Dayalbagh Educational Institute

Anthropogenic biomass burning (ABB) is a major yet least understood source of atmospheric pollution, with significant implications for air quality, visibility, atmospheric chemistry, the Earth’s radiation budget and biogeochemical cycling. This study examines the chemical composition of aerosols during biomass burning (BB) and non-biomass burning (NBB) periods to assess their impact on aerosol composition and their contribution in generating oxidative stress. Sampling of PM₂.₅ and PM₁₀ was carried out at two contrasting sites one of the highly polluted cities of Indo-Gangetic Plain (IGP), Agra: Dayalbagh (suburban) and Rambagh (urban), from December, 2022 to November, 2024. The samples were analyzed for dicarboxylic acids (DCAs), sugars, organic and elemental carbon (OC/EC), water-soluble inorganic ions (WSIIs), metals and polycyclic aromatic hydrocarbons (PAHs). Moderate Resolution Imaging Spectroradiometer (MODIS) and Fire Information for Resource Management System (FIRMS) data identified intense fire hotspots over northwestern India (Punjab–Haryana) during BB, coinciding with elevated particulate concentrations. The mean concentration of PM₂.₅ and PM₁₀ increased considerably during BB (PM₂.₅: 101.1 ± 74.4 µg m⁻³ at Dayalbagh, 120.0 ± 57.5 µg m⁻³ at Rambagh; PM₁₀: 161.2 ± 67.3 and 184.4 ± 61.7 µg m⁻³, respectively) compared to NBB. The Pearson correlation analysis showed that carbonaceous species and biomass tracers (DCAs, K⁺, levoglucosan) showed strong positive correlations (r > 0.8), confirming the influence of agricultural residue and biofuel combustion. Secondary ions (SO₄²⁻, NO₃⁻, NH₄⁺) displayed enhanced interrelationships (r = 0.75–0.77) during NBB, indicating increased secondary aerosol formation. The oxidative potential, assessed using the dithiothreitol (DTT) assay, exhibited markedly higher activity in fine particles (r2 = 0.70 and 0.79) and during BB (DTTv: 18.2 ± 10.1 pmol min⁻¹ m⁻³ at Dayalbagh, 17.1 ± 15.0 pmol min⁻¹ m⁻³ at Rambagh) compared to NBB (7.0 ± 4.0 and 10.0 ± 2.4 pmol min⁻¹ m⁻³, respectively). The correlation between DTTv and biomass tracers (oxalic acid (C2), malonic acid (C3), adipic acid (C6), levoglucosan (Lev), arabitol (Arab); r = 0.50-0.83), OC/EC (r = 0.52-0.70), metals (Fe, Mn, K, Na, Ni, Mn, Cr; r = 0.52-0.64) during BB period suggest higher redox activity during this period. Positive Matrix Factorization (PMF) identified four dominant sources: biomass burning (31–35%), vehicular emissions (26–30%), industrial activities (18–22%), and crustal dust (10–13%). Thus, BB emissions significantly enhanced PM loadings and oxidative potential, posing elevated health risks. The findings highlight the synergistic role of biomass combustion and urban emissions in amplifying aerosol toxicity and degrading air quality over the IGP.

Effective suppression of cardiac tachyarrhythmias and shortening of action potential duration (APD) remain major challenges in cardiac optogenetics due to the non-selectivity of ion-conducting channelrhodopsins (ChRs). We present a comprehensive theoretical analysis of optogenetic suppression of electrical activity and temporally precise shortening of APD in human ventricular cardiomyocytes (HVCMs) expressed with WiChR and HcKCR1 potassium (K⁺)-selective ChRs that exhibit reversal potentials close to the diastolic membrane potential of the targeted cardiac cells. Our simulations show that K⁺-selective ChRs provide more efficient hyperpolarizing control compared to cation-selective variants. Our computational simulations show that WiChR exhibits action-potential suppression at very low irradiance (4 × 10⁻² mW/mm²), while HcKCR1 could not achieve full suppression at 10 mW/mm². This makes WiChR a much stronger and more effective option for optical inhibition in human ventricular cardiomyocytes. At irradiance of 1 × 101 mW/mm², our simulations show that WiChR hyperpolarizes the membrane to -83.14 mV, closely approaching the K⁺ equilibrium potential, whereas HcKCR1 reached − 82.39 mV. The results are important for the treatment of cardiac arrhythmias and long QT syndrome. This work provides mechanistic insight into low-power optogenetic control of cardiac electrophysiology and offers quantitative predictions to guide future in vitro and in vivo investigations aimed at developing safer and more precise alternatives to electrical pacing and defibrillation technologies.

Tryptamine, an indoleamine, is recognized as an internal signaling molecule that enables plants to coordinate their responses under environmental stresses. However, its functional role in cyanobacteria remains largely unexplored. Therefore, the present study investigates the regulatory effects of exogenous tryptamine (10 µM) on mitigating salt (NaCl) induced toxicity in nitrogen-fixing cyanobacterium Anabaena sp. PCC7120. Salt (NaCl) at 60 mM markedly impaired growth, reduced photosynthetic pigment contents, suppressed photosynthetic activity including PSII photochemistry while increasing the respiratory rate in Anabaena sp. Further, NaCl also caused a significant rise in oxidative stress biomarkers (O₂·¯, H₂O₂, and malondialdehyde equivalents; MDA). Exogenously applied tryptamine enhanced the salt stress tolerance, leading to an increase in growth, photosynthetic pigment content, photosynthetic efficiency, along with an appreciable rise in antioxidative enzymes (SOD, POD, CAT and GST) that reduced the oxidative damage. Overall, the study signifies the role of tryptamine as a growth promoter substance to salt stressed Anabaena PCC7120 by enhancing the antioxidants defense system that results in better photosynthetic efficiency, increased biomass and increased synthesis of protein. Hence, supplementation of tryptamine can increase the fertility of salt-affected paddy fields, which is a low-input, sustainable agricultural practice. Although tryptamine, a tryptophan-derived monoamine, has been shown to influence cyanobacterial growth and stress responses, there is a need for a deeper understanding of its functions; this study aims to explore its underlying regulatory mechanisms.

The insulators used for the transmission lines are important for ensuring that the electrical power system is stable through insulation. The insulators tend to degenerate, either through the growth of cracks, fouling, or deterioration, leading to costly power failures. The process through which the insulators are checked manually is time-consuming and tends to lack accuracy, thereby creating the need for the use of accurate automated systems for detection. This study introduced a multi-label deep learning approach that can identify a maximum of three insulator defects that occur at the same time for each image. The study adopted six different architectures for convolutional neural networks (CNN) models, namely ResNet50, VGG16, MobileNetV2, InceptionV3, XceptionV1, and finally EfficientNetB3, and then fine-tuned them for evaluation on the IFDD dataset by utilizing five-fold cross-validation for accuracy and reliability. The study adopted different parameters, namely Exact Match Accuracy, Hamming Loss, Jaccard Similarity, Precision, Recall, and F1-Score, for evaluating the accuracy of the proposed model. Among them, the highest balance for accuracy and loss was obtained by the model called 'EfficientNetB3,' whose accuracy was 96.25% and loss was 1.46%. The model performed better than the other architectures for each fold. For better explainability, the model used Gradient-weighted Class Activation Mapping (Grad-CAM) to highlight the input image area from which the model made decisions. The outcomes of the experimental analysis prove that the model maximizes both accuracy and efficiency. Hence, it can be conclusively ascertained that it is appropriate for incorporation into drone inspection mechanisms for power transmission infrastructure.

Medicinal plants are integral to traditional healthcare practices like Ayurveda, offering therapeutic value at minimal side effects. The identification of medicinal plant species, however, presents the biggest challenge as the species exhibit morphological similarity, few botanists, and the risk of misplacement. In response, the current study puts forth an extensive deep learning-based transfer learning framework for the automatic classification of medicinal leaves. The BDMediLeaves dataset, made available for the public, consisting of 38,606 augmented images of ten species, was adopted in the development and validation of the framework. Five pre-trained CNN models—VGG16, DenseNet201, InceptionResNetV2, MobileNetV2, and Xception—were experimented in an overall experiment setup to evaluate their performance based on accuracy, precision, recall, and F1-score as the basis of their performance. The MobileNetV2-based proposed model delivered the maximum accuracy of 99.43%, performing better than the other models while ensuring computational efficacy. The study’s novelty lies in establishing a systematic and deployment-ready benchmarking framework for medicinal plant identification, demonstrating that lightweight CNNs can achieve state-of-the-art accuracy suitable for mobile and web-based healthcare applications.