Guoyue Shi

Precise discrimination of structural homologues, namely molecules with highly similar physicochemical properties, remains a critical bottleneck in artificial olfaction and limits its application in precision medicine and environmental monitoring. While deep learning has enhanced pattern recognition, current purely data-driven models often lack physical interpretability and may show limited transferability to homologous analytes not included during training. Here, we present a physics-data dual-driven deep learning framework that integrates time-resolved sensor responses with DFT-derived molecular descriptors, thereby bridging atomic-level electronic information and macroscopic sensing responses for VOC classification. The model achieved direction-averaged accuracy and macro-F1 scores of 83.88 and 83.53%, respectively, in bidirectional validation involving unseen homologues from five chemical categories. Interpretability and ablation analyses further confirmed that the complete DFT-assisted framework contributed substantially to classification. These results demonstrate the potential of integrating microscopic physicochemical descriptors with dynamic sensing signals to recognize unseen VOCs according to their chemical categories, providing a promising route toward interpretable VOC sensing within defined chemical categories.

Continuous creatinine monitoring is essential for the precise diagnosis, treatment, and personalized management of chronic kidney disease (CKD). However, existing diagnostic strategies suffer from invasive procedures and fail to capture dynamic profiles. In addition, the limited long-term stability of conventional enzymatic sensing interfaces remains a major challenge for wearable applications. To address these issues, this study developed a hollow microneedle-based (MN) creatinine sensor for in situ detection of creatinine. The hollow MN was modified with a multilayered core–shell structured Prussian blue analogue (PBA) NiCoMnPB, followed by the fabrication of a creatinine-specific recognition layer using a molecular imprinting polymer (MIP). The ternary core–shell NiCoMnPB can be synthesized via a one-step method, combining the structural stability of NiPB, the high electrochemical activity of CoPB, and the high capacity of MnPB, thereby significantly enhancing the sensor’s conductivity and electrochemical stability. The sensor enables real-time, continuous monitoring of creatinine in interstitial fluid (ISF), exhibiting high sensitivity, selectivity, reproducibility, and long-term stability. Animal experiments using a kidney injury model demonstrate that the dynamic changes of creatinine levels in ISF measured by the sensor were in high agreement with those obtained using commercial kits, effectively distinguishing creatinine levels under normal, injured, and recovery states, confirming its feasibility for in vivo application. This study provides an innovative solution for continuous monitoring of creatinine, offers a promising route for clinical translation for dynamic assessment of kidney damage, highlighting the sensor’s potential for personalized CKD management.

Given the inherent complexity of metabolic pathways and disease-associated agents, next-generation healthcare necessitates wearable, non-invasive, and customized approaches to continuously monitor a broad spectrum of physiologically relevant biomarkers for personalized health management. Moreover, existing data-based analytical strategies remain inadequate for delivering quantitative and predictive evaluations of health status in real-life settings. Here, we report an electronic multiplexed microneedle-based biosensor patch (eMPatch) that enables real-time, minimally invasive monitoring of key metabolic biomarkers in interstitial fluid, including glucose, uric acid, cholesterol, sodium, potassium, and pH. By integrating modular microneedle (MN) sensors into a skin-interfaced flexible platform, the eMPatch achieves robust mechanical stability and seamless skin conformity, thereby ensuring reliable and continuous sensing within the dermal space. In vivo validation in animal models under metabolic intervention highlights the strong capability of the eMPatch for real-time physiological tracking across diverse daily activities. Implemented with a machine learning algorithm, the eMPatch enables automatic feature extraction and multi-task health assessment, achieving a classification accuracy of 0.996 in distinguishing normal and diet-induced metabolic disorder for health condition identification and an R ² score of 0.977 for the corresponding degree evaluation. This study highlights the potential of the MN-integrated, machine learning-enhanced biosensing platform toward personalized health management.

Nowadays, specific biomarkers and detection methods for the early, rapid and non-invasive diagnosis of Alzheimer’s disease (AD) are scarce. MiR-34a a

Volatile organic compounds (VOCs) serve as key markers in environmental monitoring, healthcare, and food safety, yet their detection remains challenging due to low concentrations and complex matrices. Porous materials, with their tunable pore structures, high specific surface areas, and modifiable surface chemistries, have emerged as ideal sensing materials for electronic nose (e-nose) systems. This comprehensive review systematically explores the classification and synthetic strategies of porous materials─including microporous zeolites, mesoporous silicas, metal-organic frameworks (MOFs), carbon-based materials, and porous polymers─and their integration into e-nose platforms. We provide a detailed account of the working principles, sensor array designs, and pattern-recognition algorithms that underpin e-nose functionality. The review highlights the application of porous materials in enhancing sensor sensitivity, selectivity, and stability through molecular sieving, surface functionalization, and hierarchical pore-structure engineering. Advanced performance-optimization strategies such as pore-structure engineering, surface-chemistry modification, and device-level innovations are discussed. Moreover, we showcase practical applications in medical diagnostics, environmental monitoring, and food-quality control, demonstrating the transition from laboratory research to real-world applications. Despite progress, challenges remain in long-term stability, cost-effective scalability, standardization, and anti-interference capabilities. Future directions point toward miniaturization, AI-driven data analysis, multimodal sensor fusion, and bioinspired design. This review underscores the transformative potential of porous-material-based e-noses in achieving precise, reliable, and scalable VOC detection for next-generation intelligent sensing systems.

This study introduces a radial-hierarchical, diffusion-enhanced spatiotemporal sensing paradigm for volatile organic compound (VOC) analysis via an integrated microchamber paper-based chromatomimetic e-nose. The proposed system synergizes interlayer spatiotemporal dynamics with planar spatial variance by employing a radially symmetric electrode array and a hierarchical porous chemoresistive ink (CuP@G). This design leverages molecular diffusion gradients across the sensing plane, enabling precise discrimination of complex VOC mixtures through multidimensional "spatiotemporal fingerprints". A physics-informed framework integrates molecular transport principles with multitask learning convolutional neural network (MTL-CNN) analytics, achieving unprecedented resolution in real-sample classification. Systematic validation demonstrates superior performance in discriminating diverse VOCs, binary mixtures, and authentic tobacco samples (origin and level classification accuracy: 92-99%). This work establishes a scalable blueprint for high-fidelity VOC analytics, bridging gas diffusion physics with intelligent signal processing to advance e-nose technology toward precision-driven design.

Personalized healthcare requires convenient and high-fidelity approaches to monitoring metabolic biomarkers, for example, uric acid (UA), toward next-generation disease management. Although wearable microneedle-based devices stand out with miniaturized needle-like sensors capable of transdermal monitoring of UA in interstitial fluid (ISF) in an imperceptible manner, their rigid nature and limited lifespan pose considerable challenge to the practicability as continuous biosensors in real-world application. Here, we present a wearable microneedle-based biosensor that integrates the strength of microneedle with a Prussian Blue analog, that is, nickel hexacyanoferrate, enhanced sensing interface to minimally invasive access UA in ISF and enable highly blood-correlated, continuous, and long-term monitoring. By incorporating conductive poly(2,6-diaminopyridine), the as-proposed microneedle biosensor renders highly sensitive monitoring of UA with a detection limit of 1.78 μM. In addition, the fully integrated device features prominent mechanical resilience and flexibility against undesirable physical deformations, ensuring reliable sensing performance of UA in daily activities. In animal models, the resulting biosensors achieved dynamic UA tracking with excellent stability and captured prolonged amperometric information interpreting the metabolic changes of UA levels during a seven-week in vivo experiment. The strong correlation between the obtained ISF and blood UA levels demonstrated the high analytical accuracy of the biosensor, which, thus, provided a powerful approach to fulfill personalized UA management in the future.

The study introduces a spatiotemporal chromatography-mimicking (SCM) e-nose that integrates laser-tailored graphene paper with a microchamber for precise volatile organic compound (VOC) discrimination. The SCM e-nose overcomes traditional array limitations with a single multifunctional component capable of accurate VOC differentiation via chromatography-mimic features. Advanced laser-engraving techniques fabricate a gas-permeable interdigitated electrode from graphene paper as the sieving framework. Key achievements include its single multifunctional component, economical and scalable design, distinct response patterns for different VOCs, remarkable ability to discriminate mixed VOCs, versatility for diverse applications including real-time on-site analysis, and ease of integration with electronic systems. The SCM e-nose represents a significant advancement in electronic nose technology, offering a compact, cost-effective solution for precise VOC detection and analysis.

The design of single-atom catalysts with dual functions has emerged as a promising strategy for developing high-performance sensing platforms. Herein, we reported a facile host-guest strategy for synthesizing an atomically dispersed Co catalyst (Co-N-C), where Co atoms were uniformly anchored on the N-doped carbon matrix derived from zeolitic imidazolate framework-8. The as-prepared Co-N-C exhibits both excellent electrochemical sensing and peroxidase-like colorimetric activities toward the detection of three important bioactive small molecules, ascorbic acid (AA), dopamine (DA), and uric acid (UA). The electrochemical sensor demonstrated ultrahigh sensitivity with detection limits of 4.83, 1.36, and 0.371 μM for AA, DA, and UA, respectively, along with outstanding selectivity against common interferents and stable performance. Meanwhile, the colorimetric method also showed analytical performance with detection limits of 2.24 μM (AA), 3.09 μM (DA), and 2.97 μM (UA). The results indicate that the electronic modulation of Co through precise nitrogen coordination enhances the affinity of Co-N_x for target reactants, thereby promoting adsorption and electron transfer throughout the reaction. This improves catalytic efficiency and selectivity, establishing Co-N-C with dual-catalytic functionality as a promising material for biosensing applications.