Browse Articles

Author Correction: Electrostatic potentials of atomic nanostructures at metal surfaces quantified by scanning quantum dot microscopy

Interpretable machine learning with Bayesian optimization for bond strength prediction of steel reinforcement in geopolymer concrete

Accurate estimation of bond strength between steel reinforcement and geopolymer concrete is essential for the reliable design of sustainable reinforced concrete structures. However, the highly nonlinear interactions reduce the applicability and accuracy of conventional empirical models. This study proposes a Bayesian-optimized interpretable machine learning framework to predict the ultimate bond strength of reinforced geopolymer concrete using a comprehensive experimental database compiled from published studies. A dataset of 238 samples with 20 influential input variables was assembled to represent material properties, geopolymer chemistry, and specimen geometry. Six advanced machine learning algorithms, including Support Vector Regression (SVR), Random Forest (RF), Extra Trees Regressor (ETR), Gradient Boosting Machine (GBM), XGBoost, and CatBoost, were developed and systematically compared. Hyperparameter tuning was performed using Bayesian optimization to improve model performance. The results indicate that all models achieved strong predictive capability, while the optimized CatBoost model (BO-CatBoost) provided the best performance with testing metrics of R² = 0.950, MAE = 1.173, MAPE = 11.608%, and RMSE = 1.669. A comparative evaluation with existing empirical equations further demonstrated the superior accuracy and lower prediction variability of the proposed model. To enhance model transparency, SHAP-based explainability analysis was conducted to quantify the contribution of each input parameter. The global importance analysis revealed that compressive strength, the embedment length-to-bar diameter ratio, and the cover-to-bar diameter ratio are the most influential factors governing bond strength. Additional mixture-related parameters, including the alkaline solution-to-binder ratio, curing temperature, CaO content in the binder, and the SiO₂/Al₂O₃ ratio, also contribute to the bond mechanism by influencing geopolymerization and matrix densification. The proposed framework provides both high predictive accuracy and interpretable insights, demonstrating the potential of Bayesian-optimized interpretable machine learning to support the design and optimization of sustainable reinforced geopolymer concrete structures.

AI can cause harm: safeguards must catch up

Neural mechanisms of time-forward predictions for naturalistic auditory tone sequences

Factors associated with women’s choice for mode of cervical sampling in future cervical cancer screening

Background Cervical cancer (CC), primarily caused by persistent high-risk human papillomavirus (HR-HPV) infection, remains a major public health challenge, with Latvia experiencing substantially higher incidence and mortality rates than the European average. HR-HPV self-sampling is an effective and acceptable screening method, yet some women continue to prefer clinician-collected sampling. Evidence on factors associated with reluctance to adopt self-sampling in the Baltic region is limited. This study examined demographic, socioeconomic, behavioral, and health-related characteristics linked to women’s preferences for future cervical cancer screening modalities. Materials and methods This cross-sectional study was conducted in Latvia between February 2021 and April 2022. Women attending general practitioners or the Colposcopy Unit of Riga East University Hospital were invited to perform unsupervised HR-HPV self-sampling using written and visual instructions, followed by a structured questionnaire. Collected data included demographic and socioeconomic characteristics, health status, health behaviours, emotional responses to self-sampling, and preferred sampling modality for the next screening round. Descriptive statistics and Chi-square tests were used to assess differences in emotional responses and preferences. Factors associated with preferred sampling modality were identified through univariate and multivariate multinomial logistic regression. Statistical significance was set at p < 0.05. Results A total of 1,217 women were included. Overall, 45.9% preferred clinician-collected cervical sampling, 42.6% preferred self-sampling, and 11.5% expressed no clear preference. Women who preferred self-sampling reported significantly lower embarrassment and discomfort during the procedure (p < 0.001). Most participants expressed positive emotional responses. The main reason for preferring clinician-collected sampling (81.0%) was a lack of confidence in performing self-sampling correctly. Multivariate analysis showed a higher preference for clinician-collection among women recruited from colposcopy settings and those with more recent gynecological visits, whereas former or occasional smokers were less likely to prefer clinician collection. Higher BMI (≥25) and non-Latvian nationality were associated with indecision regarding future sampling preference. Conclusion HR-HPV self-sampling was generally well accepted and may help engage women less likely to attend traditional screening. Preference for clinician-collection was shaped mainly by medical history and nationality. Emotional factors may also influence screening choices and should be considered in future implementation strategies.

How to widen access to the critical minerals that the world needs

Author Correction: Coupled sulfur-silicon isotopes reveal supracrustal origin of Archean continents

Effectiveness and safety of intra-articular hyaluronic acid SEMICAL GEL-B CROSS therapy in knee osteoarthritis (SEM-ART1): Study protocol for a randomized, placebo controlled, double-blind, cross-over clinical trial

Objective This study aims to evaluate the effectiveness and safety of a single intra-articular injection of high-dose cross-linked hyaluronic acid on pain, physical function, muscle strength, and quality of life in patients with knee osteoarthritis. Methods This randomized, double-blind, placebo-controlled, cross-over clinical trial includes 102 adults with symptomatic knee osteoarthritis (Kellgren–Lawrence grade II–III). Participants will be allocated using stratified randomization based on radiographic severity and laterality. Patients will receive either intra-articular hyaluronic acid (Semical Gel-B Cross, 90 mg/3 mL) or isotonic saline (3 mL of 0.9% NaCl), followed by cross-over at the 3-month evaluation. Outcomes will be assessed at baseline and at 3, 6, 9, and 12 months. The primary efficacy endpoint is the between-group difference in WOMAC pain score at the 3-month assessment prior to cross-over. Secondary outcomes include VAS pain, WOMAC physical function and stiffness scores, isometric quadriceps and hamstring strength, quality of life (SF-36), the 6-minute walk test, the 5-times sit-to-stand test, analgesic use, and adverse events. Treatment response will additionally be categorized as “absolute responders” (≥50% and ≥20-point improvement in VAS) and “responders” (≥20% and ≥10-point improvement). Sample size was calculated using G*Power (effect size 0.5, 80% power), yielding 102 participants. Statistical analyses will include both intention-to-treat and per-protocol approaches. Expected results and conclusions This study will provide 3-month efficacy data and long-term (3–12 months) safety outcomes for high-dose intra-articular hyaluronic acid compared with placebo. Findings are expected to inform clinical decision-making regarding symptom relief, functional improvement, and the safety profile of cross-linked HA formulations in knee osteoarthritis. Trial registration ClinicalTrials.gov: NCT06141018

Author Correction: Long-duration electricity storage needs for coping with Dunkelflaute events in Europe

Genetic surveillance of first- and second-line drug-resistant isolates of Mycobacterium tuberculosis in Peru

Tuberculosis (TB) remains a critical public health crisis in Peru, which bears the highest multidrug-resistant TB (MDR-TB) burden in the Americas. The objective of this study was to conduct a nationwide genetic surveillance of first- and second-line drug-resistant Mycobacterium tuberculosis isolates in Peru to characterize their structural and geographic distribution using a multilocus genotype approach. We analyzed 38,339 valid results from first- and second-line line probe assays (LPA 1L, GenoType MTBDR plus v2; and LPA 2L, GenoType MTBDR sl v2) routinely collected between 2019 and 2022 by the Peruvian National Institute of Health. First-line (G1) and second-line (G2) multilocus genotypes were generated by concatenating specific hybridization banding patterns across resistance-associated loci. Additionally, targeted next-generation sequencing was performed on a stratified proportional sample of 170 isolates exhibiting “inferred” resistance genotypes, defined by a missing wild-type signal without a corresponding mutant probe, to elucidate their underlying mutational profiles. Resistance genotypes were identified in 13.5% and 11.8% of LPA 1L and LPA 2L results, respectively. The genotypic landscape was highly concentrated: 13 dominant G1 genotypes and 19 dominant G2 genotypes accounted for over 85% of their respective resistant cohorts. Structurally, these dominant genotypes mainly carried globally successful, canonical mutations (e.g., katG S315T1, rpoB S450L). Geographically, dominant genotypes were heavily consolidated within the hyper-endemic central coastal axis (Lima, Callao, and Ica) and the Amazonian department of Ucayali, maintaining highly stable temporal frequencies. Conversely, heteroresistance and complex multi-mutational patterns were exclusively restricted to low-frequency rare and orphan genotypes. Targeted sequencing of inferred patterns successfully expanded the national catalog of resistance variants, identifying non-canonical determinants (e.g., inhA g-17t, gyrB T500N) and preventing false-positive resistance calls driven by synonymous mutations. This large-scale mapping demonstrates that Peru’s drug-resistant TB epidemic is characterized by the widespread consolidation of a restricted number of canonical multilocus genotypes. Integrating targeted sequencing with routine molecular screening is essential to resolve unclassified hybridization patterns and optimize epidemiological management in high-burden settings.

In-depth characterization of the CaV2.2-knockout mouse line

Abstract The voltage-gated calcium channel Ca V 2.2 is predominantly located in neuronal synapses, where it plays a role in pain signaling and neurotransmitter release. Furthermore, Ca V 2.2 is important during early central nervous system development, as it is subsequently replaced by other calcium channels. A number of mouse lines with disrupted Ca V 2.2 channels have been developed through the utilization of a gene-targeting vector. The Ca V 2.2-deficient mice primarily exhibited reduced anxiety-linked behavior and response to pain stimuli. In a previous study, we observed hyperactivity and seizures in Ca V 2.2-knockout mice. Given the absence of prior descriptions of these characteristics, an in-depth characterization was conducted. This characterization included longitudinal behavioral studies, as well as histological, proteomic, and metabolomic analyses. Irrespective of the frequency with which testing was conducted, a variety of behaviors were observed (e.g., hyperactivity and increased exploratory behavior) in comparison with wild-type mice. The frequency and severity of the performed tests demonstrated a positive correlation with the magnitude of the difference in molecular markers between Ca V 2.2-knockout and wild-type mice. Seizures were observed but were not common. These results suggest that the Ca V 2.2-knockout mouse lines are susceptible to repeated stress stimuli during experimental studies. The mice are characterized by hyperactivity, explorative behavior, and sometimes experience seizures.

Quantifying institutional gender inequality in contemporary visual art

A distributionally-robust bayesian adaptive EWMA chart for joint surveillance of lognormal process location and scale

Reliability surveillance of safety-critical systems often involves monitoring positively skewed characteristics, such as failure rates, repair times, or material degradation paths, which are robustly modeled by the lognormal distribution. We propose a Distributionally-Robust Bayesian Adaptive EWMA (DR-BAEWMA) framework for the joint surveillance of the log-scale mean and variance under practical model and information uncertainty. Decisively departing from standard Bayesian charts that assume a perfectly specified likelihood, our architecture integrates a distributional-robustification layer using Wasserstein ambiguity sets to protect online estimates against heavy-tailed contamination and misspecified sensor noise. Operating on the log scale, the method treats process health as a latent variable within a non-stationary state-space formulation, decoupling true signals from instrumentation noise via variance-inflation surrogates derived from distributionally robust optimization (DRO) duality results. Decision-theoretic point estimates under both symmetric squared-error loss (SELF) and asymmetric Linex loss (LLF) are incorporated to support risk-sensitive monitoring priorities integrated into the alarm thresholding. We present an adaptive sampling approach that minimizes a principled cost-delay objective to estimate the optimal inspection effort online in order to account for practical problems. Robust univariate and multivariate alarms are provided by a scalar Max statistic and a multivariate Mahalanobis intensity; control limits are obtained by nested Monte Carlo calibration to guarantee that the framework maintains its specified in-control average run length ( A R L 0 ) under ambiguity and variable information density. Extensive simulation across steady-state, polluted, and cross-distributional regimes demonstrates a significantly reduced worst-case detection time as compared to conventional Bayesian and frequentist EWMA approaches. An industrial semiconductor hard-bake case study is used to illustrate implementation, robustness diagnostics utilizing the information ratio, and the effectiveness of adaptive maintenance-on-demand bursts. For real-time monitoring in safety-critical engineering applications, the framework offers an operationally tractable and mathematically rigorous instrument.

Beyond the procedure: a game-theoretic analysis of multilevel determinants shaping doctor-patient engagement after angioplasty

Time to give hydration breaks the red card? What science says about keeping cool

Climatic reach of small-scale turbulence in the ocean interior

Horn trace element profiles of managed rhinoceros species and comparisons to matched livers and wild African rhinoceros horns

Rhino horn is composed of keratin which contains numerous minerals and metals that may provide insight into the species’ environment and health. Our goals were to identify horn trace elements that differ according to sex, species, and continent, and those that reflect liver loads. Samples from 82 horns representing (male.female): 25.16 black rhino, 13.16 white rhino, 6.3 greater one-horned (GOH) rhino, 0.1 Sumatran rhino, and 2.0 undetermined were analyzed for 14 trace elements: arsenic, barium, cadmium, chromium, cobalt, copper, iron, lead, manganese, mercury, molybdenum, selenium, thallium, and zinc. A subset of matched liver tissues (n = 21) also was analyzed for up to 12 of the elements. Horn trace element concentrations did not differ by sex ( p  ≥ 0.05), but several were impacted by species and continent. Copper concentrations in white rhino horn trended higher compared to those measured in black rhino horn ( p  = 0.051), whereas selenium and molybdenum concentrations were higher in black rhino horn compared to those in white rhino horn ( p  < 0.05), and cobalt was higher in black rhino horn compared to GOH rhino horn ( p  < 0.05). Horns from Africa (n = 7) were higher in arsenic and barium and lower in copper and zinc ( p  < 0.05) compared to those from the United States (n = 52). Both arsenic and selenium were positively correlated with matched liver tissue values; arsenic across all species (Spearman’s p  = 0.017, rho = 0.659) and selenium within black rhinos ( p  = 0.004, rho = 0.790). Rhino horn trace element concentrations differed by species, and preliminarily between continents, reflecting population trends in element status, but most were not significantly correlated to individual liver loads.

Identification of PDE10A binding ligands as potential therapeutics for the Parkinson’s disease through a computational approach

Author Correction: Mitochondrial adaptor TRAK2 activates and functionally links opposing kinesin and dynein motors

Association of systemic immune-inflammatory index with stress urinary incontinence in U.S. Adults: NHANES 2007–2016

Background The aim of this study is to explore the association between the systemic immune-inflammatory index (SII) and stress urinary incontinence (SUI). Methods We used information obtained from the National Health and Nutrition Examination Survey (NHANES) that was conducted between 2007 and 2016. The weighted multivariate logistic regression model was used in order to evaluate the relationship that exists between the SII index and the SUI measurement. In addition, a technique from the field of smooth curve fitting was used to investigate the linear connection that exists between these variables. In order to determine whether or not the SII index and SUI relationship remained stable across a variety of demographic strata, subgroup analyses were carried out. Results The research had a grand total of 20,849 individuals. After full adjustment for confounders, each 1-unit increase in log₁₀(SII) was associated with a 29.7% increase in the odds of SUI (OR=1.297, 95%CI 1.028–1.636, P = 0.032). Categorizing the SII index into quartiles revealed that even in the highest SII quartile, there remained a significant positive association with SUI when compared to the lowest quartile (OR =1.177, 95%CI 1.013–1.368, P  = 0.038). Furthermore, an augmented correlation between alcohol consumption and SUI was detected among drinkers compared to non-drinkers ( P  < 0.05). Conclusions Higher systemic inflammatory burden reflected by SII is statistically significantly associated with increased odds of SUI in the U.S. adult population, though the effect size is small; this association is particularly prominent in alcohol drinkers.