GCN) to boost driver gene identification. MLGCN-Driver employs multi-layer GCN with initial residual connections and identity mappings to learn biological multi-omics features within biological networks. In addition, node2vec algorithm is used to extract the topological structure features of the biological network, and then the features are fed into another multi-layer GCN for feature learning. Meanwhile, the initial residual connections and identity mappings mitigate the over-smooth of features. Finally, the probability of each gene being a driver gene is calculated based on low-dimensional biological features and topological features. CONCLUSION: We applied the MLGCN-Driver on pan-cancer dataset and cancer type-specific datasets. Experimental results demonstrate the excellent performance of MLGCN-Driver in terms of the area under the ROC curve (AUC) and the area under the precision-recall curve (AUPRC) when compared with state-of-the-art approaches.

Intracranial aneurysm (IA) is a vascular disease of the brain arteries caused by pathological vascular dilation, which can result in subarachnoid hemorrhage if ruptured. Automatically classification and segmentation of intracranial aneurysms are essential for their diagnosis and treatment. However, the majority of current research is focused on two-dimensional images, ignoring the 3D spatial information that is also critical. In this work, we propose a novel dual-branch fusion network called the Point Cloud and Multi-View Medical Neural Network (PMMNet) for IA classification and segmentation. Specifically, one branch based on 3D point clouds serves the purpose of extracting spatial features, whereas the other branch based on multi-view images acquires 2D pixel features. Ultimately, the two types of features are fused for IA classification and segmentation. To extract both local and global features from 3D point clouds, Multilayer Perceptron (MLP) and the attention mechanism are used in parallel. In addition, a SPSA module is proposed for multi-view image feature learning, which extracts more exquisite channel and spatial multi-scale features from 2D images. Experiments conducted on the IntrA dataset outperform other state-of-the-art methods, demonstrating that the proposed PMMNet exhibits strong superiority on the medical 3D dataset. We also obtain competitive results on public datasets, including ModelNet40, ModelNet10, and ShapeNetPart, which further validate the robustness and generality of the PMMNet.

Segmentation of thoracic and abdominal organs is crucial for accurate disease diagnosis, surgical planning, and long-term health management of patients. Deep learning often depends on large quantities of high-quality training data to achieve superior results. However, the inherent complexity and sparsity of medical images require advanced models with greater learning capabilities. Additionally, significant morphological differences between organs often result in inaccurate and false-positive segmentation. To alleviate these issues, we propose a Dual Attention (DA) and Multi-Resolution Inputs (MI) based network (DAMNet) for the segmentation of thoracic and abdominal organs. Specially, DAMNet utilizes a single encoder and dual decoders. The encoder integrates MI with the Transformer to achieve multi-level feature fusion and capture global image relationships, improving the model’s capability to process intricate image data. In the decoders, Residual U-blocks (RSU) and the DA module consisting of Spatial Multi-Scale Cross-Axis Attention (SMCA) and Convolutional Self-Attention (CSA) are used as two decoder branches, respectively. The design of decoders allows the model to extract detailed information from different encoding layers from two perspectives, thereby reducing inaccuracies in segmentation. We perform thorough experiments and evaluations using three publicly available datasets: Synapse, SegTHOR, and THoracic. The experimental results indicate that our proposed DAMNet model demonstrates exceptional proficiency in segmenting thoracic and abdominal organs.

The inference of gene regulatory networks (GRNs) from gene expression profiles has been a key issue in systems biology, prompting many researchers to develop diverse computational methods. However, most of these methods do not reconstruct directed GRNs with regulatory types because of the lack of benchmark datasets or defects in the computational methods. Here, we collect benchmark datasets and propose a deep learning-based model, DeepFGRN, for reconstructing fine gene regulatory networks (FGRNs) with both regulation types and directions. In addition, the GRNs of real species are always large graphs with direction and high sparsity, which impede the advancement of GRN inference. Therefore, DeepFGRN builds a node bidirectional representation module to capture the directed graph embedding representation of the GRN. Specifically, the source and target generators are designed to learn the low-dimensional dense embedding of the source and target neighbors of a gene, respectively. An adversarial learning strategy is applied to iteratively learn the real neighbors of each gene. In addition, because the expression profiles of genes with regulatory associations are correlative, a correlation analysis module is designed. Specifically, this module not only fully extracts gene expression features, but also captures the correlation between regulators and target genes. Experimental results show that DeepFGRN has a competitive capability for both GRN and FGRN inference. Potential biomarkers and therapeutic drugs for breast cancer, liver cancer, lung cancer and coronavirus disease 2019 are identified based on the candidate FGRNs, providing a possible opportunity to advance our knowledge of disease treatments.

Inferring gene regulatory network (GRN) is one of the important challenges in systems biology, and many outstanding computational methods have been proposed; however there remains some challenges especially in real datasets. In this study, we propose Directed Graph Convolutional neural network-based method for GRN inference (DGCGRN). To better understand and process the directed graph structure data of GRN, a directed graph convolutional neural network is conducted which retains the structural information of the directed graph while also making full use of neighbor node features. The local augmentation strategy is adopted in graph neural network to solve the problem of poor prediction accuracy caused by a large number of low-degree nodes in GRN. In addition, for real data such as E.coli, sequence features are obtained by extracting hidden features using Bi-GRU and calculating the statistical physicochemical characteristics of gene sequence. At the training stage, a dynamic update strategy is used to convert the obtained edge prediction scores into edge weights to guide the subsequent training process of the model. The results on synthetic benchmark datasets and real datasets show that the prediction performance of DGCGRN is significantly better than existing models. Furthermore, the case studies on bladder uroepithelial carcinoma and lung cancer cells also illustrate the performance of the proposed model.

Abstract MicroRNAs (miRNAs) are closely related to numerous complex human diseases, therefore, exploring miRNA-disease associations (MDAs) can help people gain a better understanding of complex disease mechanism. An increasing number of computational methods have been developed to predict MDAs. However, the sparsity of the MDAs may hinder the performance of many methods. In addition, many methods fail to capture the nonlinear relationships of miRNA-disease network and inadequately leverage the features of network and neighbor nodes. In this study, we propose a deep matrix factorization model with variational autoencoder (DMFVAE) to predict potential MDAs. DMFVAE first decomposes the original association matrix and the enhanced association matrix, in which the enhanced association matrix is enhanced by self-adjusting the nearest neighbor method, to obtain sparse vectors and dense vectors, respectively. Then, the variational encoder is employed to obtain the nonlinear latent vectors of miRNA and disease for the sparse vectors, and meanwhile, node2vec is used to obtain the network structure embedding vectors of miRNA and disease for the dense vectors. Finally, sample features are acquired by combining the latent vectors and network structure embedding vectors, and the final prediction is implemented by convolutional neural network with channel attention. To evaluate the performance of DMFVAE, we conduct five-fold cross validation on the HMDD v2.0 and HMDD v3.2 datasets and the results show that DMFVAE performs well. Furthermore, case studies on lung neoplasms, colon neoplasms, and esophageal neoplasms confirm the ability of DMFVAE in identifying potential miRNAs for human diseases.

Antiviral peptides (AVPs) are widely found in animals and plants, with high specificity and strong sensitivity to drug-resistant viruses. However, due to the great heterogeneity of different viruses, most of the AVPs have specific antiviral activities. Therefore, it is necessary to identify the specific activities of AVPs on virus types. Most existing studies only identify AVPs, with only a few studies identifying subclasses by training multiple binary classifiers. We develop a two-stage prediction tool named FFMAVP that can simultaneously predict AVPs and their subclasses. In the first stage, we identify whether a peptide is AVP or not. In the second stage, we predict the six virus families and eight species specifically targeted by AVPs based on two multiclass tasks. Specifically, the feature extraction module in the two-stage task of FFMAVP adopts the same neural network structure, in which one branch extracts features based on amino acid feature descriptors and the other branch extracts sequence features. Then, the two types of features are fused for the following task. Considering the correlation between the two tasks of the second stage, a multitask learning model is constructed to improve the effectiveness of the two multiclass tasks. In addition, to improve the effectiveness of the second stage, the network parameters trained through the first-stage data are used to initialize the network parameters in the second stage. As a demonstration, the cross-validation results, independent test results and visualization results show that FFMAVP achieves great advantages in both stages.

Gene regulatory networks (GRNs) participate in many biological processes, and reconstructing them plays an important role in systems biology. Although many advanced methods have been proposed for GRN reconstruction, their predictive performance is far from the ideal standard, so it is urgent to design a more effective method to reconstruct GRN. Moreover, most methods only consider the gene expression data, ignoring the network structure information contained in GRN. In this study, we propose a supervised model named CNNGRN, which infers GRN from bulk time-series expression data via convolutional neural network (CNN) model, with a more informative feature. Bulk time series gene expression data imply the intricate regulatory associations between genes, and the network structure feature of ground-truth GRN contains rich neighbor information. Hence, CNNGRN integrates the above two features as model inputs. In addition, CNN is adopted to extract intricate features of genes and infer the potential associations between regulators and target genes. Moreover, feature importance visualization experiments are implemented to seek the key features. Experimental results show that CNNGRN achieved competitive performance on benchmark datasets compared to the state-of-the-art computational methods. Finally, hub genes identified based on CNNGRN have been confirmed to be involved in biological processes through literature.

Immunotherapy is an excellent treatment option for many solid tumors, and the therapeutic effect has been proved in clinical. Microsatellite instability (MSI) has been an important predictive marker for response to immune checkpoint inhibitors for Colorectal cancer (CRC). CRC patients with high-level MSI can be provided immunotherapy and benefit from it. Some studies have attempted to predict MSI using pathological images based on deep learning, but the accuracy of the prediction model needs to be improved. In this study, considering that different scales of pathological images contain various levels of information, we propose a novel method, named MSIUMP, to predict microsatellite instability using multi-scale pathological images. We first predict the MSI of a patient based on deep learning model using different scale pathological images, and then integrate the results by ensemble learning to improve the generalization and robustness. In addition, a convolutional neural network model modified on the basis of EfficientNet is used to extract the information of patches from pathological images at different scales. Our method achieved the areas under the receiver operating characteristic curves (AUC) of 0.9096 in the internal test dataset TCGA-CRC and the AUC of 0.9619 in the external independent validation dataset PAIP2020. These results demonstrate the potential of our method as a prediction tool for microsatellite instability in colorectal cancer.

BACKGROUND: As one of the deadliest diseases in the world, cancer is driven by a few somatic mutations that disrupt the normal growth of cells, and leads to abnormal proliferation and tumor development. The vast majority of somatic mutations did not affect the occurrence and development of cancer; thus, identifying the mutations responsible for tumor occurrence and development is one of the main targets of current cancer treatments. RESULTS: To effectively identify driver genes, we adopted a semi-local centrality measure and gene mutation effect function to assess the effect of gene mutations on changes in gene expression patterns. Firstly, we calculated the mutation score for each gene. Secondly, we identified differentially expressed genes (DEGs) in the cohort by comparing the expression profiles of tumor samples and normal samples, and then constructed a local network for each mutation gene using DEGs and mutant genes according to the protein-protein interaction network. Finally, we calculated the score of each mutant gene according to the objective function. The top-ranking mutant genes were selected as driver genes. We name the proposed method as mutations effect and network centrality. CONCLUSIONS: Four types of cancer data in The Cancer Genome Atlas were tested. The experimental data proved that our method was superior to the existing network-centric method, as it was able to quickly and easily identify driver genes and rare driver factors.

BACKGROUND: Circular RNAs (circRNAs) are a class of single-stranded RNA molecules with a closed-loop structure. A growing body of research has shown that circRNAs are closely related to the development of diseases. Because biological experiments to verify circRNA-disease associations are time-consuming and wasteful of resources, it is necessary to propose a reliable computational method to predict the potential candidate circRNA-disease associations for biological experiments to make them more efficient. RESULTS: In this paper, we propose a double matrix completion method (DMCCDA) for predicting potential circRNA-disease associations. First, we constructed a similarity matrix of circRNA and disease according to circRNA sequence information and semantic disease information. We also built a Gauss interaction profile similarity matrix for circRNA and disease based on experimentally verified circRNA-disease associations. Then, the corresponding circRNA sequence similarity and semantic similarity of disease are used to update the association matrix from the perspective of circRNA and disease, respectively, by matrix multiplication. Finally, from the perspective of circRNA and disease, matrix completion is used to update the matrix block, which is formed by splicing the association matrix obtained in the previous step with the corresponding Gaussian similarity matrix. Compared with other approaches, the model of DMCCDA has a relatively good result in leave-one-out cross-validation and five-fold cross-validation. Additionally, the results of the case studies illustrate the effectiveness of the DMCCDA model. CONCLUSION: The results show that our method works well for recommending the potential circRNAs for a disease for biological experiments.

In this paper, we presented a network-based method, named DriverFinder, by filtering frequently mutated genes just because of their large size, and comparing tumor expression with normal expression data to obtain gene expression outliers which are more likely to be cancer genes. Then greedy algorithm was applied to prioritize candidate driver genes. The proposed method can not only indentify frequently mutated genes, but also novel and infrequently mutated driver genes.

BACKGROUND: Cancer is a complex disease which is characterized by the accumulation of genetic alterations during the patient's lifetime. With the development of the next-generation sequencing technology, multiple omics data, such as cancer genomic, epigenomic and transcriptomic data etc., can be measured from each individual. Correspondingly, one of the key challenges is to pinpoint functional driver mutations or pathways, which contributes to tumorigenesis, from millions of functional neutral passenger mutations. RESULTS: In this paper, in order to identify driver genes effectively, we applied a generalized additive model to mutation profiles to filter genes with long length and constructed a new gene-gene interaction network. Then we integrated the mutation data and expression data into the gene-gene interaction network. Lastly, greedy algorithm was used to prioritize candidate driver genes from the integrated data. We named the proposed method Length-Net-Driver (LNDriver). CONCLUSIONS: Experiments on three TCGA datasets, i.e., head and neck squamous cell carcinoma, kidney renal clear cell carcinoma and thyroid carcinoma, demonstrated that the proposed method was effective. Also, it can identify not only frequently mutated drivers, but also rare candidate driver genes.