Melochia (Byttnerioideae - Malvaceae) is widely distributed across tropical and subtropical regions. In Brazil, however, its distribution, species richness, diversity, and endemism remain poorly understood. The present study aimed to analyze the biogeography of Melochia and infer which distribution patterns best reflect its geographic distribution in Brazil and where its centers of diversity and endemism are located. Species occurrence data were compiled from online databases, resulting in 4,136 cleaned records representing 27 species. Richness and diversity analyses were conducted using DIVA-GIS, while QGIS was used to create the distribution maps. Parsimony Analysis of Endemicity was performed with PAUP. Our results indicate that most Melochia species are widely distributed. Seven areas of diversity were identified: Pantanal, Central Brazilian Cerrado, Northern and Southern Espinhaço Range, Serra dos Carajás, Tapajós, and Raposa Serra do Sol Indigenous Land. Two areas of endemism were detected: Chapada Diamantina, and Chapada do Araripe. These patterns highlight the ecological relevance of Melochia across varied environmental gradients and its biogeographical structure, particularly in the Central-West and Espinhaço Range. The study also reveals significant knowledge gaps and underrepresentation of certain species in herbarium collections and databases, stressing the need for increased botanical exploration and habitat protection.

Accurate characterization of the optical properties of blood is essential for biomedical optics, diagnostic imaging, and therapeutic applications. conventional measurement techniques are frequently constrained by complex setups and idealized assumptions, highlighting the need for robust predictive models. This study presents a comparative machine learning framework to predict three blood optical properties: the absorption coefficient (µa), the scattering coefficient (µs), and the refractive index (RI). We developed and systematically optimized support vector regression (SVR), random forest (RF), and extreme gradient boosting (XGBoost) models using the Optuna framework, training them on a comprehensive dataset utilizing wavelength, hematocrit (Hct%), and oxygen saturation (SO2%) as input features. On the test dataset, the Optuna-optimized RF model achieved the highest predictive accuracy, with an R2 of 0.9981 and a root mean squared error (RMSE) of 1.3680. It significantly outperformed both SVR (R2 = 0.9736) and XGBoost (R2 = 0.9865), demonstrating superior generalization capability. SHAP analysis provided model interpretability, confirming that the predictions were based on a physically meaningful relationship. Wavelength was the dominant predictor, while hematocrit and oxygen saturation contributed with light-tissue interaction principles, reinforcing model validity. Additionally, the analysis demonstrated that tree-based models exhibit superior feature utilization with clearer decision boundaries compared to SVR, more effectively capturing the nonlinear, hierarchical relationships inherent in biological systems. Finally, the optimized models were deployed via a user-friendly web interface, providing the biomedical optics community with accessible, interpretable tools for rapid optical property prediction to advance fundamental research and clinical applications.

Optimization methods play a vital role in solving large-scale problems across engineering and environmental fields. Among them, the conjugate gradient (CG) method is popular for its computational efficiency, and recent developments-such as the three-term CG variant-have shown improved convergence and numerical stability. This paper proposes an improved three-term CG method for unconstrained optimization. Inspired by the Three-Term Zheng-Huang-Shi (ZHS) CG parameterization and Three-Term Rivaie-Mustafa-Ismail-Leong (TTRMIL), this method is specifically designed to improve the computational performance of CG methods. In the proposed method, sufficient descent conditions and global convergence properties for general functions are mathematically proven under assumption and criteria from the strong Wolfe line search. Numerical experiments conducted on several unconstrained optimization problems highlight the superiority of the new method over certain CG methods with similar characteristics. In real-world applications, the proposed method is extended to address the challenge of reducing carbon emissions for sustainability, particularly within the Low Carbon Supply Chain (LCSC) optimization problem.

Accurate prediction of corrosion depth in buried pipelines is essential for integrity assessment, reducing inspection and maintenance costs, and managing risk in pipeline transportation systems. Although machine learning (ML) and deep learning (DL) techniques have been widely applied to corrosion prediction, their reliability depends strongly on the quality, representativeness, and balance of the training data. In many previous studies, data imbalance has remained a key limitation, leading to biased models that mainly learn the majority class, underestimate severe corrosion cases, and exhibit reduced overall performance. In this study, a polynomial-based Synthetic Minority Over-sampling Technique (SMOTE) is employed as a data-level solution to alleviate class imbalance and improve model learning. SMOTE is integrated with several predictive models, including a deep learning neural network (DLNN), eXtreme Gradient Boosting (XGBoost), support vector regression (SVR), and multiple linear regression (MLR). The models are developed and evaluated using both the original and the SMOTE-oversampled datasets, and their performance is compared using several statistical metrics. The results show that incorporating SMOTE significantly enhances the predictive capability and robustness of the ML and DL models, attaining a significant reduction of 24.94% in root mean square error (RMSE) compared with the corresponding models trained without SMOTE. The DLNN-SMOTE model provides the best performance, achieving a lower RMSE (0.624) and a higher correlation coefficient (R = 0.940) than other models. These findings demonstrate that SMOTE-enhanced AI models offer an efficient and cost-effective framework for corrosion depth prediction and support more reliable and sustainable integrity management of buried pipelines.

Child undernutrition remains a significant population health challenge in Nigeria and sub-Saharan Africa. Existing research has established an association between socioeconomic factors and child undernutrition, but less attention has been paid to whether the socioeconomic gradient is similar across indicators that capture different underlying processes. Drawing on fundamental cause theory, this study examines the association between household wealth and child stunting and wasting in Nigeria, and the extent to which that association is attenuated by other socioeconomic and geographic factors. Using the 2024 Nigeria Demographic and Health Survey (n = 9,410 children for stunting; n = 9,464 for wasting), Survey-weighted logistic regression models were fitted, using stepwise model building and joint tests of interaction. Household wealth was associated with stunting in a graded pattern but showed no association with wasting. The findings suggest that chronic and acute undernutrition may not share the same social determinants and may call for different interventions.

Computational optical imaging reconstructs signals from limited and noisy measurements, leading to ill-posed inverse problems that require regularization. Graph regularization (GR) has shown promise for modeling complex spatial correlations, but its relationship to classical Total Variation (TV) regularization remains unclear. Using ghost imaging as a representative inverse imaging application, this paper establishes a unified theoretical framework for graph-regularized estimation. We derive explicit expressions for the bias and variance, obtain an upper bound on the mean squared error (MSE), and develop an information-theoretic minimax lower bound for TV regularization. The analysis identifies regimes in which GR can be advantageous relative to TV, highlighting the different inductive biases of graph-based smoothness and TV-based local sparsity. In particular, GR is especially effective under low-frequency priors, high noise, strong graph connectivity, and severe undersampling. We also implement a graph-regularized reconstruction framework based on spatial and gradient similarities. Simulations and experiments show that the proposed method achieves competitive and often improved reconstruction quality relative to TV and sparsity-based baselines, especially under noisy and undersampled conditions. These results clarify the role of GR in computational optical imaging.

Parameter-efficient fine-tuning (PEFT) aims to adapt pre-trained vision models to downstream tasks. Among PEFT paradigms, sparse tuning achieves remarkable performance by adjusting only the weights most relevant to downstream tasks, rather than densely tuning the entire weight matrix. Current sparse tuning methods follow a two-stage paradigm. First, it locates task-relevant weights by gradient information, which overlooks the parameter adjustments during fine-tuning and limits the performance. Second, it updates only the located weights by applying a sparse mask to the gradient of the weight matrix, which results in high memory usage due to the storage of all weight matrices in the optimizer. In this paper, we propose a one-stage method named SNELLA (Sparse tuning with kerNELized LoRA and Adaptive bi-level sparsity allocation) to overcome the above limitations. For memory usage, SNELLA selectively updates the weight matrix by adding it to another sparse matrix merged by two learnable low-rank matrices. We extend the low-rank decomposition by replacing the inner product with our proposed nonlinear kernel Mix-K, which largely increases the rank of the merged matrix while preserving training stability. The increase in rank prevents the interdependency among weight updates and enables better adaptation to downstream tasks. For locating task-relevant weights, we propose an adaptive bi-level sparsity allocation mechanism that encourages weights to compete across and inside layers based on their importance scores in an end-to-end manner under a predefined overall budget of weight updating, ensuring task-relevant weights to attain higher scores and be updated more likely. Extensive experiments are conducted on classification, segmentation, and generation tasks using pre-trained vision models with different parameter scales, architectures, and pre-training strategies. The results show that SNELLA achieves state-of-the-art performance with low memory usage. Notably, SNELLA obtains 1.8% (91.9% v.s. 90.1%) higher Top-1 accuracy on the FGVC benchmark compared to SPT-LoRA. Compared to previous sparse tuning methods, SNELLA achieves a memory reduction of 31.1%-39.9% across models with parameter scales from 86 M to 632 M.

Understanding and optimizing blood flow behavior in curved vessels is crucial for improving cardiovascular treatment planning and reducing complications such as thrombosis and impaired oxygen transport. Although recent advances in computational modeling have enhanced the simulation of blood rheology, systematic optimization of flow parameters under controlled geometric conditions remains limited. This study addresses this gap by optimizing the hemodynamic performance of curved blood vessels using numerical modeling and multi-objective optimization. Three key parameters hematocrit level, kinematic viscosity, and curvature degree were analyzed to enhance flow velocity and minimize shear stress, balancing oxygen delivery and thrombus formation risks. A numerical model based on the HemoCell framework and the Lattice Boltzmann-Immersed Boundary Method simulated 250 scenarios across ten vessel geometries with varying curvature. Predictive modeling was performed using gradient boosting and support vector regression integrated with bio-inspired optimization algorithms. The Sparrow Search Algorithm combined with gradient boosting achieved the highest velocity prediction accuracy (R2 = 0.99902), while the Hunger Game Search Algorithm demonstrated superior performance in shear stress prediction (R2 = 0.99577). These results confirmed the reliability of the surrogate models for optimization. Pareto fronts and optimal points were then identified, revealing a shear stress of 587.70 Pa and a velocity of 0.039 m/s as one of the optimal trade-offs. The proposed approach provides a validated framework for hemodynamic optimization and provides preliminary insights that could inform future studies on thrombus risk assessment and the bio-inspired design of bypass grafts.

To develop a machine learning (ML)-based workflow for virtual patient-specific quality assurance (PSQA) to predict local gamma passing rates (GPRs) and QA pass/fail outcomes for Volumetric Modulated Arc Therapy (VMAT) plans. Treatment plan files from 156 double-arc VMAT cases were used, resulting in 312 beam-specific datasets. A total of 48 plan complexity metrics (PCMs), derived from DICOM RT plan files based on the UCoMX framework, were calculated for each arc. Feature selection was performed using a blended ensemble of Random Forest (RF), Gradient Boosting Decision Tree (GBDT), eXtreme Gradient Boost (XGBoost), and linear Support Vector Machine (SVM) models, yielding 15 key predictive features. Regression models using RF, GBDT, XGBoost, SVR, and an averaging ensemble were trained to predict local GPR values. Patient-level grouped train-test splitting and grouped 5-fold cross-validation were employed to prevent data leakage. Classification models were developed using a 90% GPR threshold to predict QA pass/fail outcomes. Model performance was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), ROC-AUC, sensitivity, specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV). Among the regression models, GBDT achieved the best overall performance with an MAE of 1.85% points and an RMSE of 2.35% points, followed closely by the averaging ensemble model. Grouped 5-fold cross-validation demonstrated stable regression performance across all models. For classification analysis, grouped 5-fold cross-validation ROC-AUC values ranged from 0.721 to 0.772, with SVM achieving the highest mean cross-validation ROC-AUC. On the independent test dataset, SVM achieved the highest ROC-AUC (0.725) and specificity (0.829), while the averaging ensemble model demonstrated balanced classification performance with improved specificity and PPV. An ML-based workflow for predicting VMAT PSQA outcomes from plan complexity metrics derived from DICOM RT plans has been developed. The results demonstrate the feasibility of predicting QA outcomes with moderate accuracy and provide a framework for further development of ML-assisted PSQA. Future studies incorporating additional plan, delivery, and machine-related features, together with larger multi-institutional datasets, may further improve predictive performance and clinical applicability.

Tick-borne pathogen systems are inherently multi-species networks, yet most predictive frameworks model vector distributions independently, thereby overlooking community structure. Here, we develop a stacked approach to infer and map tick communities from species distribution models (SDMs), leveraging habitat suitability to reconstruct patterns of co-occurrence across large spatial extents. We applied this framework to eight tick species across the Iberian Peninsula using high-resolution climatic and vegetation data (1 km resolution; years 2010-2025). Species-level predictions were integrated and clustered using Gaussian Mixture Models (GMMs) to identify statistically supported communities, followed by downstream analyses of dominance and specialization. Climatic variables alone achieved high predictive performance, with limited statistical gains observed from the inclusion of vegetation descriptors. Species distributions exhibited extensive spatial overlap, revealing that most of the study area supports multi-species assemblages rather than isolated populations. Community reconstruction identified a structured gradient extending from Atlantic systems dominated by Ixodes ricinus to Mediterranean assemblages characterized by Hyalomma lusitanicum and Rhipicephalus bursa. Transitional communities formed a continuous ecological spectrum, demonstrating strong environmental filtering within Mediterranean assemblages. Our results show that SDM-derived stacked models can be leveraged to infer biologically meaningful community structures, providing a practical and scalable alternative to data-intensive joint modelling approaches. This framework offers a robust tool to integrate community ecology into tick surveillance and to improve the spatial risk assessment of tick-borne diseases.

Mountain headwater streams are highly sensitive freshwater ecosystems where hydromorphological gradients strongly influence habitat structure and biological communities, highlighting the need for integrated ecological assessment. This study uses Israfil Creek (Denizli, Türkiye) as a model snowmelt-fed mountain headwater to investigate how hydromorphological gradients and metric sensitivity structure ecological patterns. A total of 55 macroinvertebrate taxa were identified, with Chironomidae as the most diverse group. Community composition showed a clear longitudinal gradient, with cold-adapted taxa (e.g., Diamesa spp.) dominating upstream and more tolerant taxa (e.g., Chironomus riparius and Gyraulus albus) increasing downstream. Substrate composition shifted from lithal dominance in headwaters to finer sediments and particulate organic matter downstream. Saprobic indices indicated predominantly xenosaprobic-oligosaprobic conditions upstream, with increasing mesosaprobic influence downstream. Although Biological Monitoring Working Party (BMWP) scores consistently indicated high ecological quality, Ephemeroptera, Plecoptera, and Trichoptera (EPT) percentages declined longitudinally, revealing differences in metric sensitivity. Hydroclimatic analysis highlighted strong elevational control on snow persistence and pronounced precipitation seasonality. Overall, ecological patterns were governed primarily by hydromorphological gradients rather than conventional water quality signals, and reliance on single metrics may obscure subtle ecological transitions. These findings emphasize the importance of integrated, multimetric approaches for assessing climatically sensitive mountain stream ecosystems.

The study investigated thermoregulatory dynamics in Murrah buffalo calves under heat stress using infrared thermography. Six calves (8-12 months old; weighing 125 ± 15.35 kg) were exposed in a psychrometric chamber to temperatures of 25, 28, 31, 34, 37, and 40 °C, corresponding to temperature-humidity indices (THIs) ranging from 73 to 90. Each exposure lasted 12 h daily (0600-1800 h) for 10 consecutive days, with animals kept at 25 °C and 65% humidity (THI: 73.49) during the remaining 12 h. On the 10th day, respiratory rate (RR), pulse rate (PR), rectal temperature (RT), feed intake and water intake were recorded and infrared thermography was done to record body surface temperatures (BST) at head, neck, eye, ear, croup, and muzzle besides blood sampling. Significant increases (p < 0.05) in RT, PR, RR, and water intake were first observed at 34 °C, while feed intake declined significantly (p < 0.05) at 34 °C. BSTs increased significantly (p < 0.05) at 28 °C compared to 25 °C and rose progressively thereafter, though values at 28 and 31 °C, and 34 and 37 °C, were similar except at muzzle. An approximate 5 °C rectal-to-surface temperature gradient was associated with effective thermoregulation under cyclic heat stress. Hematological changes included leukocytosis (p < 0.05) at 34 °C, along with lymphocytosis, granulocytopenia (p < 0.05) and elevated plasma cortisol (p < 0.05) at 31 °C. Among anatomical sites, muzzle temperature exhibited strongest correlation with ambient temperature (p < 0.01), while ear and neck temperatures correlated most strongly with RT (p < 0.01). Correlation analysis further revealed that all studied physiological and hematological parameters, except erythrocytic indices and granulocyte percentage, were significantly (p < 0.05) associated with BST at one or more body sites. In conclusion, progressive cyclic heat exposure induced staged thermophysiological adjustments in Murrah buffalo calves, with 31-34 °C representing a transitional range of systemic activation. Maintenance of an approximate 5 °C rectal-to-surface temperature gradient reflected effective thermoregulatory reserve, whereas its narrowing indicated increasing thermal strain. Ear and neck surface temperatures showed strong associations with rectal temperature, supporting their potential as non-invasive indicators of thermal status. Infrared thermography may therefore aid precision-based heat stress monitoring in buffalo.

Assessment of environmental pollution in the Arctic is becoming increasingly important in the context of active industrial development and global climate change. In this study, soils and bottom sediments from the northern regions of Western and Central Siberia were analyzed for persistent organic pollutants including polycyclic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs). Total contaminant concentrations were quantified and spatial trends in the relative distribution of individual PAH and PCB groups were identified across Arctic subregions defined by the CAFF working group. Potential pollution sources were evaluated using non-metric multidimensional scaling (nMDS) and ANOSIM. Furthermore, the specific characteristics of lake bottom sediment pollution under the influence of active oil and gas production were examined. Total PAH concentrations in background soils and sediments ranged from 0.3 to 112.6 µg/kg dry weight, exhibiting a significant northward increasing gradient, heavily dominated by low-molecular-weight (2-3 ring) compounds. In contrast, total PCB concentrations remained uniformly low, ranging from 0.2 to 9.0 µg/kg dry weight regardless of the subregion. The nMDS ordination demonstrated a pronounced decrease in the diversity of PAH compositions with increasing latitude and proximity to nearest industrial center, reflecting the combined effects of temperature-dependent atmospheric fractionation and regional emissions. Compositional fingerprinting confirmed that the background PCB pollution was predominantly driven by the historical regional footprint of Soviet "Sovol" technical mixtures. Compared to pristine background baselines, lake sediments within active oil and gas fields exhibited an increase in both the total content and the relative abundance of heavy PAHs. In contrast, local hydrocarbon extraction activities did not influence either the total loads or the homologous profiles of PCBs, leaving the stabilized regional historical background unaltered.

The development of efficient sensing platforms for highly toxic gases remains a critical challenge in environmental monitoring and industrial safety. Despite the remarkable progress of two-dimensional heterostructures, their temperature-dependent gas-sensing behavior remains poorly understood, limiting their practical applicability. To address this gap, the gas-sensing performance of in-plane In2SSe/Al2SSe (X/Y), In2SSe/Al2SSe/In2SSe (X/Y/X), and Al2SSe/In2SSe/Al2SSe (Y/X/Y) heterostructures was systematically investigated over a wide temperature range. Interface engineering effectively modulates the electronic structure and adsorption behavior, resulting in enhanced sensing performance. The Y/X/Y heterostructure exhibits the highest response toward Cl2 (17,521.68 at 298 K), whereas ClCN shows the strongest adsorption (- 2.46 eV). These findings demonstrate the potential of interface-engineered Janus heterostructures as thermally robust platforms for next-generation toxic-gas sensing.