Cell-cell communication plays a pivotal role in tissue development, homeostasis, and disease progression. However, existing approaches often lack systematic modeling of networklevel structure and often overlook spatial modular organization. To address these limitations, we propose CCNVGAE, an unsupervised framework based on variational graph autoencoder (VGAE) that integrates spatial coordinates, ligand-receptor gene expression and graph structural information for functional module identification. Experiments on spatial transcriptomic datasets from mouse brain and human lung cancer demonstrate that CCNVGAE achieves competitive performance in spatial segmentation and biological interpretability. Moreover, it effectively uncovers module specific ligand-receptor interactions and their potential regulatory mechanisms, providing a powerful tool for elucidating spatial communication structures in complex tissues.

Accurately annotating non-coding regulatory DNA, critical for gene regulation and implicated in numerous diseases, remains a major challenge. Current deep learning models struggle to integrate local motifs with long-range dependencies due to limited architectures and simplistic feature fusion. We propose DCAM, a dual-branch neural network where a global branch captures long-range context via a CNN-BiLSTM, while a local branch extracts fine-grained motifs using dna2vec and CNNs. These features are dynamically integrated via a sequential channel-spatial attention mechanism. Evaluated on diverse benchmark datasets, DCAM significantly outperforms state-of-the-art models. DCAM also accurately predicts the functional impact of single nucleotide variants (SNVs), establishing it as a powerful tool for prioritizing pathogenic non-coding variants.

Inferring gene regulatory networks (GRNs) allows us to obtain a deeper understanding of cellular function and disease pathogenesis. Recent advances in single-cell RNA sequencing (scRNA-seq) technology have improved the accuracy of GRN inference. However, many methods for inferring individual GRNs from scRNA-seq data are limited because they overlook intercellular heterogeneity and similarities between different cell subpopulations, which are often present in the data. Here, we propose a deep learning-based framework, DeepGRNCS, for jointly inferring GRNs across cell subpopulations. We follow the commonly accepted hypothesis that the expression of a target gene can be predicted based on the expression of transcription factors (TFs) due to underlying regulatory relationships. We initially processed scRNA-seq data by discretizing data scattering using the equal-width method. Then, we trained deep learning models to predict target gene expression from TFs. By individually removing each TF from the expression matrix, we used pre-trained deep model predictions to infer regulatory relationships between TFs and genes, thereby constructing the GRN. Our method outperforms existing GRN inference methods for various simulated and real scRNA-seq datasets. Finally, we applied DeepGRNCS to non-small cell lung cancer scRNA-seq data to identify key genes in each cell subpopulation and analyzed their biological relevance. In conclusion, DeepGRNCS effectively predicts cell subpopulation-specific GRNs. The source code is available at https://github.com/Nastume777/DeepGRNCS.

Identification of disease-associated genes is a crucial step in unraveling the underlying molecular mechanisms of complex diseases. As a potent computational tool for recognizing potential associations between genes and diseases, heterogeneous graphs allow for more comprehensive and accurate modeling and analysis of intricate molecular interactions within unified networks. In this study, a method based on Heterogeneous Graph Transformer (HGT) is proposed for mining associations between genes and diseases from heterogeneous graphs. Six types of association data are collected among four biological components: Protein-coding (PC) genes, long non-coding RNAs (lncRNAs), microRNAs, and diseases. A series of processing and integration steps are performed on these raw data to construct an extensive heterogeneous graph. For the vertexes in the graph, their description texts are collected and specialized natural language processing model, BioBERT, is employed to generate semantic vectors as the initial feature vectors for the vertexes. The topological structure of this heterogeneous graph along with the nodes' initial features collectively serve as input to the HGT, enabling information aggregation. Finally, the gene and disease vector representations produced by HGT are used to compute the association scores, which are further applied in the association prediction. Experimental results demonstrate that the proposed method outperforms existing state-of-the-art methods in terms of accuracy and exhibits a remarkably high degree of reliability. Moreover, the method has the capability to implicitly capture meaningful biological mechanisms through learned implicit meta-paths, which can offer substantial guidance for subsequent biomedical validations.

BACKGROUND: Precise disease module is conducive to understanding the molecular mechanism of disease causation and identifying drug targets. However, due to the fragmentization of disease module in incomplete human interactome, how to determine connectivity pattern and detect a complete neighbourhood of disease based on this is still an open question. RESULTS: In this paper, we perform exploratory analysis leading to an important observation that through a few intermediate nodes, most separate connected components formed by disease-associated proteins can be effectively connected and eventually form a complete disease module. And based on the topological properties of these intermediate nodes, we propose a connect separate connected components (C3) method to detect a succinct disease module by introducing a relatively small number of intermediate nodes, which allows us to obtain more pure disease module than other methods. Then we apply C3 across a large corpus of diseases to validate this connectivity pattern of disease module. Furthermore, the connectivity of the perturbed genes in multi-omics data such as The Cancer Genome Atlas also fits this pattern. CONCLUSIONS: C3 tool is not only useful in detecting a clearly-defined connected disease neighbourhood of 299 diseases and cancer with multi-omics data, but also helpful in better understanding the interconnection of phenotypically related genes in different omics data and studying complex pathological processes.

BACKGROUND: In recent years, increasing evidences have indicated that long non-coding RNAs (lncRNAs) are deeply involved in a wide range of human biological pathways. The mutations and disorders of lncRNAs are closely associated with many human diseases. Therefore, it is of great importance to predict potential associations between lncRNAs and complex diseases for the diagnosis and cure of complex diseases. However, the functional mechanisms of the majority of lncRNAs are still remain unclear. As a result, it remains a great challenge to predict potential associations between lncRNAs and diseases. RESULTS: Here, we proposed a new method to predict potential lncRNA-disease associations. First, we constructed a bipartite network based on known associations between diseases and lncRNAs/protein coding genes. Then the cluster association scores were calculated to evaluate the strength of the inner relationships between disease clusters and gene clusters. Finally, the gene-disease association scores are defined based on disease-gene cluster association scores and used to measure the strength for potential gene-disease associations. CONCLUSIONS: Leave-One Out Cross Validation (LOOCV) and 5-fold cross validation tests were implemented to evaluate the performance of our method. As a result, our method achieved reliable performance in the LOOCV (AUCs of 0.8169 and 0.8410 based on Yang's dataset and Lnc2cancer 2.0 database, respectively), and 5-fold cross validation (AUCs of 0.7573 and 0.8198 based on Yang's dataset and Lnc2cancer 2.0 database, respectively), which were significantly higher than the other three comparative methods. Furthermore, our method is simple and efficient. Only the known gene-disease associations are exploited in a graph manner and further new gene-disease associations can be easily incorporated in our model. The results for melanoma and ovarian cancer have been verified by other researches. The case studies indicated that our method can provide informative clues for further investigation.

MOTIVATION: Genome-scale CRISPR/Cas9 system has been a democratized gene editing technique and widely used to investigate gene functions in some biological processes and diseases especially cancers. Aiming to characterize gene aberrations and assess their effects on cancer, we designed a pipeline to identify the essential genes for pan-cancer. METHODS: CRISPR screening data were used to identify the essential genes that were collected from published data and integrated by Robust Rank Aggregation algorithm. Then, hypergeometrics test and random walks with restart (RWR) were used to predict additional essential genes on broader scale. Finally, the expression status and potential roles of these genes were explored based on TCGA portal and regulatory network analysis. RESULTS: We collected 926 samples from 10 CRISPR-based screening studies involving 33 different types of cancer to identify cancer-essential genes, which consists of 799 protein-coding genes (PCGs) and 97 long non-coding RNAs (lncRNAs). Then, we constructed a 'bi-colored' network with both PCGs and lncRNAs and applied it to predict additional essential genes including 495 PCGs and 280 lncRNAs on a broader scale using hypergeometrics test and RWR. After obtaining all essential genes, we further investigated their potential roles in cancer and found that essential genes have higher and more stable expression levels, and are associated with multiple cancer-associated biological processes and survival time. The regulatory network analysis detected two intriguing modules of essential genes participating in the regulation of cell cycle and ribosome biogenesis in cancer. AVAILABILITY AND IMPLEMENTATION: . SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Determining the dynamics of pathways associated with cancer progression is critical for understanding the etiology of diseases. Advances in biological technology have facilitated the simultaneous genomic profiling of multiple patients at different clinical stages, thus generating the dynamic genomic data for cancers. Such data provide enable investigation of the dynamics of related pathways. However, methods for integrative analysis of dynamic genomic data are inadequate. In this study, we develop a novel nonnegative matrix factorization algorithm for dynamic modules ( NMF-DM), which simultaneously analyzes multiple networks for the identification of stage-specific and dynamic modules. NMF-DM applies the temporal smoothness framework by balancing the networks at the current stage and the previous stage. Experimental results indicate that the NMF-DM algorithm is more accurate than the state-of-the-art methods in artificial dynamic networks. In breast cancer networks, NMF-DM reveals the dynamic modules that are important for cancer stage transitions. Furthermore, the stage-specific and dynamic modules have distinct topological and biochemical properties. Finally, we demonstrate that the stage-specific modules significantly improve the accuracy of cancer stage prediction. The proposed algorithm provides an effective way to explore the time-dependent cancer genomic data.