Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy最新文献

Background: Near-infrared (NIR) spectroscopy combined with chemometric modeling is a widely used rapid and non-destructive analytical technique, showing promise in estimating soil organic matter (SOM). However, acquiring a sufficient number of experimental spectra and corresponding SOM values is time-consuming, expensive, and often impractical, which can compromise the accuracy of regression models. Therefore, there is an urgent need for a strategy that can overcome data scarcity while maintaining robust estimation.

Results: This study development a hybrid framework that integrates a conditional variational autoencoder (CVAE) with a one-dimensional convolutional neural network (1D-CNN) for spectral data generation and regression modeling. The CVAE generated realistic spectra conditioned on target SOM content, and the generated spectra were combined with measured spectra to form an augmented dataset for regression modeling. The results showed that the CVAE accurately reproduced key spectral features and generated spectra consistent with the specified SOM values. The augmented dataset improved the estimation performance of the regression models. Among these models, the 1D-CNN outperformed partial least squares regression (PLSR) and random forest (RF), highlighting its superior ability to extract informative features from spectral data.

Significance: A novel spectral analytical methodology that help alleviate data scarcity and enhances regression performance under limited-sample conditions was established. By combining data augmentation with advanced regression models, the approach advances rapid and non-destructive soil analysis and provides a useful reference for other spectroscopic applications facing sampling limitations.

Accurate differentiation of benign and malignant thyroid lesions continues to pose a significant clinical challenge. Raman spectroscopy offers label-free molecular fingerprints of cells, yet the identification of diagnostic spectral patterns remains challenging. While artificial intelligence has been applied to analyze Raman data as one-dimensional (1D) signals, such approaches may overlook subtle nonlinear relationships across wavenumbers, particularly in cases involving spectrally similar constituents. Converting 1D spectral data into two-dimensional (2D) representations can preserve both amplitude and positional correlations, thereby uncovering latent temporal and structural features. However, such transformations risk incurring information loss, the extent of which is contingent upon the encoding strategy employed. To address this, we propose a novel multimodal deep learning framework that synergistically integrates 1D spectral and 2D spatiotemporal features, representing the first application in Raman-based thyroid cancer detection. Our model uniquely combines a Transformer to capture global dependencies in 1D spectra and a 3D-CNN to extract local spatial patterns from multiple 2D spectral transformations. These dual-modality features are adaptively fused through a multi-head cross-attention mechanism, enabling dynamic feature integration. The multimodal model ultimately achieves an accuracy of 94.7% in the identification of thyroid lesions, outperforming the unimodal Transformer and 3D-CNN models, which achieve accuracies of 91.0% and 89.4%, respectively. Notably, the multimodal model enhances interpretability by identifying contributions of key Raman peaks to the classification decision. Thus, the integration of SERS with explainable deep learning establishes a novel method for thyroid cancer diagnosis, achieving both exceptional diagnostic performance and significantly enhanced model interpretability.

Blood species identification is crucial for forensic investigations. Surface-enhanced Raman scattering (SERS) offers promising advantages including rapid detection, non-destructive analysis, and high sensitivity, but faces challenges from direct metal-biomolecule interactions that can cause spectral interference and sample degradation. Here, we present a novel approach using graphene/Ag/Styrene Ethylene Butylene Styrene (SEBS) flexible SERS substrates for rapid and non-destructive blood species determination. The innovative "sandwich" structure features single-layer graphene (0.34 nm) positioned between silver nanoparticles and blood samples, effectively preventing direct metal-blood contact while maintaining the superior SERS enhancement. The flexible substrate demonstrates good performance including high detection sensitivity (the detection sensitivity of 4-Mercaptobenzoic Acid (4-MBA) is 10^-12 M), excellent storage stability (>30 days), and remarkable mechanical durability (>100 folding cycles). We collected 682 SERS spectra from five animal species and employed a convolutional neural network algorithm for classification, achieving the classification performance with 98.54% accuracy, 98.29% precision, and 98.38% recall. The flexible substrate enables rapid analysis and maintains the spectral integrity of archived samples for up to 31 days. This flexible graphene /Ag/SEBS platform has reliable capabilities and offers accurate new means for on-site forensic applications: non-destructive analysis, field portability, rapid results, and high accuracy.

Serum albumin, the most abundant plasma protein, provides neuroprotection in Parkinson's disease (PD) via glial-neuronal signaling modulation and antioxidant activity. Meanwhile, lysosomes play an indispensable role in maintaining neuronal homeostasis. Therefore, investigating the dynamic changes of HSA in lysosomes is critical not only for elucidating the pathophysiological mechanisms of PD but also for offering a potential target for developing novel diagnostic strategies. However, existing fluorescent probes for lysosomal HSA imaging exhibit limitations in their response speed, detection range, aqueous solubility, and emission wavelength. Here, we develop SQ-1, a probe with a rapid response (<60 s), high sensitivity (LOD = 82 nM), excellent aqueous solubility, a long emission wavelength (>600 nm), and a wide detection range (0-60 μM). The probe SQ-1 proved to be a reliable tool capable of detecting a specific decrease in lysosomal albumin levels in a cellular PD model. Furthermore, in vivo imaging in a PD rat model uncovered elevated albumin levels in the brain. Crucially, SQ-1 enabled the quantitative detection of albumin in clinically relevant biofluids, including urine and CSF from PD model animals. The SQ-1 probe thus provides a powerful tool for detecting lysosomal HSA, holding broad potential for applications in neurobiological research and the diagnosis of neurological disorders.

The competition between the intramolecular hydrogen bonds (intra-HBs) in astaxanthin (AST) and the intermolecular hydrogen bonds (inter-HBs) between AST and ethanol (EtOH) controls the structural stability of AST in the solution during the cooling process. This study employs an integrated approach combining Moving window Two-Dimensional (MW2D) correlation Raman spectroscopy and density functional theory (DFT) simulations to investigate the interaction between ethanol (EtOH) and astaxanthin (AST) across a temperature range of 303 K-83 K. The MW2D analysis indicates that, below 253 K, EtOH reorganizes into cooperative multimers prior to any conformational change in AST. DFT simulations demonstrate that EtOH multimers weaken the intramolecular hydrogen bonds (intra-HBs) in AST through OH⋯OC interactions. These findings clarify the apparent contradiction between the extended conjugation length of AST and weakened inter-HBs.

The global rise in firearms-related crimes underscores the urgent need for rapid, sensitive, and field-deployable analytical tools to detect gunshot residues (GSR). Unlike earlier studies that utilized Lanthanide Metal-Organic Frameworks (Ln-MOFs) primarily as optical taggants incorporated in non-toxic ammunition, this work introduces a direct analytical interface between a separate luminescent Ln-MOF pellet and Organic analytes of GSR (OGSR) generated by firing of an optical taggant-free conventional ammunition, bridging molecular spectroscopy with forensic microanalysis. Nine luminescent Ln-MOFs were synthesized via the solvothermal method using various lanthanide metal precursors to explore their potential in OGSR identification. Among them, the gadolinium-based framework [Gd(NDC)] exhibited significant fluorescence quenching upon the addition of GSR, achieving a detection limit as low as ∼1 ppm. To enhance field applicability, a low-cost polystyrene-[Gd(NDC)] composite pellet was developed, enabling rapid, portable, and post-incident OGSR detection with minimal sample preparation. Comprehensive photophysical, microstructural, and surface analyses of Gd(NDC), including FTIR, XRD, BET, TGA, and DLS, revealed a clear correlation between the framework's structural parameters and its luminescent sensing performance.

Ruxolitinib (RUX), a selective JAK1/JAK2 inhibitor, is considered a therapeutic option for childhood B-cell precursor acute lymphoblastic leukemia (B-ALL) with JAK2 gain-of-function mutations. This study aimed to evaluate whether Raman spectroscopy combined with chemometric analysis can monitor the biochemical effects of RUX treatment in B-ALL cell lines. We employed single-cell confocal Raman imaging, flow cytometry, and Western blotting to assess the response of JAK2-mutated (MUTZ-5 and MHH-CALL-4) and wild-type (SEM) B-ALL cells to 10 μM RUX treatment over 48 h. Dimensionality reduction methods (PCA, t-SNE) and classification approach (o-PLS-DA) were applied to the spectral data to identify treatment-induced changes. RUX selectively reduced STAT5 phosphorylation and induced distinct Raman spectral shifts in JAK2-mutant cells, particularly in DNA- and protein-related bands. No significant changes were observed in JAK2 wild-type cells. The results demonstrate that Raman spectroscopy, when integrated with multivariate analysis, enables the non-destructive tracking of leukemia cell responses to targeted therapy and may support the development of phenotyping tools for drug monitoring in precision oncology.

Cardiovascular diseases (CVD) are becoming a serious threat to human health. These are considered the leading causes of mortality. Abnormal lipid concentrations in the body, such as High-density lipoproteins (HDL) and Low-density lipoproteins (LDL), are major factors that contribute to CVD. Surface-enhanced Raman spectroscopy (SERS) has the potential to be used to compare HDL and LDL cholesterol. In the current study, SERS was employed for the comparative profiling of HDL and LDL cholesterol using clinical blood serum samples along with silver nanoparticles (Ag-NPs) as the SERS substrate. The SERS spectral features of HDL and LDL cholesterol were clearly identified by applying various chemometric statistical tools. The Principal Component Analysis (PCA) was employed for the differentiation of blood serum samples of HDL and LDL cholesterol. Moreover, the support vector machine-Synthetic minority over-sampling technique (SVM-SMOTE) was used to accurately address the different imbalanced concentration of HDL and LDL cholesterol in order to reduce the risk of overfitting as compared to traditional machine learning algorithms. The SMOTE algorithm improves the interpretability of SVM by analyzing the minority classes of data sets. The macro-average Area Under the Curve (AUC) increased slightly from 0.97 to 0.98 with SMOTE, though the test Area Under the Curve was the same as 0.95. These results showed the accuracy and validation of the SMOTE model for the comparison of blood serum samples of HDL and LDL cholesterol.

The reprogramming of fatty acid metabolism in cancer cells holds significant importance in tumor research. This article focuses on the major applications of Raman imaging of single-cell phenotypes techniques in the study of metabolism in five types of cancer-prostate cancer, breast cancer, glioblastoma, cervical cancer, and colon cancer-at the single-cell level. These applications include noninvasive discrimination of cancer cell phenotypes by distinguishing intracellular fatty acid types, visualization of lipid distribution and differentiation within cancer cells, semi-quantitative assessment of total cellular lipid content levels, and detection of lipid saturation. Raman imaging of single-cell lipid metabolism phenotypes offers exciting new possibilities for tumor research, including advanced imaging capabilities and biorthogonal-labeled Raman Tag. Additionally, the present paper discusses the theoretical foundations and applications of coherent anti-Stokes Raman spectroscopy, stimulated Raman spectroscopy, and novel multimodal single-cell phenotypes chemical imaging in lipid metabolism research, which have opened vast new opportunities for diagnosing and treating cancer.

Bombyx mori nucleopolyhedrovirus (BmNPV) accounts for ∼70% of annual disease losses in sericulture, demanding rapid, accurate, and field-compatible diagnostics. We propose Fourier Transform Infrared (FT-IR) spectroscopy combined with a tiered chemometric approach for early BmNPV detection using minimal hemolymph volumes. A total of 840 samples collected over five days post-infection were analyzed. Unsupervised Principal Component Analysis (PCA) showed extensive overlap among infection stages, indicating limited inherent separability. Linear Discriminant Analysis (LDA) still produced misclassifications. The k-Nearest Neighbors (kNN) algorithm, applied after baseline correction and mean centering, achieved 99.29% accuracy, with sensitivity and specificity both exceeding 99%. Highest performance was obtained with Partial Least Squares Discriminant Analysis (PLS-DA) using first-derivative and mean center signal correction preprocessing, yielding perfect classification (sensitivity, specificity, and accuracy = 1.000) in both calibration and prediction sets. Notably, PLS-DA latent variable score plots (LV1: 19.71%; LV2: 23.44%) tracked the temporal progression of infection, revealing stage-specific metabolic shifts-from early energy mobilization to late-stage systemic collapse. This work demonstrates that FT-IR spectroscopy, when integrated with an optimized chemometric pipeline, provides a rapid, low-cost, and highly accurate diagnostic platform amenable to real-world sericulture. By enabling early detection and infection staging, the method supports timely intervention, effective disease containment, and enhanced sustainability in silk production.