Abstract: Not all sleep loss is the same, and failing to recognise this is the biggest barrier to advancing research in sleep and neurological diseases. This review systematically compares nine rodent sleep deprivation paradigms: gentle handling, the multiple platform method and its variants, the disk-over-water method, the Unpredictable Chronic Sleep Deprivation (UCSD) paradigm developed in our laboratory, novel object introduction, the curling prevention by water approach, automated mechanical systems, and the head-lifting method. It evaluates each for stress confound profile, sleep stage specificity, chronicity, and the neurobiological outcome domains to which it is appropriately suited. We describe the neuroimmune and neurochemical consequences of sleep loss across these models, covering hippocampal synaptic plasticity, prefrontal neurochemistry, glymphatic waste clearance, neuroinflammation, oxidative stress, hippocampal neurogenesis, and circadian clock gene regulation, and situate these findings within the translational context of Alzheimer's disease, Parkinson's disease, and neuropsychiatric comorbidities. Special attention is given to the UCSD paradigm, in which five established sleep disrupters, gentle handling, 24/0 h light/dark cycle, platform-over-water, crowded cage, and stroboscopic light, were applied in daily rotation without repetition across fourteen days. Using this paradigm, our group showed that chronic unpredictable sleep disruption, especially when combined with high-dose caffeine, causes prefrontal antioxidant depletion, serotonin loss, acetylcholinesterase upregulation, and synaptophysin reduction, confirmed through immunohistochemistry in Long-Evans rats, a neurochemical signature that aligns with early markers of neurodegeneration. We propose a disease-target-driven model selection framework, a six-priority translational research agenda, and minimum reporting standards for the field.

Abstract: Background: Patients with Kawasaki disease (KD) who develop coronary artery aneurysms (CAAs) are at increased risk of future fatal coronary events. Pharmacotherapeutic strategies to prevent coronary stenosis are still lacking. This study investigated the therapeutic effect of the angiotensin receptor blocker (ARB) losartan, on coronary artery stenosis in a murine model. Methods: Five-week-old male C57BL/6J mice were intraperitoneally injected with 1,000 μg of Lactobacillus casei cell wall extract (LCWE) (n=12) to induce coronary artery stenosis. Two weeks later, LCWE-injected mice (n=12) were divided into two groups: six received drinking water containing losartan (100 mg/L) (LCWE+ARB), while six received normal drinking water (LCWE group). A control group (n=5) received phosphate-buffered saline (PBS) instead of LCWE. Sixteen weeks after LCWE administration—corresponding to the peak of coronary artery stenosis and 14 weeks after treatment initiation—the mice were euthanized for histological evaluation of the coronary arteries. Results: Losartan treatment significantly reduced the coronary arteritis score (4.3±3.3 vs. 19.3±2.8, p=0.003). LCWE-induced neointimal formation with vascular smooth muscle cell (VSMC) proliferation and subsequent coronary artery stenosis were markedly attenuated in losartan-treated mice (25% vs. 100%, p<0.001). Moreover, losartan inhibited coronary artery stenosis, at least in part, by preventing the phenotypic switch of vascular VSMCs from a contractile to a synthetic phenotype. Conclusions: Losartan is a potential therapeutic agent for preventing coronary events in KD by suppressing intimal proliferation and modulating the VSMC phenotype.

Abstract: Background: Adults with congenital heart disease (CHD) are at markedly increased risk of infective endocarditis (IE); however, data comparing clinical characteristics and outcomes in sur-gically treated IE patients with and without CHD remain limited. This study aimed to evaluate differences in clinical profile, microbiology, complications, and outcomes be-tween these groups. Methods: We conducted a retrospective cohort study of 773 adult patients who underwent surgery for IE at a tertiary center in China between October 2013 and August 2025. Patients were categorized into CHD (n = 188) and non-CHD (n = 585) groups. Baseline characteristics, microbiological findings, operative data, and postoperative outcomes were compared. Inverse probability of treatment weighting (IPTW) was applied to adjust for baseline differences. Long-term survival was assessed using Kaplan–Meier analysis. Results: Patients with CHD were significantly younger and had fewer cardiovascular comorbid-ities than non-CHD patients. CHD was associated with a higher prevalence of right-sided and multivalvular infection, whereas non-CHD patients predominantly had left-sided disease. Streptococcus species were the most common pathogens in both groups, with no significant intergroup differences in microbiological profiles. After IPTW adjustment, no significant differences were observed in major postoperative complications, length of stay, or early mortality. Overall and in left-sided IE, long-term survival was comparable between groups, whereas in right-sided IE, patients with CHD exhibited significantly better long-term survival (HR = 0.17, 95% CI: 0.04–0.66, P = 0.01). Conclusions: Despite distinct clinical characteristics, adults with and without CHD undergoing surgery for IE had similar overall outcomes, although CHD was associated with better long-term survival in right-sided IE.

Abstract: Conventional tunnelling, CT, in densely populated urban areas requires ensuring that both serviceability and failure limit states are satisfied throughout construction, to avoid excessive ground deformations and catastrophic failures, while maintaining optimum support and excavation lengths. To achieve both goals, continuous monitoring of the tunnel-ground system behavior and adaptation of the tunneling process through calibrated numerical models becomes critical. This paper documents the field performance and construction optimization of a 4.5 km long tunnel excavated by CT in the stiff soils of the northwestern Mexico City region. A five-step monitoring and back-analyses procedure is introduced for risk reduction and optimization of the tunneling process during CT. The monitoring information, adopted support types, and excavation lengths demonstrate that construction times can be shortened through the proposed approach, enhancing construction processes with the corresponding cost reduction. Both three-dimensional numerical models and geotechnical instrumentation, including convergences, surface topographic references, extensometers, and pressure cells, were implemented throughout construction. The numerical models were continuously calibrated against field measurements to increase their predictive capability, accounting for actual subsoil conditions, tunnel geometries, and construction procedures. From the results gathered here, the benefits of using this integral approach to ensure good tunnel performance during excavation are established, in particular when the tunnel is excavated below densely populated areas in brittle cemented fine-grained soils

Abstract: Background/Objectives: Breast cancer represents a major public health and economic challenge, generating substantial costs for healthcare systems, patients, and the broader economy. In Poland, comprehensive assessments capturing the full societal burden remain limited. This study aimed to estimate the cost of illness of breast cancer in Poland in 2024 from a societal perspective, including direct and indirect costs, and to assess their distribution across cost bearers. Methods: A cost-of-illness analysis was conducted using a societal perspective. Data were derived from administrative sources, including the National Health Fund (NFZ), Social Insurance Institution (ZUS), Central Statistical Office (GUS), and National Cancer Registry (NCR), supplemented with a patient survey (n = 289). Direct medical, direct non-medical, and indirect costs were estimated. Productivity losses were valued using a human capital approach with a GDP-based productivity metric adjusted by a correction factor. Results: Total costs of breast cancer in Poland in 2024 were dominated by indirect costs, which accounted for approximately 73% of the total burden. Direct costs represented 27% and included €837.7 million in public healthcare expenditures and €322.4 million in patient-borne costs. Among indirect costs, absenteeism, presenteeism, unpaid work, and informal caregiving contributed substantially, with productivity losses exceeding several hundred million euros in each category. The largest single indirect component was absenteeism, followed by presenteeism and unpaid work. Conclusions: Breast cancer imposes a substantial societal burden in Poland, driven predominantly by indirect costs. These findings highlight the importance of adopting a societal perspective in economic evaluations and support the inclusion of productivity losses in health policy decision-making.

Abstract: Neonatal calf diarrhea (NCD) is one of the most important problems of calf breeding across the world. It causes deaths in calves in the first 10 days of their life and it is mainly caused by E. coli, Bovine Rotavirus (BRV) and Bovine Coronavirus (BCoV). Ab-sence of an effective vaccine targeting the main causes of NCD makes disease control highly challenging. The current study aims to design multi-epitope mRNA based vac-cine targeting the major pathogens responsible for NCD using Immunoinformatic tools and molecular modelling approaches. BRV capsid protein VP6, BCoV Spike glycopro-tein and E. coli F5 fimbrial protein were used as antigenic proteins to predict potential epitopes. Fifteen selected epitopes were linked with suitable linkers and conjugated with build adjuvant, resulting in designing of stable, antigenic and non-allergenic vac-cine candidate against NCD pathogens. Furthermore, Molecular docking analysis shows strong binding affinity between the vaccine candidate and bovine Toll-like re-ceptors TLR2 and TLR4 at low energy and high stability. Based on these findings, the proposed multi-epitope vaccine represents a promising approach for prevention and control of neonatal calf diarrhea and provides a solid scientific foundation for future experimental studies to validate its efficacy and safety in vivo.

Abstract: This study investigates the energy-efficient mechanochemical activation of fly ash derived from Kazakh coals for the development of sustainable cementitious composites. The ap-proach aims to enhance the reactivity of aluminosilicate materials while reducing the en-ergy demand and carbon footprint associated with conventional clinker-based cement production. Mechanochemical activation was performed to increase the specific surface area and in-duce structural defects in the glassy phase of fly ash, thereby improving its reactivity. Chemical activation using sodium hydroxide (NaOH) was applied to promote intensive pozzolanic reactions and accelerate dissolution kinetics. The optimal activation conditions were identified as 15 min of mechanical treatment com-bined with 4% NaOH. Under these conditions, the compressive strength reached 35.5 MPa at 28 days, exceeding that of the reference cement (35.0 MPa). At fly ash contents of 15–20%, the composites maintained or improved strength, whereas an increase to 30% resulted in a reduction to 31.5 MPa. Mechanical activation increased the specific surface area to approximately 4800–5000 cm²/g; however, prolonged grinding (up to 30 min) led to particle agglomeration and a de-crease in strength to about 28 MPa. Chemical activation enhanced reaction kinetics without significantly affecting particle fineness. Microstructural analysis revealed the formation of a dense and homogeneous matrix dom-inated by C–S–H, C–A–S–H, and N–A–S–H gel phases with reduced porosity. The com-bined activation approach demonstrated a clear synergistic effect, enabling up to 20% ce-ment replacement without loss of performance. Importantly, the proposed method provides a low-energy pathway for the utilization of industrial waste, contributing to reduced clinker consumption and lower CO₂ emissions. The results highlight the significant potential of Kazakhstan’s industrial by-products for the production of energy-efficient, environmentally friendly, and cost-effective construction materials.

Abstract: Physics makes two questionable assumptions: (1) Distant galaxies are accelerating relative to Earth. (2) Entangled objects are spatially separated from each other. Why questionable? Acceleration relative to Earth has never been observed in a single galaxy. Observers perceive entangled objects as spatially separated, yet 3D space is relative. We show that physical realities are projections of a mathematical background reality: 4D Euclidean space (ES). In Euclidean relativity (ER), all objects move through ES at the speed C. There is no time coordinate in ES. All action is due to a monotonically increasing, absolute, external evolution parameter θ. An observer experiences two projections of ES as space and time. The axis of his current 4D motion is his proper time τ. Three orthogonal axes form his 3D space x₁, x₂, x₃. His physical reality is his spacetime x₁(ϑ), x₂(ϑ), x₃(ϑ), τ(ϑ), where τ is a natural time coordinate and θ converts to absolute parameter time ϑ. Without gravity, his spacetime is Minkowski-like. As in general relativity (GR), gravity in ER is the curvature of spacetime. Since coordinates in GR are merely labels, the Einstein field equations also hold in systems that use τ as the time coordinate. ER predicts time’s arrow, relativistic effects, galactic motion, the Hubble tension, and entanglement. Remarkably, ER manages without cosmic inflation, expanding space, dark energy, and non-locality. ER tells us: (1) Distant galaxies maintain their recession speeds. (2) From their perspective, entangled objects have never been spatially separated, yet their proper time flows in opposite 4D directions.

Abstract: The COVID-19 pandemic has disrupted the lives of the world's population, resulting in over 7 million deaths. It was immediately noted that obese and/or diabetic subjects and frail elderly individuals with multiple comorbidities were more likely to have a more severe disease course. The cause of the increased morbidity and mortality in obese and/or diabetic subjects was found to be related to the presence of insulin resistance in these individuals. Furthermore, it was also discovered that COVID-19, particularly in its more severe forms, was capable of causing de novo type 1 and type 2 diabetes as well as worsening the disease course, if already present. This review aims to highlight the most accredited possible mechanisms by which subjects with insulin resistance may have a more severe disease course and those by which SARS-CoV-2 infection may cause new onset of diabetes or worsening of existing diabetes. To write this manuscript, the authors independently reviewed and compared the results of peer-reviewed and impacted journal publications, written in English, selected from the most well-known search platforms such as PubMed, Scopus, Science Direct, Google Scholar, and ResearchGate, using the following keywords: SARS-CoV-2, COVID-19, Insulin resistance, Glucose metabolism, Obesity, Diabetes, Hospitalization, Mortality.

Abstract:

The increasing global frequency of harmful cyanobacterial blooms (CyanoHABs), driven by nutrient enrichment and climate change, poses a severe threat to aquatic ecosystems and public health. This study evaluates the effectiveness of novel clay-polymer nanocomposites—combining the charge-neutralizing capabilities of polydiallyldimethylammonium chloride (PolyDADMAC) with the high density of clay minerals (kaolinite and sepiolite) for the rapid removal of toxic cyanobacteria from water. Laboratory-scale experiments were conducted using Microcystis aeruginosa, Aphanizomenon ovalisporum, and Chlorella sp., with treatment doses determined by particle charge detector (PCD) measurements to identify the "nominal dose" required for full charge neutralization. Results demonstrate that clay-polymer nanocomposites achieve over 95% removal of turbidity and chlorophyll in M. aeruginosa at doses significantly lower (15–20%) than the calculated nominal dose, likely due to specific physical bridging interactions with the cyanobacteria’s external exopolysaccharide fibers. In contrast, A. ovalisporum and Chlorella sp. required doses closer to full charge neutralization for optimal removal. Among the materials tested, kaolinite-based nanocomposites (DKG24) showed slightly superior and more stable performance than sepiolite-based versions. Notably, application at or above the nominal dose was associated with increased soluble microcystin levels, suggesting that excessive polymer concentrations may compromise cell integrity and lead to toxin leakage. These findings suggest that engineered nanocomposites offer highly efficient, scalable technology for CyanoHAB management, provided that operational doses are carefully optimized to maximize biomass removal while minimizing toxin release.

Abstract: Overhead sports place high demands on the shoulder complex, making warm-up specificity relevant for acute readiness. This randomized controlled pilot trial compared the immediate effects of a shoulder-specific warm-up with a habitual routine in 24 youth competitive overhead athletes (14–20 years), allocated to an experimental group (EG = 12) and a standard warm-up group (SWG = 12). Outcome measures were collected before and immediately after warm-up and included shoulder flexion range of motion (ROM), handgrip strength, Closed Kinetic Chain Upper Extremity Stability (CKCUES) performance, and post-warm-up Rating of Perceived Exertion (RPE; Borg CR-10). A significant group-by-time interaction was found for right shoulder flexion ROM (p = 0.003, η²p = 0.346), with a significant increase in the EG from baseline to post-test (p = 0.008). No significant effects were observed for left shoulder flexion ROM, handgrip strength, or CKCUES performance. Post-warm-up RPE was significantly higher in the EG than in the SWG (p = 0.041). These preliminary findings support the practical value of more targeted warm-up strategies in overhead sports, while larger longitudinal studies are needed to confirm their broader functional relevance.

Abstract: The discrepancy between galactic rotation curves and visible baryonic mass persists despite empirical scaling relations like the Radial Acceleration Relation (RAR) and Baryonic Tully-Fisher Relation (BTFR). We explore a phenomenological framework where this discrepancy arises from the geometric misinterpretation of observables. Inspired by Painlevé-Gullstrand coordinates, we model the vacuum as a radially infalling compliant medium that induces an apparent compression of radial coordinates for distant observers, the "Mezzi effect". Assuming Newtonian dynamics govern an undistorted "true frame", we developed a discrete shell reconstruction method parameterized by a single universal compliance constant, tested against photometric and kinematic data from 175 late type galaxies in the SPARC database. This single parameter model yields universal scaling relations of Σtrue/Σobs∝(Rtrue/Robs)−0.5and Mobs/M.And reproduces observed rotation curves (RMS residual ∼ 34km/s). The geometric projection recovers the empirical RAR and shifts the BTFR slope from ∼ 2.8in the true frame to ∼ 3.7in the observer frame, and eliminating the normalization offset. Furthermore, The Mezzi scale factor ζ governs mass and lensing corrections via distinct power laws: Mtrue/Mtrue and αtrue/αobs∝ζ−1.26 , revealing that geometric scaling affects dynamical mass more strongly than lensing mass. These results indicate that geometric projection effects may offer a viable phenomenological explanation for galactic dynamics while remaining consistent with both Newtonian gravity and weak field general relativity. For reproducibility, the code used for this analysis is publicly available at https://github.com/Brahim-Benaissa/Zeta

Abstract: Gas hydrates are ice-like compounds formed from water and methane under high-pressure, low-temperature conditions in marine sediments. They influence sediment stability, fluid flow, and hydrocarbon distribution in continental margin settings. This study employs advanced seismic attribute analysis to investigate the gas hydrate stability zone (GHSZ) in the Gulf of Mexico and to assess the relationship between hydrate presence, subsurface fluid flow, and sediment deformation.Seismic attributes, including coherence, amplitude, and spectral decomposition, were applied to 3D seismic reflection datasets covering structurally complex regions of the northern Gulf of Mexico. These attributes were used to map bottom-simulating reflectors (BSRs), gas chimneys, and fault/fracture systems. Results indicate that gas hydrate stability zones are strongly associated with structural highs, fault intersections, and areas of enhanced deformation.The study finds that fault-controlled fluid pathways significantly influence hydrate distribution and sediment deformation patterns, highlighting the need to integrate seismic attribute analysis in hydrate resource assessment and geohazard evaluation. These findings provide new insights into fluid migration mechanisms and sediment dynamics in hydrate-bearing marine environments.

Abstract: The composition of partially ionised plasmas is investigated for densities and temperatures at which the free electrons are degenerate. Based on a quantum statistical approach, the effect of Pauli blocking is addressed. Specifically, one- and two-electron ions are studied. New results regarding the degree of ionisation and the Mott effect are presented. Standard codes for plasma properties do not take Pauli blocking effects into account and are therefore unable to explain the experiments in the high-density regime, where the electrons are degenerate.

Abstract: Extracting Alpha in extreme low signal-to-noise ratio (SNR) environments, such as the Chinese A share market, remains a notoriously unsolved challenge for deep learning. Traditional heavily parameterized models, including Transformers, inevitably fall into the ”dimensionality disaster,” memorizing market noise rather than fundamental mechanics. To break this overfitting curse, we propose a novel, ultra-lightweight architecture inspired by thermody namics and Neural Turing Machines (NTMs): the Physics-Informed Ghost Operator. By mapping the cross-sectional stock market into a high-dimensional physical manifold, our Ghost Operators navigate the feature space driven by gravitational routing. Crucially, we enforce a minimal action principle via a Boltzmann-distributed temperature scaling and Pauli-exclusion-like potential well clamping. Walk forward validation on 10 years of real A-share data reveals that our architecture achieves a substantial Sharpe ratio improvement (up to 3.2x) and cuts the maximum drawdown by nearly half compared to native NTMs. Furthermore, network sparsity is reduced by 66%, proving that physical constraints compel the model to aggressively filter noise and focus strictly on high-potential Alpha regions.

Abstract: Low-light video enhancement aims to restore clear, color-faithful, and temporally consistent visual content from video sequences captured under extremely low signal-to-noise ratios and high dynamic range constraints. Existing multi-frame enhancement methods typically adopt uniform spatio-temporal sampling and feature extraction strategies for all frames, making it challenging to simultaneously achieve long-range temporal denoising and accurate fast-motion modeling. To address this trade-off, we propose a low-light video enhancement framework based on a Fast–Slow dual-branch architecture. The video signal is decomposed into two complementary feature streams: a Slow branch with sparse temporal sampling and high spatial resolution, built on a Vision Transformer backbone, which focuses on long-range temporal denoising and high-frequency texture restoration for static and slow-moving regions; and a Fast branch with dense temporal sampling and low spatial resolution, built on a ViT-Tiny backbone, which efficiently captures large-scale motion and rapid illumination changes. To mitigate the discrepancy in sampling rates and spatial resolutions between the two branches, we further introduce a flow branch based on a pre-trained StreamFlow model and design a Flow-Guided Cross-Attention (FGCA) module. FGCA first uses optical flow to geometrically modulate and progressively align Fast-branch features, and then injects the flow-enhanced Fast features into the Slow branch at each space-time location via lightweight pixel-wise cross-attention. This mechanism achieves a cascade of coarse geometric alignment and fine semantic fusion. Experiments on two real-world low-light video datasets, SDSD-indoor and SDSD-outdoor, demonstrate that our method consistently outperforms several representative approaches in terms of PSNR, SSIM, AB(Var), and MABD, while effectively suppressing motion blur and ghosting artifacts in dynamic night scenes, yielding temporally stable and perceptually pleasing results.

Abstract: The optical counterpart of the gravitational wave event GW170817, known as kilonova, has provided strong evidence that binary neutron star mergers are favourable sites to host the r-process nucleosynthesis. Kilonova is a quasi-thermal electromagnetic emission powered by the radioactive decay of heavy neutron-rich nuclei produced by the r-process. Considering the variety of elements contributing to kilonova ejecta, essential information about its composition can be achieved through spectral characterisation, radiative transfer simulations, and opacities. The latter represents one of the most challenging aspects of the modelling, as it relies on accurate atomic structure calculations of energy levels and transitions. Since light r-process elements are major opacity contributors in early (< 2 days) scenario, this work focuses on atomic calculations for Zr I-IV. Energy levels and bound-bound transitions are determined using the GRASP2018 code, assuming two different multi-reference sets for each ionisation stage: one including, and one excluding core-core and core-valence correlations. Results demonstrate that the inclusion of f shell and core correlations impacts on both energy levels and transitions. A systematic assessment of the accuracy is performed through detailed comparisons with the NIST ASD. Finally, these Zr data are integrated on the open access MARTINI platform.

Abstract: Background/Objectives: Large language models (LLMs) are increasingly applied to medical image interpretation; however, their diagnostic accuracy and reliability in musculoskeletal radiology remain uncertain. This study evaluates the diagnostic performance and confidence calibration of LLMs in detecting and classifying bone tumors on radiographs. Methods: This retrospective observational study analyzed a dataset of 257 radiographs with confirmed diagnoses obtained from Radiopaedia, including normal studies and a spectrum of benign and malignant bone tumors. Cases were selected to ensure representation across multiple tumor types. Three LLMs (ChatGPT 5.3, X-ray Interpreter GPT-4.1, and X-ray Interpreter Gemini) evaluated each image using a standardized prompt assessing abnormality detection, tumor detection, classification, and confidence. Outcomes included diagnostic accuracy, false positive abnormality rates, false negative rates, tumor hallucination rates, and confidence calibration. Results: Abnormality detection was high across models, with Gemini demonstrating the highest sensitivity (up to 100%). Tumor detection was strongest in lesions with characteristic features, including osteosarcoma and osteochondroma. False negative rates varied substantially, with GPT-4.1 demonstrating the highest rate (29.9%), followed by ChatGPT (24.8%) and Gemini (6.6%). Primary diagnostic accuracy was highest for osteosarcoma in GPT-4.1 (80%), while ChatGPT 5.3 performed best in benign lesions, including osteochondroma (84.6%) and non-ossifying fibroma (76.9%). Tumor subtype classification remained limited across all models and was poorest for Ewing sarcoma (0% in ChatGPT and GPT-4.1; 10.3% in Gemini). False positive abnormality rates were highest in GPT-4.1 (40.7%), followed by Gemini (25.9%) and ChatGPT (13.5%). Tumor hallucination occurred only in Gemini (12.3%). All models demonstrated confidence miscalibration, with higher confidence observed in incorrect predictions and in tumor-negative cases. Conclusions: LLMs demonstrate strong performance in detecting radiographic abnormalities but remain limited in tumor subtype classification, particularly for diagnostically challenging lesions such as Ewing sarcoma. Elevated false positive and false negative rates, along with systematic overconfidence—especially in GPT-4.1—highlight important limitations for clinical use. These findings support the role of LLMs as adjunctive tools rather than independent diagnostic systems.

Abstract: Background: Genetic variations such as single nucleotide polymorphisms (SNPs) as part of pharmacogenomics play an important role in the metabolism of drug and hence their active concentrations in blood plasma. Objectives: The aim of this study is to select candidate compounds from a TCM dataset that may be repurposed for arterial and venous thromboses management. This shall be achieved through development and evaluation of an ensemble deep learning model that taking into account the genetic variations in protein sequences. Methods: BIOSNAP dataset was supplemented with 321,657 drug–target pairs consisting of SNP variants of wild-type proteins. The application dataset consisted of a TCM dataset containing 35,553 ingredients. The control group was set as the pathogenic group, whilst the treatment group was set as the non-pathogenic group. Contrastive and non-contrastive deep cross-modal attention ensemble modelling was developed, evaluated and applied. Results: Contrastive regularisation effect improved the performance of the Contrastive Learning (CL) Ensemble over the Non-CL Ensemble model as well as the Dong et al. (May 2025) CL model in the test set (AUPR 0.919 vs. 0.894 vs. 0.813). Safflower yellow A, Paeoniflorin and Notoginsenoside R6 were associated with existing TCM and highly ranked for interaction with Factor Xa genetic variants. Highly interacting protein targets were identified. Conclusions: Ensemble modelling with contrastive learning resulted in performance improvements and can be useful for selecting TCM compounds for antithrombotic management. This is a step towards personalised drug selection and can simultaneously facilitate interpretation of the biological rationales for risk vs benefit evaluations during decision making.

Abstract: Background: The coexistence of Myasthenia Gravis (MG) and Rheumatoid Arthritis (RA) represents a rare but clinically challenging form of polyautoimmunity, raising interesting questions about shared immunopathogenic mechanisms and the safety of long-term immunomodulatory therapies. Methods: The article describes a case report of a 66-year-old female with a 12-year history of seropositive RA who subsequently developed seropositive MG during long-term exposure to hydroxychloroquine (HCQ) therapy. Following discontinuation of HCQ, methotrexate (MTX) therapy was initiated and stable control of both diseases was temporally obtained. Results: Three years later, the patient presented with upper gastrointestinal bleeding and severe microcytic anemia. Further evaluation revealed advanced liver fibrosis (F4) and severe gastropathy, consistent with Child–Pugh class A cirrhosis. Viral, alcoholic, and autoimmune causes of chronic liver disease were excluded. In the absence of alternative etiologies, this was considered possibly associated with MTX therapy, in the context of additional metabolic risk factors, including type 2 diabetes mellitus and increased body mass index. Conclusions: The complex interplay between polyautoimmunity and treatment-related toxicity is underscored in this article. Overlapping autoimmune diseases may arise on a shared immunological background, while therapeutic agents may contribute to disease expression or long-term complications. These findings highlight the need for individualized therapeutic strategies and vigilant monitoring, particularly in patients with coexisting metabolic risk factors.