Search Results (2,199)

Autoimmune diseases arise from complex interactions between genetic susceptibility, immune regulation, and tissue-specific inflammatory processes, yet most risk variants identified by genome-wide association studies occur in non-coding regions with poorly defined biological functions. This review addresses the challenge of interpreting non-coding regulatory variants in autoimmunity by synthesizing emerging analytical frameworks that integrate functional genomics, single-cell profiling, spatial transcriptomics, and multi-omics data. We describe stepwise strategies that refine statistical associations through regulatory annotation, immune cell–state resolution, and perturbational evidence, highlighting complementary approaches such as massively parallel reporter assays, transcriptome-wide association studies, and single-cell expression quantitative trait locus mapping. These methods demonstrate that many autoimmune risk variants exert context-dependent effects that emerge only in specific immune cell states, activation trajectories, or tissue microenvironments. Advances in spatial and chromatin-informed technologies further clarify how regulatory variation shapes immune circuits in diseases such as systemic lupus erythematosus and rheumatoid arthritis. Finally, we discuss how machine learning-enabled multi-omics integration supports molecular endotyping and therapeutic inference while emphasizing interpretability and reproducibility. Collectively, this review highlights a shift from static variant annotation toward dynamic, context-aware analytical frameworks that enable mechanism-informed interpretation of genetic risk in autoimmune disease. Full article

Background: Right-sided colon cancer (RCC) is clinically aggressive and prone to liver metastasis, yet the cellular basis underlying its metastatic potential remains unclear. This study aimed to delineate the single-cell landscape of primary RCC tumors with and without liver metastasis. Methods: Public single-cell RNA sequencing datasets of primary right-sided colon tumors from eight patients (five with liver metastasis and three without metastasis) were integrated and analyzed. Malignant cells were identified by copy number variation inference. Tumor subclusters, differential gene expression, pathway enrichment, metabolic activity, and pseudotime trajectories were systematically compared between RCC with liver metastasis (RCC_LM) and without metastasis (RCC_noM). Results: RCC_LM tumors exhibited higher genomic instability and a significantly higher proportion of cells in G1 phase, suggesting that altered cell cycle progression is a key feature of tumors with metastatic potential. Five tumor subclusters were identified, with stem-like tumor cells significantly enriched in RCC_LM, whereas enterocyte-like cells predominated in RCC_noM. The primary tumor samples from tumors that metastasized displayed transcriptional programs indicative of epithelial–mesenchymal transition, extracellular matrix remodeling, inflammatory signaling, and metabolic reprogramming involving glycolysis and oxidative phosphorylation. Trajectory analyses indicated that RCC_LM tumors were enriched in early pseudotime states, suggesting increased cellular plasticity. Conclusions: These findings indicate that liver metastatic potential in RCC is marked by stem-like tumor states, metabolic plasticity, and microenvironmental remodeling, providing insight into the cellular mechanisms underlying RCC progression. Full article

Ferroptosis is an iron-dependent programmed cell death (PCD) implicated in cancer therapy response, yet its transcriptional control remains unevenly characterized and often centered on a limited subset of transcription factors (TFs) rather than systematically addressing TF families. The Activating enhancer-binding Protein-2 (AP-2) family of TFs is a plausible but understudied regulatory node linking oncogenic programs to ferroptosis, with prior research limited to AP-2α and AP-2γ, suggesting anti-ferroptotic and pro-tumorigenic roles. Thus, the present study aimed to provide a family-wide analysis of the relationships between AP-2 and ferroptosis across tumors in which this PCD type is considered biologically and clinically relevant. The research integrates ferroptosis gene modules with AP-2 targetomes, tumor–normal expression comparisons, survival stratification, ferroptosis scoring, cross-cohort functional analyses, and signaling pathway projection extending canonical ferroptosis circuits with AP-2–associated non-canonical elements. Consistent associations between AP-2 expression, prognosis, and ferroptosis score were observed in five tumor cohorts: cervical squamous cell carcinoma, glioblastoma, ovarian serous cystadenocarcinoma, pancreatic adenocarcinoma, and thyroid carcinoma. In addition, cross-cohort clustering highlighted genes enriched in redox- and lipid-metabolism programs linked to apoptosis and autophagy-dependent death. Among the candidates emerging from these analyses, ferroptotic markers (LOX, PTGS2, and NQO1) and AP-2–linked nodes such as CD36, DUOX1, EPHA2, MUC1, PTPRC, SNAI2, and TP63 warrant targeted functional and binding validation to infer whether these associations reflect direct AP-2 regulatory mechanisms. Most importantly, AP-2–centered research appears to be a valuable area for guiding studies of tumor-specific ferroptosis vulnerability or resistance. Full article

Active inference provides a unified framework for perception and action as processes of minimizing prediction error given a generative model of the environment. Whilst standard formulations assume continuous inference and control, empirical evidence indicates that humans update their control strategies intermittently, which reduces computational demands and mitigates propagation of correlated noise in closed feedback loops. To address this, we introduce Intermittent Active Inference (IAIF), a novel variant in which sensing, inference, planning, or acting can occur intermittently. This paper investigates intermittent planning, where IAIF agents follow their current plan and only re-plan when the prediction error exceeds a predefined threshold or the Expected Free Energy associated with the current plan surpasses prior estimates. We evaluate intermittent planning in a mouse pointing task, comparing against continuous planning while examining the impact of different threshold parameters on performance and efficiency. The findings indicate that IAIF reduces computation time whilst maintaining task performance, particularly when the number of plans sampled during planning is increased. In case of the proposed trigger based on Expected Free Energy, no additional calibration is required for this. The straightforward integration of IAIF makes it valuable in practical modelling workflows. Full article

Background: Myocardial ischemia–reperfusion (IR) injury is associated with dysregulated Ca²⁺ handling and oxidative stress, particularly in the middle-aged heart. Sarcoplasmic reticulum (SR)–mitochondria communication via mitochondria-associated membranes (MAMs) is essential for coordinating Ca²⁺ transfer and redox signaling; however, its role in IR injury in the middle-aged myocardium remains incompletely understood. This study investigated changes in cardiac MAM protein composition and associated functional and oxidative parameters during ischemia and IR. Methods: Middle-aged rat hearts were subjected to global ischemia or IR using the Langendorff perfusion model. Mitochondrial, MAM, and homogenate fractions were analyzed using biochemical, proteomic, and functional assays to assess Ca²⁺-handling proteins, redox enzymes, lipid peroxidation markers, and mitochondrial antioxidant defenses. Results: Myocardial ischemia and IR disrupted SR–mitochondria communication in middle-aged hearts, leading to impaired Ca²⁺ handling, redox imbalance, and reduced contractile recovery. Ischemia induced significant MAM remodeling, characterized by reduced mitofusin 2 levels and increased enrichment of voltage-dependent anion channel 1. These changes were associated with disturbed mitochondrial Ca²⁺ signaling, impaired SR Ca²⁺ sequestration. Although mitochondrial antioxidant defenses, including MnSOD, were largely preserved, IR was associated with compartment-specific redox alterations within MAMs, as inferred from altered redox enzyme activity and enhanced lipid peroxidation. Conclusions: Disruption of SR–mitochondria coupling and MAM-associated redox regulation represents a key mechanism underlying increased vulnerability to IR injury in the middle-aged heart. Targeting MAM integrity and modulating Ca²⁺-redox cross-talk may improve cardiac resilience in elderly populations. Full article

Satellite glial cells (SGCs) in sensory ganglia are increasingly recognized as active regulators of neuronal excitability and inflammatory signaling involved in pain conditions. In craniofacial and orofacial pain, trigeminal SGCs exhibit stimulus-dependent responses that develop over time and contribute to disease-related plasticity. Additionally, advances in experimental modeling, computational analysis, and data integration have fueled interest in “digital twins” as tools for hypothesis generation and decision support in biomedicine. However, most current biomedical applications are loosely defined and rarely explicitly address glial biology. Here, we propose a digital twin-inspired framework focused on trigeminal satellite glial cells to combine stimulus-response experiments with computational state modeling. Instead of claiming a fully developed digital twin, we describe a hybrid experimental–computational approach where glial activation states are inferred from measurable outputs, iteratively refined, and used to explore what-if scenarios related to pain mechanisms and treatments. These scenarios are intended to guide experimental design and hypothesis prioritization rather than to generate clinical predictions. We detail how this framework could enhance understanding of underlying mechanisms, prioritize potential interventions, and align with New Approach Methodologies (NAMs) and the 3Rs by reducing exploratory animal use. We also discuss key limitations, including biological simplification, uncertainty, and translational challenges. By viewing glial systems as dynamic, updateable entities rather than static readouts, this approach offers a practical and ethically grounded pathway toward more integrated research on craniofacial pain. Full article

Background: Psychosocial well-being has been increasingly recognized as a relevant factor in cardiometabolic health; however, evidence linking Purpose in Life with type 2 diabetes risk across validated prediction tools remains limited. This study examined the association between Purpose in Life and estimated diabetes risk using three established risk scores. Methods: A cross-sectional analysis was performed in 93,077 Spanish working adults aged 18–69 years participating in routine occupational health assessments. Purpose in Life was measured with the 10-item Purpose in Life scale and categorized into high, moderate, and low levels. Estimated type 2 diabetes risk was evaluated using QDScore, FINDRISC, and CANRISK. Multivariable logistic regression models were applied to calculate odds ratios (ORs) and 95% confidence intervals (CIs), adjusting for age, sex, occupational social class, smoking status, dietary pattern, physical activity, and body mass index. Results: Lower levels of Purpose in Life were consistently associated with greater likelihood of high estimated diabetes risk across all three instruments. Compared with participants reporting high Purpose in Life, those with low Purpose in Life showed increased odds of high-risk classification for QDScore (OR 2.38; 95% CI 2.19–2.57), FINDRISC (OR 2.49; 95% CI 2.08–2.89), and CANRISK (OR 2.79; 95% CI 2.50–3.09). Clear dose–response patterns were observed across Purpose in Life categories, and associations were similar in men and women as well as across lifestyle strata. Conclusions: Reduced Purpose in Life is strongly associated with higher estimated type 2 diabetes risk across multiple validated screening tools. Although causal direction cannot be inferred from this cross-sectional design, these findings suggest that psychosocial dimensions may provide complementary information for cardiometabolic risk assessment and prevention strategies. Full article

Background/Objectives: The primary outcome of this 8-week randomized, controlled, parallel trial was to assess longitudinal shifts in gut microbiota structure and predicted metabolic potential in 45 elite football players following protein supplementation. Methods: Participants combined resistance training with daily intake (30 g) of whey protein concentrate (WPC), pea protein isolate (PPI), rice protein isolate (RPI), or a plant-protein blend (MIX). For the acquisition of prokaryotic metataxonomic data, the V3–V8 region of the 16S rRNA gene was sequenced using Oxford Nanopore Technology (ONT). Functional potential was inferred through the MACADAM database and STAMP software. Strict dietary monitoring and gravimetric adherence checks were performed to isolate the intervention effect. Results: While microbial alpha-diversity indices (Chao1, Shannon, Simpson) remained stable across all groups, significant source-specific shifts in taxonomic structure and predicted metabolic activity were identified. Whey protein concentrate (WPC) was associated with an increase in Bacteroidetes abundance and greater balance within the microbial community structure, whereas pea protein isolate (PPI) and the MIX correlated with reduced fermentative bacteria and elevated taxa potentially involved in cadaverine biosynthesis. Rice protein isolate (RPI) supplementation was associated with a higher predicted representation of taxa involved in succinate-to-butyrate fermentation pathways. These functional markers and differential responses of selected bacterial groups to particular protein types were observed. Conclusions: The data indicate complex interactions between supplement type, exposure duration, and microbiome response, underscoring the necessity for individualized dietary recommendations and supplementation strategies to optimize gut health and training adaptation in professional football players. Full article

Cyclin-dependent kinase 1 (CDK1) regulates multiple cellular processes that HSV-1 can exploit to promote its own replication, particularly during the early steps of lytic infection. We investigated whether CDK1 inhibition disrupts immediate-early (IE) gene expression and analyzed the host phosphoproteome early in infection to identify putative host factors and mechanisms that facilitate HSV-1 IE gene expression and are controlled by CDK1. Human foreskin fibroblasts (HFFs) were pre-treated with a CDK1 inhibitor and showed a 1000-fold reduction in HSV-1 replication and significant reductions in IE mRNAs and protein levels at 4 hpi. We characterized cells after CDK1 inhibition and HSV-1 infection at 3 hpi by tandem mass spectrometry and identified >5500 phosphopetides (~2600 proteins), analyzing differential phosphorylation and protein–protein interactions. We validated CDK1 inhibition by detecting phosphorylation-specific decreases in known CDK1 substrates, as well as Robust Kinase Activity Inference. Rank- and network-based analyses of our dataset highlighted several candidate proteins, linking their CDK-directed phosphorylation to HSV-1 IE gene expression. Notably, the C-terminal domain of the large subunit of RNA polymerase II (RNAPII), POLR2A, is extensively phosphorylated, and its phosphorylation is significantly reduced upon CDK1 inhibition during viral infection. Taken together, these data support a model in which CDK1 activity maintains a transcriptionally permissive cellular state required for efficient HSV-1 IE gene expression. Our data suggest that when CDK1 is pharmacologically inhibited, key transcriptional facilitators are dysregulated, impairing viral transcription and replication. Full article

Human Activity Recognition (HAR) on wearable and IoT devices must jointly satisfy four requirements: high accuracy, the ability to recognize previously unseen activities, strict memory and latency constraints, and interpretable decisions. In this work, we address all four by introducing an explainable Kolmogorov–Arnold Network for Human Activity Recognition (TinyKAN-HAR) with a zero-shot learning (ZSL) module, designed specifically for TinyML edge devices. The proposed KAN replaces fixed activation functions by learnable one-dimensional spline operators applied after linear mixing, yielding compact yet expressive feature extractors whose internal nonlinearities can be directly visualized. On top of the KAN latent space, we learn a semantic projection and cosine-based compatibility function that align sensor features with class-level semantic embeddings, enabling both pure and generalized zero-shot recognition of unseen activities. We evaluate our method on three benchmark datasets (UCI HAR, WISDM, PAMAP2) under subject-disjoint and zero-shot splits. TinyKAN-HAR consistently achieves over 97% macro-F1 on seen classes and over 96% accuracy on unseen activities, with harmonic mean above 96% in the generalized ZSL setting, outperforming CNN, LSTM and Transformer-based ZSL baselines. For explainability, we combine gradient-based attributions, SHAP-style global relevance scores and inspection of the learned spline functions to provide sensor-level, temporal and neuron-level insights into each prediction. After 8-bit quantization and TinyML-oriented optimizations, the deployed model occupies only 145 kB of flash and 26 kB of RAM, and achieves an average inference latency of 4.1 ms (about 0.32 mJ per window) on a Cortex-M4F-class microcontroller, while preserving accuracy within 0.2% of the full-precision model. These results demonstrate that explainable, zero-shot HAR with near state-of-the-art accuracy is feasible on severely resource-constrained TinyML edge devices. Full article

The scientific study of affect has been historically characterized by a profound lack of terminological consensus, leading to a state of conceptual fragmentation that persists in psychology, neuroscience and many other fields. This ambiguity is not merely an academic concern; it has significant consequences for the development of artificial intelligence (AI) systems designed to recognize and respond to human “emotions”. In fact, it has an influence on the entire field of affective computing. The problem is obvious. Without a distinct definition of “emotion” it is difficult to train an algorithm to recognize it. The Walla Emotion Model, also known as the ESCAPE (Emotions Convey Affective Processing Effects) model, provides a potentially helpful and neurobiologically grounded framework to resolve this impasse and to improve any discourse about it, for businesses and even lawmakers aiming at healthy societies. By establishing clear, non-overlapping definitions for affective processing, feelings, and emotions, this model offers a path toward more precise research and more ethically sound affective computing including AI-driven “emotion” recognition. It introduces a concept that allows for the detection of incongruences between internal states and external signals with a very clear terminology supporting understandable communication. This is critical for identifying feigned or socially masked inner affective states, a challenge that traditional “face-reading” AI models frequently fail to address. Even tone of voice and body postures as well as gestures can be and are often voluntarily modified. Through the separation of subcortical affective processing (evaluation of valence; neural activity) from subjective experience (feeling) and external communication (emotion), the Walla model provides a helpful framework for AI-designs meant to have the capacity to infer an internal affective state from collected signals in the wild bypassing verbal self-report. This paper is purely theoretical; it does not provide any algorithm models or other distinct suggestions to train a software package. Its main purpose is the introduction of a new emotion model, particularly a new terminology that is considered helpful in order to proceed with this endeavor. It is considered important to first enable the clearest-possible form of communication about anything related to the term emotion across all disciplines dealing with it. Only then can progress be made. Full article

In contemporary 5G network environments, intrusion detection systems must balance detection accuracy with operational efficiency, as improvements in one dimension are often achieved at the expense of the other. This study addresses this trade-off by proposing a lightweight two-stage intrusion detection architecture that augments a standard decision-tree classifier with a conditional counter-inspection mechanism. At inference time, a global decision tree produces an initial classification for each traffic record, which is selectively validated by a small set of class-biased expert trees trained under controlled minority exposure. Only experts associated with the opposite class of the initial prediction are activated, and decision revision is governed by a unanimous-dissent rule, ensuring conservative and deterministic correction while avoiding over-correction. Experiments conducted on the 5G-NIDD dataset in a binary benign/malicious setting show that the proposed architecture consistently improves upon the standalone decision tree, reducing false negatives from 51 to 27 (−47.1%) and false positives from 48 to 30 (−37.5%), and achieving an F1-score of 0.99981 on a held-out test set. Ablation and paired statistical tests confirm that these gains arise from selective validation and the unanimous-dissent mechanism rather than from uniform ensembling. The complete pipeline operates in the microsecond inference regime per record, evaluates fewer models on average than flat voting strategies, and preserves full interpretability through deterministic decision paths, making it suitable for practical and resource-constrained 5G intrusion detection deployments. Full article

Skin lesion segmentation has achieved notable progress in recent years; however, accurate delineation remains challenging due to complex backgrounds, ambiguous boundaries, and low lesion-to-skin contrast. To address these issues, we propose the lesion-aware triple-path feature fusion network (LTPNet), an end-to-end framework that progressively processes features through extraction, refinement, and aggregation stages. In the extraction stage, we incorporate a general foreground–background attention to suppress background interference and accelerate model convergence. In the refinement stage, we introduce an attentive spatial modulator (ASM) to jointly exploit local structural cues and global semantic context for precise spatial modulation. We further develop a lesion-aware lite-gate attention (LALGA) module that performs local spatial feature modulation and global channel recalibration tailored to lesion characteristics. In the aggregation stage, we propose a triple-path feature fusion (TPFF) module that explicitly models feature relationships across scales via three complementary pathways: a common path (CP) for semantic consistency, a saliency path (SP) for highlighting co-activated regions, and a difference path (DP) for accentuating structural discrepancies. Extensive experiments on in-domain and cross-domain datasets show that LTPNet achieves superior segmentation accuracy with reasonable inference efficiency and model complexity, demonstrating its potential for efficient and reliable clinical decision support. Full article

Wolves (Canis lupus) have recolonized Belgium after more than a century of absence, raising concerns about interactions with livestock in densely populated regions such as Flanders. Empirical field-based documentation of wolf behavior near protected livestock in such landscapes remains limited. This study presents a short-term, descriptive camera-trap case study documenting wolf presence near a temporary sheep pasture protected by electric fencing and livestock guardian dogs (LGDs). Nineteen camera traps monitored the pasture perimeter within a military training area in northeastern Flanders over a 16-day period in September 2023. Sheep were present for 11 days and accompanied by six LGDs. Twenty-three wolf images were recorded, corresponding to eight distinct detection events. Wolves were detected shortly after fence installation and following sheep removal. Occasional close approaches and fence inspection behavior were observed, but no fence crossings or predation events occurred. Most wolf detections occurred when sheep and LGDs were absent, although wolves were also recorded near periods of human activity. Given the observational design, causal inference is not possible. The study provides baseline documentation of wolf–livestock–LGD interactions in a densely populated European landscape. Full article

Background/Objectives: Despite the publication of several systematic reviews on adolescent health literacy, comprehensive evaluations of the relationship between health literacy and health-related behaviors are still limited. This systematic review sought to synthesize and critically appraise the available evidence on associations between health literacy and health behaviors among adolescents. Methods. A systematic search of three databases (Scopus, PubMed, and PsycINFO) was conducted in accordance with PRISMA guidelines. Thirty-seven eligible cross-sectional studies were selected for qualitative synthesis. Methodological quality was evaluated using the Newcastle–Ottawa Scale adapted for cross-sectional studies. Results: The 37 included studies encompassed 71,558 adolescents (mean age range 11.0–17.0 years) and were conducted primarily in Europe (n = 22), with additional studies from the USA (n = 5), Asia (n = 8), and cross-cultural settings (n = 2). Across studies, 11 HL instruments were used (including two eHealth literacy measures), most commonly the Health Literacy for School-aged Children scale (n = 14). Physical activity (n = 22), nutrition-related indicators (n = 26), and smoking/alcohol/drug outcomes (n = 16) were assessed using heterogeneous operationalisations. Overall, higher HL was more often associated with healthier behavioral profiles, with more consistent patterns for nutrition-related outcomes. Findings for physical activity and substance use were more heterogeneous and, in some cases, varied depending on the HL measurement approach (subjective vs. objective) and the behavioral reference period. Conclusions: Current evidence indicates that higher health literacy in adolescents is generally associated with more favorable health behaviors, particularly regarding nutrition-related indicators. However, study heterogeneity and the predominance of cross-sectional designs limit comparability and causal inference. Future research should prioritize standardized measurement tools and longitudinal designs to clarify directionality and underlying mechanisms. Full article