Browse Articles
Discover research articles across all indexed journals
Median Raphe projection into the dorsal dentate gyrus modulates anxiety behavior and coping with learning under stress
CBEC inventory optimization model design based on spatiotemporal attention and transformer architecture
To solve the problem of inaccurate long-term prediction of cross-border inventory encountered by cross-border enterprises in their experience, this paper proposes a cross-border inventory prediction model based on a spatiotemporal perception Transformer. Specifically, firstly, an improved temporal aware self-attention mechanism is adopted to mine potential temporal trends and spatial heterogeneity features in cross-border inventory, and an accurate spatiotemporal correlation matrix is established to obtain global spatiotemporal features. Secondly, we simulate the multi-level diffusion process of inventory data in the road network using multi-scale diffusion convolution, which captures the local spatial features of nodes across multiple neighborhood ranges. Finally, a multi-dimensional feature fusion module is used to adaptively fuse the captured spatiotemporal features and output prediction results. The experimental results show that compared with the ASTGNN model with the highest prediction accuracy, the method proposed in this paper performs better in MAE, MAPE, and RMSE, which are reduced by 7.6%, 4.2%, and 1.1%, respectively.
Dual Janus foam for directional thermal management
A hybrid deep learning and fuzzy logic framework for robust tomato disease detection and classification
Comparative assessment of the impact of iron deficiency on HbA1c accuracy in non-anaemic individuals with type 2 diabetes: A secondary data analysis
Purpose Results from a few studies have been conflicting regarding whether iron deficiency affects HbA1c reliability, and the mechanisms by which iron might influence HbA1c are not fully understood. We aimed to compare the relationship between HbA1c and average glucose levels measured by continuous glucose monitoring, retrospectively, in iron-replete and iron-deplete states among non-anemic type 2 diabetes mellitus (T2DM) patients. Methods We compared the differences in HbA1c between iron-replete and iron-deplete groups using the Chi-square test for categorical data and the Mann-Whitney U test for continuous data. We also evaluated the correlation between HbA1c and mean plasma glucose for both iron-replete and iron-deplete individuals using Pearson’s correlation and linear regression. Results A total of 146 of the 213 participants screened had complete data and were included in the final analysis. 43 out of 146 (29.5%) had iron deficiency, and 103 were iron-replete. No significant difference was observed in HbA1c levels between iron-replete and iron-deplete individuals: 69 (51.0, 85.0) vs 62 (46.0, 83.0) mmol/mol, P = 0.291). There was a strong positive correlation between HbA1c and mean plasma glucose concentration for both iron-replete and iron-deplete individuals (Pearson Correlation coefficient: 0.88 (0.83–0.92) and 0.93 (0.88–0.98), respectively). Conclusions HbA1c correlates well with mean blood glucose even in the iron-deplete state amongst non-anaemic T2DM individuals. However, larger studies are needed to confirm these findings, particularly at screening and diagnostic thresholds.
Biomarker-integrated prognostic stagings for Alzheimer’s Disease
2D Trefftz method in identification of flow boiling heat transfer coefficient in horizontal minichannel
Large-scale multi-omic biosequence transformers for modeling protein–nucleic acid interactions
The transformer architecture has revolutionized bioinformatics and driven progress in the understanding and prediction of the properties of biomolecules. To date, most biosequence transformers have been trained on single-omic data—either proteins or nucleic acids—and have seen incredible success in downstream tasks in each domain, with particularly noteworthy breakthroughs in protein structural modeling. However, single-omic pretraining limits the ability of these models to capture cross-modal interactions. Here we present OmniBioTE, the largest open-source multi-omic model trained on over 250 billion tokens of mixed protein and nucleic acid data. We show that despite only being trained on unlabeled sequence data, OmniBioTE learns joint representations mapping genes to their corresponding protein sequences. We further demonstrate that OmniBioTE achieves state-of-the-art results predicting the change in Gibbs free energy ( Δ G ) of the binding interaction between a given nucleic acid and protein. Remarkably, we show that multi-omic biosequence transformers emergently learn useful structural information without any a priori structural training, allowing us to predict which protein residues are most involved in the protein–nucleic acid binding interaction. Compared to single-omic controls trained with identical compute, OmniBioTE also demonstrates superior performance-per-FLOP across both multi-omic and single-omic benchmarks. Together, these results highlight the power of a unified modeling approach for biological sequences and establish OmniBioTE as a foundation model for multi-omic discovery.
PRMT3-mediated post-translational adaptation to fasting regulates metabolic flexibility
Chronic early life stress alters the microbial and transcriptional profile of the zebrafish gut
Modeling HBV transmission dynamics in Indonesia (2024–2030) using a SIVRM model: Evaluating optimal control strategies for elimination by 2030
Hepatitis B remains a global health concern. Achieving the World Health Organization’s (WHO) goal of eliminating the disease by 2030 requires a comprehensive understanding of its transmission dynamics. This study aimed to develop and apply an extended SIVRM (Susceptible–Infected–Vaccinated–Recovered–Mortality) model to simulate hepatitis B transmission in Indonesia and to evaluate optimal vaccination strategies. The model comprises 14 compartments that distinguish between vertical and horizontal transmission, account for vaccination and loss of immunity, and incorporate hepatitis B virus (HBV) reactivation among recovered individuals, a novel feature of this model. Parameters were estimated using data from the Social Security Administrator for Health (BPJS Kesehatan) from 2019 to 2023 through a least-squares fitting approach. The basic reproduction number ( ℛ 0 ) and disease-free equilibrium (DFE) were analytically derived. Simulations were conducted using MATLAB 2018 to project hepatitis B trends from 2024 to 2030 and to evaluate scenarios of adult and newborn vaccination coverage. A key finding from the parameter estimation was an HBV reactivation rate of 0.30, indicating that 30% of recovered individuals remain at risk. The model estimated a baseline ℛ 0 of 4.39, indicating that current control strategies in Indonesia are insufficient to achieve the WHO elimination goal. However, scenario-based analysis revealed that increasing adult vaccination coverage to at least 59%, while maintaining newborn vaccination at 70%, could reduce ℛ 0 to 0.90 and substantially lower the disease burden. These findings underscore the urgent need to expand adult vaccination programs and strengthen post-recovery monitoring to advance hepatitis B elimination in Indonesia.
Continuous fabrication of Janus liquid crystal elastomer fibers with programmable actuation
Origin of giant dielectric permittivity and localized polaron-supported electrical conduction in CaCu3Ti4O12 for extreme environment energy storage applications
Systematic analysis of proton pump inhibitors-related adverse reactions using the FDA adverse event reporting system database
Background Proton pump inhibitors (PPIs), known for their potent acid-suppressing effects, are widely used in various clinical settings, including treatment and prevention. Understanding their adverse effects is crucial. This study, utilizing the FDA Adverse Event Reporting System (FAERS) database, comprehensively analyzes PPI-related adverse events to guide clinical medication practices. Methods This study analyzed suspected adverse drug reactions (ADRs) related to specific PPI drugs using data from the FAERS database, covering Q1 2004 to Q4 2024. Multiple statistical methods, including ROR, PRR, IC025, and EBGM, were employed for evaluation, with ADRs defined according to System Organ Class (SOC) and Preferred Term (PT). A comparative analysis was conducted to assess potential differences in ADR profiles among different PPI drugs. Results This study analyzed 176,680 cases of PPI-related adverse events, with a total of 632,468 adverse reaction reports recorded when PPIs were designated as the primary suspected drug (PS). PPIs showed significantly elevated risks in the renal/urinary and gastrointestinal systems, with other common adverse reactions including hypomagnesemia, hypocalcemia, and renal anemia. Most adverse reactions occurred either within the first 0–30 days of use or after prolonged exposure (>6 months), and elderly patients (≥65 years) were disproportionately affected. Conclusions For high-risk populations using PPIs long-term (such as elderly patients or those with pre-existing renal impairment), continuous monitoring is essential to mitigate potential complications. Unnecessary use should be strictly avoided, and long-term medication should be minimized to ensure safety and appropriateness.
A digital platform with activity tracking for energy management support in long COVID: a randomised controlled trial
Abstract In a 6-month pragmatic randomised controlled trial (RCT; ISRCTN16033549), we compared a just-in-time intervention to support energy management in adults with long COVID (LC) to standard care. Participants received either the ‘Pace Me’ app and a wearable activity tracker (intervention) or an app only with data entry screens (control). The intervention group received just-in-time messages on energy management when they reached 50%, 75%, and 100% of their daily ‘activity allowance’. The primary outcome was post-exertional malaise (PEM) measured by the DePaul Symptom Questionnaire-PEM (DSQ-PEM). Of 369 participants assessed for eligibility, 250 participants were randomised 1:1, and 77 controls and 84 intervention participants were included in the final per-protocol analysis. There was no time by group interaction for the DSQ-PEM. The intervention group value was 48 (95% CI 44-53) at baseline and 46 (95% CI 41-51) post-intervention (arbitrary units). The control group value was 47 (95% CI 42-52) at baseline and 44 (95% CI 39-49) at follow-up (interaction effect p = 0.614, η²p = 0.002; trivial). No individual question exhibited an interaction effect (p > 0.05). Although the intervention had minimal effect compared to control, the substantial recovery rates previously reported in LC, coupled with our wide inclusion criteria may have masked intervention effects. Therefore, future studies should consider this energy management framework in conditions without such recovery rates, such as CFS.
New perspectives on the ozonated high-oleic sunflower oil physico-chemical behaviour
Recycling of polyolefins using recycled PVC under sunlight
Quartz porosity in amorphous SiO2 of granitic shear bands
Abstract Important porosity is commonly observed in quartz-rich rocks that viscously deformed at depths of the metamorphic continental crust. Although the presence of such pores – often occurring with angular, pyramidal shapes – has major implications for fluid circulation, mass transfer and rock strength, whether or not they are directly produced by deformation remains unclear. Here we provide detailed observations of plastically deformed, pure quartz aggregates decorated by (sub)micrometric pores in granitic shear bands (western granite, Naxos, Greece). Using electron microscopy, we demonstrate that pores (1) decorate both grain boundaries and intra-grain substructures, (2) do not necessarily align with substructures when decorating grain boundaries, and (3) interconnect through layers of amorphous SiO 2 . Transmission electron microscopy further reveals dislocation densities one (or two) order(s) of magnitude below the predictions based on the lattice curvature gradients, suggesting the occurrence of residual stress along pore-decorated, partly amorphized substructures. Challenging long-lasting hypotheses, these features are here proposed to result from stress concentration at grain and subgrain boundaries, followed by fluid exsolution into mechanically amorphized quartz. Although involving syn-kinematic processes that remain to be fully understood, our findings corroborate recent studies that emphasise stress-induced amorphization as a general process of lithospheric rock deformation.