Search Results (471)

Automatic detection of sensitive fields in structured data is a critical prerequisite for privacy compliance and data governance. However, existing approaches face severe cross-domain generalization challenges. Hand-crafted pattern rules often fail under highly heterogeneous naming conventions, while single statistical models tend to overfit and degrade sharply under distribution shifts between training and deployment domains. These limitations stem from the weak semantic signals and distributional heterogeneity of structured data, which make it difficult to simultaneously capture explicit rules and latent, variant-sensitive attributes. To address these challenges, we propose a detection framework based on multi-view complementary features and a Hierarchical Gated Residual Network (HGRN). The framework first constructs a full-spectrum feature system that integrates explicit rules and implicit statistical fingerprints (e.g., entropy and character texture) to fill the semantic gap. It then introduces a decision mechanism combining robust priors with dynamic residual calibration: a random forest provides a stable probabilistic anchor, which is further nonlinearly corrected by a learnable gating-and-expert network. This design explicitly resolves the cognitive conflict between rule-dominated regions and complex distributional regions. Experiments on multiple real-world datasets—including DeSSI, CMS Open Payments and Home Credit—show that the proposed method achieves a Macro-F1 of 0.9408 on DeSSI and exhibits strong in-domain performance. Under strict frozen-model cross-domain transfer, HGRN mitigates the catastrophic collapse observed in pure neural baselines and maintains moderate detection capability, offering interpretable trust allocation between rule-based priors and data-driven correction in both financial and healthcare scenarios. Full article

WiFi sensing has gained widespread attention as a promising technology, owing to its non-intrusiveness, strong privacy-preserving characteristics, and cost-effective deployment, enabling diverse application scenarios. In addition, the stable spatial characteristics and symmetry-related patterns exhibited by human body postures in WiFi signal propagation provide new possibilities for robust person identification. In traditional WiFi-based person identification technologies, although gait recognition has achieved certain success, it is complex to operate and limited in application scenarios, increasing the constraints on recognition. This issue becomes more pronounced in large-scale user scenarios, where the system performance tends to degrade and exhibit instability. To overcome these challenges, we introduce a new person identification system called WiPID. The WiFi signals extracted from the static postures of users are treated as a “biometric fingerprint” for identity verification. An end-to-end deep learning framework is utilized by WiPID to process WiFi signals, and a convolutional autoencoder is adopted to preprocess the signals directly, effectively reducing redundant information and greatly simplifying the WiFi data processing. Furthermore, the integration of a multi-scale feature extraction module improves the system’s ability to capture discriminative features. The proposed system not only reduces operational complexity but also extends its applicability to a wider range of scenarios, thereby enhancing recognition performance. In an experiment involving 50 volunteers, WiPID achieved an average recognition accuracy of up to 98%, demonstrating the method’s suitability for large-scale person identification scenarios. In addition, a real-time identification experiment has been conducted on PCs and commercial WiFi devices. Experiments have proven that WiPID can achieve real-time person identification on Internet of Things devices, further validating its feasibility and stability in practical applications. Full article

Accurate characterization of closely related Ilex taxa is essential for the conservation, documentation, and utilization of plant genetic resources. In this study, five Ilex taxa from eastern China (Ilex rotunda Thunb., Ilex chinensis, Ilex cornuta Lindl. & Paxt., Ilex cornuta ‘Fortunei’, and Ilex latifolia Thunb.) were evaluated using an integrated framework combining fruit morphometric traits, CIELAB color parameters, and electronic-nose (E-nose) volatile fingerprints. Fruit transverse diameter, longitudinal diameter, single-fruit weight, fruit shape index, and peel color traits (L*, a*, b*, and chroma, C*) differed significantly among taxa (one-way ANOVA, all p < 0.001). I. cornuta produced the largest and heaviest fruits, I. chinensis showed the most elongated fruit shape, and I. rotunda exhibited the highest redness and chroma values. Chemometric analyses of E-nose responses further improved taxon discrimination and revealed clear divergence in volatile-response patterns. Trait-space relationships were broadly consistent with the preset phylogenetic framework, with I. rotunda showing the greatest divergence and I. cornuta and I. cornuta ‘Fortunei’ showing the closest similarity. These findings indicate that integrated fruit phenotyping and rapid volatile profiling provide a practical approach for Ilex germplasm identification, comparative evaluation, and resource documentation, with potential value for conservation planning and horticultural utilization. Full article

Interpretation of pharmacometabolomics results, aiming particularly at biomarker (sets) discovery for drug exposure, remains a major challenge. The metabotyping of drug exposure depends on resolution of specific metabolomics techniques and comprises individual metabolic phenotypes (“metabotypes”), disease-, drug- and microbiome-specific patterns, as well as conditional metabolic states (e. g. fasting). In this clinical trial with 16 healthy participants, an exploratory objective was to evaluate the untargeted urinary metabolomics of voriconazole, administered in four single doses, using proton nuclear magnetic resonance (¹H-NMR) spectroscopy. Voriconazole is a second-generation triazole and a potent inhibitor of drug-metabolizing enzymes such as cytochrome P450 (CYP) isozymes CYP3A4 and CYP2C19. Therefore, identification of metabolites reflecting acute CYP3A4 inhibition was of particular interest. On two treatment days without and with voriconazole (with background microdosed midazolam and omeprazole administration for CYP3A4 and CYP2C19 phenotyping, respectively), spot urine was collected after overnight fasting (predose) and 4 h later (postdose fasting). In the postdose versus predose fingerprints, most changes at the annotated metabolite level were attributable to fasting metabolomics or potential confounders. ¹H-NMR spectroscopy identified neither a short-term voriconazole-specific signature nor patterns or metabolites potentially reflecting acute CYP3A4 inhibition. Our study emphasizes crucial significance of strict standardization of fasting time and minimization of confounder influences by clinical trial design as well as selection of adequate baselines and high-resolution analytical techniques in pharmacometabolomics research, especially for biomarker discovery. Full article

Selective inhibition of poly(ADP-ribose) polymerase 1 (PARP1) may reduce the hematologic toxicity associated with dual PARP1/PARP2 inhibition. We performed molecular dynamics simulations for five selective inhibitors in complexes with PARP1 and PARP2, using three independent 50 ns runs per complex after docking and equilibration, followed by protein–ligand interaction fingerprint and statistical analyses. All complexes remained dynamically stable, with ligand root-mean-square deviation values generally within 0.3 nm. Comparative analysis identified three αF-helix residue pairs with nominally reduced interaction frequencies in PARP2: Asn767/Ala336, Leu769/Gly338, and Asp770/Asp339 (p < 0.05). After Benjamini–Hochberg correction for multiple comparisons, Leu769/Gly338 remained significant (q < 0.05), indicating that this pair represents the most statistically robust interaction difference within this region. Using palacaparib as the most selective inhibitor, these differences were associated with weakened or lost hydrophobic, van der Waals, and cation–π interactions in PARP2. Selective binding of modern PARP1 inhibitors appears to be associated with αF-helix-dependent interaction patterns, providing a mechanistic basis for the rational design of next-generation selective inhibitors with improved selectivity and potentially reduced toxicity. Full article

Background: Endogenous opioids, such as endomorphin-2, are key regulators of the body’s pain pathways and mediate analgesia by engaging the $\mu$-opioid receptor. This class of opioids are distinguished by their transient activation of the $\mu$-opioid receptor, which is attributed to rapid enzymatic degradation. Methods: To understand how degradation of endomorphin-2 by the enzyme DPP IV affects its interaction with the $\mu$-opioid receptor, we analyzed the ligand–receptor conformational dynamics and interaction patterns of molecular dynamics simulations data of morphine, fentanyl and endomorphin-2 and one degradation product Phe-Phe-NH₂, using molecular fingerprints and the mathematical framework ISOKANN. Results: Our analyses revealed that both the clinically relevant opioids, morphine and fentanyl, as well as the endogenous opioid endomorphin-2, adopt a set of recurring binding conformations within the $\mu$-opioid receptor binding pocket, maintaining overlapping interaction motifs throughout the simulations. In contrast, Phe-Phe-NH₂ failed to maintain a persistent binding mode over the simulated timescale. This instability arises from the dipeptidyl peptidase IV mediated cleavage of endomorphin-2, which generates Phe-Phe-NH₂ and removes critical proline and tyrosine residues, thereby leading to the loss of stabilizing hydrophobic contacts with receptor residues Tyr150^3,33, Val238^5,43 and Val302^6,55. Conclusion: By mapping structural interaction motifs essential for stable $\mu$-opioid receptor binding, this study provides mechanistic insights into how endogenous degradation reshapes ligand–receptor interactions. Full article

Clinical prediction from electronic health records (EHRs) is complicated by pervasive missingness and label scarcity, which make performance sensitive to the match between data conditions and pipeline choice. Choosing the best pipeline for a new incomplete dataset still requires costly trial-and-error. We cast this as an algorithm selection problem and address two bottlenecks—instance scarcity and distance quality—that have so far prevented meta-learning from reaching clinical settings. Graph neural networks offer diverse strategies (patient similarity networks, bipartite imputation graphs, attention-driven feature interaction), yet no single architecture dominates across missingness patterns, and selecting the best pipeline for a new dataset remains a trial-and-error approach. Formal algorithm selection could automate this choice but requires many characterized meta-instances—more than clinical settings typically provide. We propose two solutions: (1) constructive instance augmentation, applying controlled quality perturbations (MCAR and MNAR missingness injection, label trimming) to 20 base EHR datasets to expand the meta-knowledge base to 83 characterized meta-instances, each described by a 10-dimensional missingness fingerprint, without additional model training; and (2) dynamic-supervised metric learning, using differential evolution to optimize fingerprint feature weights so that static distances preserve method-performance similarity captured by dynamic fingerprints, which require model sweeps and are unavailable at deployment. Under base-dataset-level leave-one-dataset-out cross-validation over 21 pipelines, the resulting metric-learned kNN recommender attains the highest win rate (20.5%) among non-oracle strategies on the augmented store, selecting the correct pipeline more often than any fixed default. At deployment, the recommender needs only the 10-dimensional static fingerprint with pre-learned weights; no sweep data is required for new datasets. Cross-domain evaluation on 25 external subsets (colorectal cancer, kidney disease, MIMIC-IV) demonstrates framework modularity: when the fingerprint module is adapted (standard meta-features in place of the missingness-specific set), the recommender achieves regret of 0.025 (55% below random selection). Full article

Background/Objectives: Natural compounds occupy a pharmacologically rich chemical space, characterized by abundant scaffolds, extensive functional group elaboration, and defined stereochemistry. In this context, Helichrysum italicum, a Mediterranean medicinal plant, represents a valuable source of polyphenols with multiple biological and pharmacological activities. Methods: Here, we introduce an inverse molecular docking fingerprint approach to systematically investigate eight major Helichrysum italicum polyphenols, including $\alpha$-pyrones (arzanol, ethylpyrone), flavonols (gnaphaliin, kaempferol, quercetin), and flavanones (naringenin, pinocembrin, hesperetin). More than 40,000 human protein structures from the Protein Data Bank were screened to generate target-based inverse docking score fingerprints for each compound. Results: Hierarchical clustering of these fingerprints revealed shared binding patterns among structurally related polyphenols and enabled hypothesis generation regarding potential synergistic effects. Notably, favorable interactions were identified with PPARG and CARM1, supporting therapeutic relevance in inflammation and cancer, alongside additional targets associated with neurodegeneration and bone metabolism. Conclusions: This study establishes inverse docking fingerprints as a robust, mechanism-oriented method for natural product research and highlights Helichrysum italicum polyphenols as starting points for medicinal chemistry and drug discovery. Full article

This study comprehensively analyzed the effects of gamma irradiation (GI) on the flavor profile of aged light-flavor tartary buckwheat Baijiu (LTB) using E-nose, E-tongue, and high-sensitivity headspace–gas chromatography–ion mobility spectrometry (HS-GC-IMS). A total of 30 volatile organic compounds (VOCs) were identified, with concentrations showing significant dose-dependent correlations with GI treatment. Aging alone reduced harsh and pungent VOCs (e.g., 1-propanol, 2-methyl butanoic acid ethyl ester), while GI followed by aging further decreased undesirable compounds (e.g., butanal-D, pyrrolidine) and enhanced beneficial flavor components, such as 1,1-diethoxy ethane-D and butanoic acid propyl ester. Notably, this treatment partially restored 1-propanol, triethylamine, and 2-butanone-M, though their levels remained significantly lower than in newly brewed LTB, achieving a more balanced purity and flavor complexity. The significantly elevated levels of tetrahydrofuran-M/D, 1,1-diethoxy ethane-D, and cyclohexane in GI-treated aged LTB, along with their dose-dependent accumulation patterns, suggest their potential as reliable markers. Multivariate analysis confirmed that all three techniques (E-nose, E-tongue, and HS-GC-IMS) effectively differentiated LTB samples, with strong correlations between E-nose and HS-GC-IMS data, as well as between E-tongue and HS-GC-IMS results. This work provides flavor fingerprints and potential markers for gamma-irradiated LTB identification, while proposing an innovative technical approach for rapid flavor assessment of light-flavor Baijiu. Full article

The Temporal Analysis of Products (TAP) pulse response technique provides valuable insights into catalytic function and reaction kinetics. However, complex fragmentation patterns in the TAP mass spectrometry signals can complicate precise quantification, particularly when analyzing transient gas flux data typical of TAP experiments. This work demonstrates a standard defragmentation method that deconvolves transient TAP signals while maintaining the temporal resolution of the experiment. First, the integrals of calibration gas fluxes are used to determine the fingerprint fragmentation pattern and construct a fragmentation matrix. This matrix is then used to defragment experimental flux data at each recorded time point via a non-negative least squares regression. The effectiveness of this method is demonstrated using virtual data and control experiments with a TAP reactor system. The defragmentation is then applied to the more complex propane dehydrogenation reaction on a chromia/alumina catalyst, which can contain up to ten significant gas species in the reactor outlet. Initial propane pulsing reveals an induction period during which propane is fully oxidized to CO_2, followed by partial reduction to CO. Afterwards, there is a transition in chemistries towards coking and propylene production. Our example illustrates a practical method for the accurate determination of the time-dependent reactant/product concentrations and rates for a thorough analysis of the propane dehydrogenation kinetics. This approach can be broadly applied to any transient mass spectrometry experiment for a better understanding of catalyst-reaction dynamics. Full article

Neurotoxicity from drugs and environmental pollutants poses serious risks to brain function, yet existing computational models mainly target general neurotoxicity and lack specialized tools for brain-specific assessment. This study aimed to develop and validate a high-performance, brain-focused neurotoxicity prediction framework to improve drug safety evaluation and toxicity screening. We systematically analyzed molecular features, clustering patterns, and target predictions of brain-toxic compounds. Multiple feature representations were compared, including traditional molecular fingerprints, knowledge-guided pre-trained graph Transformer (KPGT) embeddings, and transformer-based MolFormer embeddings, combined with machine learning classifiers. Model performance was evaluated using multiple metrics, and SHAP analysis was conducted to identify influential molecular substructures. Toxic molecules showed physicochemical properties favoring central nervous system (CNS) penetration, including lower molecular weight, lower LogP, fewer hydrogen bond donors/acceptors, fewer rotatable bonds, and lower polar surface area (PSA). The KPGT-MLP model achieved the best balanced performance, with an accuracy (ACC) of 0.8928 and an ROC-AUC of 0.9459, clearly outperforming traditional fingerprint-based models, MolFormer-based models, and general prediction tools such as DI-NeuroT and ADMETlab 3.0. Overall, this study establishes a robust framework for brain-specific neurotoxicity prediction, with the KPGT-MLP model demonstrating strong accuracy and robustness. The proposed approach provides an effective strategy for early neurotoxicity screening and risk assessment, offering valuable insights for safer drug design and advancing computational toxicology and drug discovery. Full article

Serum contains diverse proteins whose concentrations vary with pathological conditions such as cancer, liver disease, neurological disorder, and infections. Conventional methods like serum protein electrophoresis (SPEP) and enzyme-linked immunosorbent assay (ELISA) are gold standards for protein identification; however, they are time-consuming and can miss abnormal serum protein levels. Inspired by chemical nose sensing based on selective sensor–analyte interactions, we synthesized five pyrene-conjugated fluorescent polymers (PFPs) with distinct side-chain head groups to construct a multichannel fluorescence sensor array. These polymers were screened for sensitivity to changes in serum protein levels using linear discriminant analysis (LDA), a machine learning method. This process led to the successful discovery of two PFPs that effectively detect protein level fluctuations. These PFPs provided a sensitive sensor array capable of generating a high-content response pattern (fingerprint) with six fluorescence channels. This sensor array successfully discriminated protein level fluctuations in serum with 98% jackknife classification accuracy and 95% unknown identification accuracy. This polymer sensor array holds strong potential as a diagnostic tool for serum-based samples and can be extended to other applications related to protein identification. Full article

Compost supplementation is widely used to improve yield and crop consistency in the cultivation of white button mushroom (Agaricus bisporus), yet practical alternatives to conventional protein-rich supplements and rapid candidate-screening approaches are still needed. In this study, plant- and byproduct-based supplements were first compared by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) to obtain qualitative fingerprints of extractable protein fractions, and were then evaluated in Phase III cultivation under both bag-based screening conditions and in a large-scale pull-mat system. Supplements differed notably in protein banding patterns and cultivation performance. In the bag trials, lupin grist and corn pellet produced the largest yield increases relative to the non-supplemented control, whereas in the commercial pull-mat trials lupin grist was the best-performing supplement, reaching 240.77 kg t⁻¹ compost. Under the present conditions, SDS-PAGE was useful as a qualitative screening aid for prioritizing candidates for cultivation trials, but not as a stand-alone predictor of yield. These results identify lupin grist as a practically relevant supplement candidate for commercial A. bisporus production. Full article

Moringa oleifera leaves are recognized as a nutrient-dense plant material of compositional and nutritional interest. This study aimed to characterize the nutritional and physicochemical properties of M. oleifera dried leaves through nutritional assessment and spectroscopic fingerprinting. Amino acid profiling, antioxidant activity assessment using ferric reducing antioxidant power (FRAP), 2,2-diphenyl-1-picrylhydrazyl (DPPH), 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), and oxygen radical absorbance capacity (ORAC) assays, chromatographic analysis of organic acids and sugars, color measurement, techno-functional characterization, and vibrational spectroscopy including Fourier Transform infrared with attenuated total reflectance (FT-IR/ATR) and Raman were employed. The crude protein content was 16.13 ± 0.43%. Moringa leaves contained all essential amino acids, with notably high tryptophan content (amino acid score, AAS = 200.00%). The amino acids limiting the nutritional value of the protein were primarily sulfur-containing amino acids (AAS = 49.57%) and lysine (AAS = 49.79%). Histidine, leucine, and valine also showed levels below the reference protein. Antioxidant activity exhibited solvent-dependent patterns: the 80% ethanolic extract demonstrated significantly higher FRAP activity (27.05 ± 1.05 mg Trolox Equivalent (TxE)/g dry matter (DM)) and ORAC values (107.24 ± 6.80 mg TxE/g DM), while no statistically significant differences between extracts were observed for DPPH, ABTS, or total phenolic content. Chromatographic profiling identified fructose and glucose as the predominant sugars, alongside citric, succinic, lactic, and acetic acids. The leaves exhibited favorable techno-functional properties, including high water holding capacity and water solubility index. Spectroscopic analysis revealed bands consistent with proteins, lipids, carbohydrates, and glycoside-related structures, while the preserved green-yellow coloration (hue angle 101.68°) indicated retention of pigment-related features during processing. These findings provide compositional and physicochemical characteristics of Moringa leaves relevant to their evaluation as a plant-derived food material. Full article

Background: Latent fingerprints are crucial forensic evidence, but their stability can be affected by environmental factors such as direct sunlight. The findings indicate that prolonged sunlight exposure may be associated with reduced fingerprint quality and forensic usability. Methods: A total of 322 groomed latent fingerprints from one volunteer were deposited on non-porous glass and exposed to direct sunlight for 1–7 weeks. A control sample was preserved without exposure. Fingerprints were developed using magnetic powder and assessed by minutiae counts. Usability was classified according to Slovak forensic standards. Statistical analysis was conducted using the Friedman test and Durbin–Conover test. Results: Significant differences in minutiae counts were observed between the control and selected exposure intervals (weeks 1, 3, 4, 6 and 7; p < 0.05). The degradation pattern was not linear, with initial decreases followed by stabilization in later weeks. Despite statistical differences, 99.38% of fingerprints remained usable for identification, and none were classified as non-usable. Conclusions: Prolonged direct sunlight exposure did not substantially reduce the identificatory value of groomed latent fingerprints on glass. Even after several weeks, most fingerprints retained sufficient ridge detail for personal identification, supporting their evidential relevance in outdoor forensic contexts. Full article