From Homeostasis to Neuroinflammation: Insights into Cellular and Molecular Interactions and Network Dynamics

Cells, Год журнала: 2025, Номер 14(1), С. 54 - 54

Опубликована: Янв. 5, 2025

Neuroinflammation is a complex and multifaceted process that involves dynamic interactions among various cellular molecular components. This sophisticated interplay supports both environmental adaptability system resilience in the central nervous (CNS) but may be disrupted during neuroinflammation. In this article, we first characterize key players neuroimmune interactions, including microglia, astrocytes, neurons, immune cells, essential signaling molecules such as cytokines, neurotransmitters, extracellular matrix (ECM) components, neurotrophic factors. Under homeostatic conditions, these elements promote cooperation stability, whereas neuroinflammatory states, they drive adaptive responses become pathological if dysregulated. We examine how mediated through actors pathways, create networks regulate CNS functionality respond to injury or inflammation. To further elucidate dynamics, provide insights using multilayer network (MLN) approach, highlighting interconnected nature of under inflammatory conditions. perspective aims enhance our understanding communication mechanisms underlying shifts from homeostasis Applying an MLN approach offers more integrative view adaptability, helping clarify processes identify novel intervention points within layered landscape responses.

Язык: Английский

---

Nonlocal Models in Biology and Life Sciences: Sources, Developments, and Applications

Physics of Life Reviews, Год журнала: 2025, Номер 53, С. 24 - 75

Опубликована: Фев. 27, 2025

Язык: Английский

---

Leveraging graph neural networks for supporting automatic triage of patients

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Май 31, 2024

Abstract Patient triage is crucial in emergency departments, ensuring timely and appropriate care based on correctly evaluating the grade of patient conditions. Triage methods are generally performed by human operator her own experience information that gathered from management process. Thus, it a process can generate errors emergency-level associations. Recently, Traditional heavily rely decisions, which be subjective prone to errors. A growing interest has recently been focused leveraging artificial intelligence (AI) develop algorithms maximize gathering minimize processing. We define implement an AI-based module manage patients’ code assignments departments. It uses historical data department train medical decision-making Data containing relevant information, such as vital signs, symptoms, history, accurately classify patients into categories. Experimental results demonstrate proposed algorithm achieved high accuracy outperforming traditional methods. By using method, we claim healthcare professionals predict severity index guide processing resource allocation.

Язык: Английский

---

Using residue interaction networks to understand protein function and evolution and to engineer new proteins

Current Opinion in Structural Biology, Год журнала: 2024, Номер 89, С. 102922 - 102922

Опубликована: Сен. 26, 2024

Язык: Английский

---

Target and Biomarker Exploration Portal for Drug Discovery

Research Square (Research Square), Год журнала: 2025, Номер unknown

Опубликована: Янв. 10, 2025

The discovery of novel drug targets and precision biomarkers remains a major challenge in development, with traditional differential expression analysis often overlooking key regulatory proteins. Here, we present novel, web-based bioinformatics tool designed to accelerate the process by integrating large-scale biomedical data network techniques. This harnesses machine-learning approaches combine multi-modal datasets, including human genetics, functional genomics, protein-protein interaction networks, decode causal disease mechanisms uncover therapeutic for specific phenotypes. A unique feature is its ability real-time, facilitated efficient cloud-based architecture. Additionally, incorporates an integrated large language model (LLM), which assists researchers exploring interpreting complex biological relationships within generated networks multi-omics data. By offering intuitive, interactive interface, LLM enhances exploration insights, making it easier scientists derive actionable conclusions. powerful integration AI-driven analysis, data, advanced models provides robust framework accelerating identification targets, ultimately advancing field medicine. publicly available at https://pdnet.missouri.edu/.

Язык: Английский

---

Build the virtual cell with artificial intelligence: a perspective for cancer research

Military Medical Research, Год журнала: 2025, Номер 12(1)

Опубликована: Янв. 27, 2025

Язык: Английский

---

Clover: An unbiased method for prioritizing differentially expressed genes using a data‐driven approach

Genes to Cells, Год журнала: 2024, Номер 29(6), С. 456 - 470

Опубликована: Апрель 11, 2024

Abstract Identifying key genes from a list of differentially expressed (DEGs) is critical step in transcriptome analysis. However, current methods, including Gene Ontology analysis and manual annotation, essentially rely on existing knowledge, which highly biased depending the extent literature. As result, understudied genes, some may be associated with important molecular mechanisms, are often ignored or remain obscure. To address this problem, we propose Clover, data‐driven scoring method to specifically highlight genes. Clover aims prioritize mechanisms by integrating three metrics: likelihood appearing DEG list, tissue specificity, number publications. We applied Alzheimer's disease data confirmed that it successfully detected known Moreover, effectively prioritized but potentially druggable Overall, our offers novel approach gene characterization has potential expand understanding functions. an open‐source software written Python3 available GitHub at https://github.com/G708/Clover .

Язык: Английский

---

Detecting tipping points of complex diseases by network information entropy

Briefings in Bioinformatics, Год журнала: 2024, Номер 25(4)

Опубликована: Май 23, 2024

The progression of complex diseases often involves abrupt and non-linear changes characterized by sudden shifts that trigger critical transformations. Identifying these states or tipping points is crucial for understanding disease developing effective interventions. To address this challenge, we have developed a model-free method named Network Information Entropy Edges (NIEE). Leveraging dynamic network biomarkers, sample-specific networks, information entropy theories, NIEE can detect in diverse data types, including bulk, single-sample expression data. By applying to real datasets, successfully identified predisease stages before onset. Our findings underscore NIEE's potential enhance comprehension development.

Язык: Английский

---

Heterogeneous network-based algorithms in the biomedical data mining: A review from technical perspective

Informatics and Health, Год журнала: 2024, Номер 1(2), С. 111 - 122

Опубликована: Авг. 24, 2024

Язык: Английский

---

Unified knowledge-driven network inference from omics data

Опубликована: Окт. 29, 2024

Abstract Analysing omics data requires computational methods to effectively handle its complexity and derive meaningful hypotheses about molecular mechanisms. While data-driven statistical machine learning can identify patterns from across multiple samples, they typically require a large number of samples often lack interpretability alignment with existing biological knowledge. In contrast, knowledge-based network integrate prior knowledge provide results that are biologically interpretable, but both unified mathematical framework, leading ad-hoc solutions specific particular types or knowledge, limiting their generalisability, common modelling interface for programmatic manipulation, restricting method extensions. Furthermore, generally cannot perform joint inference conditions, which restricts capacity capture shared mechanisms, making these more sensitive noise prone overfitting. To address limitations, we introduce CORNETO (Constrained Optimisation the Recovery NETworks Omics), framework knowledge-driven inference. redefines task as constrained optimisation problem penalty induces structured sparsity, allowing simultaneous samples. The is highly flexible supports wide variety networks—undirected, directed signed graphs, well hypergraphs—enabling generalisation improvement many methods, despite seemingly different assumptions. We demonstrate utility by presenting novel extensions signalling, metabolism protein-protein interactions. show how new improve performance traditional techniques on diverse set tasks using simulated real data. available an open-source Python package ( github.com/saezlab/corneto ), facilitating researchers in extending, reusing, harmonising

Язык: Английский

---