Search Results (441)

Corrosion, a naturally occurring process leading to the deterioration of metallic materials, demands diligent detection for quality control and the preservation of metal-based objects, especially within industrial contexts. Traditional techniques for corrosion identification, including ultrasonic testing, radiographic testing, and magnetic flux leakage, necessitate the deployment of expensive and bulky equipment on-site for effective data acquisition. An unexplored alternative involves employing lightweight, conventional camera systems and state-of-the-art computer vision methods for its identification. In this work, we propose a complete system for semi-automated corrosion identification and mapping in industrial environments. We leverage recent advances in three-dimensional (3D) point-cloud-based methods for localization and mapping, with vision-based semantic segmentation deep learning techniques, in order to build semantic–geometric maps of industrial environments. Unlike the previous corrosion identification systems available in the literature, which are either intrusive (e.g., electrochemical testing) or based on costly equipment (e.g., ultrasonic sensors), our designed multi-modal vision-based system is low cost, portable, and semi-autonomous and allows the collection of large datasets by untrained personnel. A set of experiments performed in relevant test environments demonstrated quantitatively the high accuracy of the employed 3D mapping and localization system, using a light detection and ranging (LiDAR) device, with less than $0.05$ m and $0.02$ m average absolute and relative pose errors. Also, our data-driven semantic segmentation model was shown to achieve $70\%$ precision in corrosion detection when trained with our pixel-wise manually annotated dataset. Full article

Background/Objectives: Identification and characterization of broadly neutralizing monoclonal antibodies from individuals exposed to SARS-CoV-2, either by infection or vaccination, can inform the development of next-generation vaccines and antibody therapeutics with pan-SARS-CoV-2 protection. Methods: Through single B cell sorting and RT-PCR, monoclonal antibodies (mAbs) were isolated from a donor who experienced a BA.5 or BF.7 breakthrough infection after three doses of inactivated vaccines. Their binding and neutralizing capacities were measured with ELISA and a pseudovirus-based neutralization assay, respectively. Their epitopes were mapped by competition ELISA and site-directed mutation. Results: Among a total of 67 spike-specific mAbs cloned from the donor, four mAbs (KXD643, KXD652, KXD681, and KXD686) can neutralize all tested SARS-CoV-2 variants from wild-type to KP.3. Moreover, KXD643, KXD652, and KXD681 belong to a clonotype encoded by IGHV5-51 and IGKV1-13 and recognize the cryptic and conserved RBD-8 epitope on the receptor-binding domain (RBD). In contrast, KXD686 is encoded by IGHV1-69 and IGKV3-20 and targets a conserved epitope (NTD Site iv) outside the antigenic supersite (NTD Site i) of the N-terminal domain (NTD). Notably, antibody cocktails containing these two groups of mAbs can neutralize SARS-CoV-2 more potently due to synergistic effects. In addition, bispecific antibodies derived from KXD643 and KXD686 demonstrate further improved neutralizing potency compared to antibody cocktails. Conclusions: These four mAbs can be developed as candidates of pan-SARS-CoV-2 antibody therapeutics through further antibody engineering. On the other hand, vaccines designed to simultaneously elicit neutralizing antibodies towards RBD-8 and NTD Site iv have the potential to provide pan-SARS-CoV-2 protection. Full article

Strong earthquake activity along fault zones can lead to the displacement of geomorphic units such as gullies and terraces while preserving earthquake event data through changes in sedimentary records near faults. The quantitative analysis of these characteristics facilitates the reconstruction of significant earthquake activity history along the fault zone. Recent advancements in acquisition technology for high-precision and high-resolution topographic data have enabled more precise identification of displacements caused by fault activity, allowing for a quantitative assessment of the characteristics of strong earthquakes on faults. The 1920 Haiyuan earthquake, which occurred on the Haiyuan fault in the northeastern Tibetan Plateau, resulted in a surface rupture zone extending nearly 240 km. Although clear traces of surface rupture have been well preserved along the fault, debate regarding the maximum displacement is ongoing. In this study, we focused on two typical offset geomorphic sites along the middle segment of the Haiyuan fault that were previously identified as having experienced the maximum displacement during the Haiyuan earthquake. High-precision geomorphologic images of the two sites were obtained through unmanned aerial vehicle (UAV) surveys, which were combined with light detection and ranging (LiDAR) data along the fault zone. Our findings revealed that the maximum horizontal displacement of the Haiyuan earthquake at the Shikaguan site was approximately 5 m, whereas, at the Tangjiapo site, it was approximately 6 m. A cumulative offset probability distribution (COPD) analysis of high-density fault displacement measurements along the ruptures indicated that the smallest offset clusters on either side of the Ganyanchi Basin were 4.5 and 5.1 m long. This analysis further indicated that the average horizontal displacements of the Haiyuan earthquake were approximately 4–6 m. Further examination of multiple gullies and geomorphic unit displacements at the Shikatougou site, along with a detailed COPD analysis of dense displacement measurements within a specified range on both sides, demonstrated that the cumulative displacement within 30 m of this section of the Haiyuan fault exhibited at least five distinct displacement clusters. These dates may represent the results of five strong earthquake events in this fault segment over the past 10,000–13,000 years. The estimated magnitude, derived from the relationship between displacement and magnitude, ranged from M_w 7.4 to 7.6, with an uneven recurrence interval of approximately 2500–3200 years. Full article

In this study, we evaluated in situ click chemistry as a platform for discovering boronic acid-based β-lactamase inhibitors (BLIs). Unlike conventional drug discovery approaches requiring multi-step synthesis, protection strategies, and extensive screening, the in situ method can allow for the generation and identification of potent β-lactamase inhibitors in a rapid, economic, and efficient way. Using KPC-2 (class A carbapenemase) and AmpC (class C cephalosporinase) as templates, we demonstrated their ability to catalyse azide-alkyne cycloaddition, facilitating the formation of triazole-based β-lactamase inhibitors. Initial screening of various β-lactamases and boronic warheads identified compound 3 (3-azidomethylphenyl boronic acid) as the most effective scaffold for kinetic target-guided synthesis (KTGS). KTGS experiments with AmpC and KPC-2 yielded triazole inhibitors with K_i values as low as 140 nM (compound 10a, AmpC) and 730 nM (compound 5, KPC-2). Competitive inhibition studies confirmed triazole formation within the active site, while an LC–MS analysis verified that the reversible covalent interaction of boronic acids did not affect detection of the in situ-synthesised product. While KTGS successfully identified potent inhibitors, limitations in amplification coefficients and spatial constraints highlight the need for optimised warhead designs. This study validates KTGS as a promising strategy for BLI discovery and provides insights for further refinement in fighting β-lactamase-mediated antibiotic resistance. Full article

We applied UAV (Unmanned Aerial Vehicle) and ALS (Airborne Laser Scanning) remote sensing methods to identify terrain obstacles encountered during timber extraction in the skidding process with the aim of proposing accessibility solutions to the inner parts of forest stands using skidding trails. At the Vítovický žleb site, located east of Brno in the South Moravian Region of the Czech Republic, we analysed the accuracy of digital terrain models (DTMs) created from UAV LiDAR (Light Detection and Ranging), RGB (Red–Green–Blue) UAV, ALS data taken on site and publicly available LiDAR data DMR 5G (Digital Model of Relief of the Czech Republic, 5th Generation, based on airborne laser scanning, providing pre-classified ground points with an average density of 1 point/m²). UAV data were obtained using two types of drones: a DJI Mavic 2 mounted with an RGB photogrammetric camera and a GeoSLAM Horizon laser scanner on a DJI M600 Pro hexacopter. We achieved the best accuracy with UAV technologies, with an average deviation of 0.06 m, compared to 0.20 m and 0.71 m for ALS and DMR 5G, respectively. The RMSE (Root Mean Square Error) values further confirm the differences in accuracy, with UAV-based models reaching as low as 0.71 m compared to over 1.0 m for ALS and DMR 5G. The results demonstrated that UAVs are well-suited for detailed analysis of rugged terrain morphology and obstacle identification during timber extraction, potentially replacing physical terrain surveys for timber extraction planning. Meanwhile, ALS and DMR 5G data showed significant potential for use in planning the placement of skidding trails and determining the direction and length of timber extraction from logging sites to forest roads, primarily due to their ability to cover large areas effectively. Differences in the analysis results obtained using GIS (Geographic Information System) cost surface solutions applied to ALS and DMR 5G data DTMs were evident on logging sites with terrain obstacles, where the site-specific ALS data proved to be more precise. While DMR 5G is based on ALS data, its generalised nature results in lower accuracy, making site-specific ALS data preferable for analysing rugged terrain and planning timber extractions. However, DMR 5G remains suitable for use in more uniform terrain without obstacles. Thus, we recommend combining UAV and ALS technologies for terrain with obstacles, as we found this approach optimal for efficiently planning the logging-transport process. Full article

Chemical warfare agents (CWAs), including hydrogen cyanide (AC), 2-[fluoro(methyl)phosphoryl]oxypropane (GB), 3-[fluoro(methyl)phosphoryl]oxy-2,2-dimethylbutane (GD), ethyl S-(2-diisopropylaminoethyl) methylphosphonothioate (VX), and di-2-chloroethyl sulfide (HD), pose a great threat to public safety; therefore, it is important to develop sensing technology for CWAs. Herein, a sensor array consisting of 24 metal oxide semiconductor (MOS)-based MEMS sensors with good gas sensing performance, a simple device structure (0.9 mm × 0.9 mm), and low power consumption (<10 mW on average) was developed. The experimental results show that there are always several sensors among the 24 sensors that show good sensing performance in relation to each CWA, such as a relatively significant response, a broad detection range (AC: 5.8–89 ppm; GB: 0.04–0.47 ppm; GD: 0.06–4.7 ppm; VX: 9.978 × 10⁻⁴–1.101 × 10⁻³; HD: 0.61–4.9 ppm), and a low detection limit that is lower than the immediately dangerous for life and health (IDLH) level of the five CWAs. This indicates that these sensors can meet the needs for qualitative detection and can provide an early warning regarding low concentrations of CWAs. In addition, features were extracted from the initial kinetic characteristics and dynamic change characteristics of the sensing response. Finally, principal component analysis (PCA) and machine learning algorithms were applied for CWA classification. The obtained PCA plots showed significant differences between groups, and the narrow neural network among the machine learning algorithms achieves a prediction accuracy of nearly 100.0%. In summary, the proposed MOS-based MEMS sensor array driven by pattern recognition algorithms can be integrated into portable devices, showing great potential and practical applications in the rapid, in situ, and on-site detection and identification of CWAs. Full article

Background/Objectives: Mild traumatic brain injury (mTBI) is a leading cause of pediatric emergency department visits, particularly among children under three years old. Although computed tomography (CT) is the gold standard for diagnosing intracranial injuries, its use in young children poses radiation risks. Identifying reliable clinical indicators that justify CT imaging is essential for optimizing both patient safety and resource utilization. Objective: This study aimed to evaluate CT findings in children under three years of age with mTBI and no focal neurological deficits, as well as to identify clinical predictors associated with skull fractures and intracranial injuries. Methods: A retrospective analysis was conducted on 224 children under 36 months who presented with mTBI to a tertiary pediatric hospital from July 2019 to July 2024. Demographic data, injury mechanisms, clinical presentation and CT findings were evaluated. Univariate and multivariate regression analyses were performed to identify risk factors associated with skull fractures and intracranial injuries. Results: Falls accounted for 96.4% of injuries, with the majority occurring from heights of 0.5–1 m. The parietal region was the most frequently affected site (38%). Skull fractures were present in 46% of cases and were primarily linear (92.8%). Intracranial hematomas were identified in 13.8% of cases, while brain edema was observed in 7.6%. Significant predictors of skull fractures included age under 12 months (p < 0.001), falls from 0.5–1 m (p = 0.005), somnolence (p = 0.030), scalp swelling (p = 0.001) and indentation of the scalp (p = 0.016). Parietal bone involvement was the strongest predictor of both skull fractures (OR = 7.116, p < 0.001) and intracranial hematomas (OR = 4.993, p < 0.001). Conversely, frontal bone involvement was associated with a lower likelihood of fractures and hematomas. Conclusions: The findings highlight key clinical indicators that can guide decision-making for CT imaging in children with mTBI. Infants under 12 months, falls from moderate heights and parietal bone involvement significantly increase the risk of fractures and intracranial injuries. A more refined diagnostic approach could help reduce unnecessary CT scans while ensuring the timely identification of clinically significant injuries. Full article

As a key global food reserve, rice disease detection technology plays an important role in promoting food production, protecting ecological balance and supporting sustainable agricultural development. However, existing rice disease identification techniques face many challenges, such as low training efficiency, insufficient model accuracy, incompatibility with mobile devices, and the need for a large number of training datasets. This study aims to develop a rice disease detection model that is highly accurate, resource efficient, and suitable for mobile deployment to address the limitations of existing technologies. We propose the Transfer Layer iRMB-YOLOv8 (TLI-YOLO) model, which modifies some components of the YOLOv8 network structure based on transfer learning. The innovation of this method is mainly reflected in four key components. First, transfer learning is used to import the pretrained model weights into the TLI-YOLO model, which significantly reduces the dataset requirements and accelerates model convergence. Secondly, it innovatively integrates a new small object detection layer into the feature fusion layer, which enhances the detection ability by combining shallow and deep feature maps so as to learn small object features more effectively. Third, this study is the first to introduce the iRMB attention mechanism, which effectively integrates Inverted Residual Blocks and Transformers, and introduces deep separable convolution to maintain the spatial integrity of features, thus improving the efficiency of computational resources on mobile platforms. Finally, this study adopted the WIoUv3 loss function and added a dynamic non-monotonic aggregation mechanism to the standard IoU calculation to more accurately evaluate and penalize the difference between the predicted and actual bounding boxes, thus improving the robustness and generalization ability of the model. The final test shows that the TLI-YOLO model achieved 93.1% precision, 88% recall, 95% mAP, and a 90.48% F1 score on the custom dataset, with only 12.60 GFLOPS of computation. Compared with YOLOv8n, the precision improved by 7.8%, the recall rate improved by 7.2%, and mAP@.5 improved by 7.6%. In addition, the model demonstrated real-time detection capability on an Android device and achieved efficiency of 30 FPS, which meets the needs of on-site diagnosis. This approach provides important support for rice disease monitoring. Full article

In response to the challenges of low accuracy in traditional pepper blight identification under natural complex conditions, particularly in detecting subtle infections on early-stage leaves, stems, and fruits. This study proposes a multi-site pepper blight disease image recognition algorithm based on YOLOv8, named MSPB-YOLO. This algorithm effectively locates different infection sites on peppers. By incorporating the RVB-EMA module into the model, we can significantly reduce interference from shallow noise in high-resolution depth layers. Additionally, the introduction of the RepGFPN network structure enhances the model’s capability for multi-scale feature fusion, resulting in a marked improvement in multi-target detection accuracy. Furthermore, we optimized CIOU to DIOU by integrating the center distance of bounding boxes into the loss function; as a result, the model achieved an impressive mAP@0.5 score of 96.4%. This represents an enhancement of 2.2% over the original algorithm’s mAP@0.5. Overall, this model provides effective technical support for promoting intelligent management and disease prevention strategies for peppers. Full article

Introduction: The increase in the rates of multidrug-resistant bacteria in healthcare environments has been recognized as a global public health problem. In view of the scarcity of data on the neonatal population, this study aimed to provide information on the genotypic and epidemiological characteristics of Gram-negative microorganisms isolated from colonization and infection sites in neonates admitted to a tertiary university center of high complexity. Methods: Enterobacterales and non-fermenting Gram-negative bacilli previously collected in a prospective cohort study were submitted to genotypic identification, detection of extended-spectrum β-lactamases (ESBL), carbapenemases and biofilm production, detection of specific virulence markers in Pseudomonas aeruginosa, and typing by pulsed-field gel electrophoresis. Results: The data found here revealed higher rates of infection by Klebsiella spp. and Serratia marcescens that caused bloodstream infection and pneumonia, respectively. In this study, high biofilm production was observed, with 95.0% of Enterobacterales and 100% of non-fermenting Gram-negative bacilli being producers. Most of the P. aeruginosa isolates carried pathogenicity factors such as alginate, hemolytic phospholipase C, exotoxin A, and rhamnolipids. The phenotypic analysis of ESBL revealed that 16 (5.3%) isolates produced these enzymes. Four of these isolates (66.7%) carried the CTX-M-9 gene, three (50%) carried the TEM gene, and one (16.7%) was positive for the SHV and CMY-2 genes. Univariate and multivariate Cox regression analyses were used to identify risk factors for colonization and infection by Gram-negative microorganisms. The results of multivariate analysis revealed that biofilm production by these microorganisms was associated with the persistence of colonization by the same pathogen in the newborn and increased by 75% the daily probability of the newborn developing infection. The production of ESBL also increased the daily probability of infection by 46.8 times. Conclusions: Enterobacterales showed average biofilm production, while the majority of non-fermenting Gram-negative bacilli were strong producers. The present data increase our knowledge of the molecular epidemiology of important Enterobacterales species, with emphasis on ESBL-producing Enterobacter cloacae and Klebsiella pneumoniae with emerging epidemiological potential in the neonatal intensive care unit of a tertiary university hospital. Furthermore, the results highlight the need for the monitoring and implementation of control measures and for restricting the use of broad-spectrum antibiotics. Full article

The identification of the most pertinent site parameters to classify soils in terms of their amplification of seismic ground motions is still of prime interest to earthquake engineering and codes. This study investigates many options for improving soil classifications in order to reduce the deviation between “exact” predictions using wave propagation and the method used in seismic codes based on amplification (site) factors. To this end, an exhaustive parametric study is carried out to obtain nonlinear responses of sets of 324 clay and sand columns and to constitute the database for neuronal network methods used to predict the regression equations of the amplification factors in terms of seismic and site parameters. A wide variety of parameters and their combinations are considered in the study, namely, soil depth, shear wave velocity, the stiffness of the underlaying bedrock, and the intensity and frequency content of the seismic excitation. A database of AFs for 324 nonlinear soil profiles of sand and clay under multiple records with different intensities and frequency contents is obtained by wave propagation, where soil nonlinearity is accounted for through the equivalent linear model and an iterative procedure. Then, a Generalized Regression Neural Network (GRNN) is used on the obtained database to determine the most significant parameters affecting the AFs. A second neural network, the Radial Basis Function (RBF) network, is used to develop simple and practical prediction equations. Both the whole period range and specific short-, mid-, and long-period ranges associated with the AFs, F_a, F_v, and F_l, respectively, are considered. The results indicate that the amplification factor of an arbitrary soil profile can be satisfactorily approximated with a limited number of sites and the seismic record parameters (two to six). The best parameter pair is (PGA; resonance frequency, f₀), which leads to a standard deviation reduction of at least 65%. For improved performance, we propose the practical triplet $\left({PGA;V}_{s30};f_0\right)$ with V_s30 being the average shear wave velocity within the upper 30 m of soil below the foundation. Most other relevant results include the fact that the AFs for long periods (F_l) can be significantly higher than those for short or mid periods for soft soils. Finally, it is recommended to further refine this study by including additional soil parameters such as spatial configuration and by adopting more refined soil models. Full article

This paper presents a novel deep learning model, RLANet, based on the ResNet-LSTM-Multihead Attention module, designed for processing and classifying one-dimensional spectral data. The model incorporates ResNet, LSTM, and attention mechanisms, omitting the traditional fully connected layer to significantly reduce the parameter count while maintaining global spectral feature extraction. This design enables RLANet to be lightweight and computationally efficient, making it suitable for real-time applications, especially in resource-constrained environments. Furthermore, this study introduces the Kepler Optimization Algorithm (KOA) for hyperparameter tuning in deep learning, demonstrating its superiority over the traditional Bayesian optimization (BO) in achieving optimal hyperparameter configurations for complex models. Experimental results indicate that the RLANet model successfully achieves accurate identification of three types of engine oil products and their mixtures, with classification accuracy approaching one. Compared to conventional deep learning models, it features a significantly reduced parameter count of only 0.09 M, enabling the deployment of compact devices for rapid on-site classification of oil spill types. Furthermore, relative to traditional machine learning models, RLANet demonstrates a lower sensitivity to preprocessing methods, with the standard deviation of classification accuracy maintained within approximately 0.001, thereby underscoring its excellent end-to-end analytical capabilities. Moreover, even under a strong noise interference at a signal-to-noise ratio of 15 dB, its classification performance declines by only 19% relative to the baseline, attesting to its robust resilience. These results highlight the model’s potential for practical deployment in end-to-end online spectral analysis, particularly in resource-constrained hardware environments. Full article

Human cysteine desulfurase (NFS1) participates in numerous critical cellular processes, including iron–sulfur (Fe-S) cluster biosynthesis and tRNA thiolation. NFS1 overexpression has been observed in a variety of cancers, and thus it has been considered a promising anti-tumor therapeutic target. To date, however, no inhibitors targeting NFS1 have been identified. Here, we report the identification of the first potent small-molecule inhibitor (Compound 53, PubChem CID 136847320) of NFS1 through a combination of virtual screening and biological validation. Compound 53 exhibited good selectivity against two other pyridoxal phosphate (PLP)-dependent enzymes. Treatment with Compound 53 inhibited the proliferation of lung cancer (A549) cells (IC₅₀ = 16.3 ± 1.92 μM) and caused an increase in cellular iron levels due to the disruption of Fe-S cluster biogenesis. Furthermore, Compound 53, in combination with 2-AAPA, an inhibitor of glutathione reductase (GR) that elevates cellular reactive oxygen species (ROS) levels, further suppressed the proliferation of A549 cells by triggering ferroptotic cell death. Additionally, the key residues involved in the binding of the inhibitor to the active center of NFS1 were identified through a combination of molecular docking and site-directed mutagenesis. Taken together, we describe the identification of the first selective small-molecule inhibitor of human NFS1. Full article

The study of marine and terrestrial palynomorphs in fluviomarine environments has been successfully used in combination with different geophysical approaches to understand high-resolution relative sea-level oscillations and to reconstruct the environmental changes affecting estuaries and adjacent inland ecosystems. However, erosion during the postglacial marine transgression frequently causes sedimentary discontinuities or may lead to the redeposition of ancient upland sediments, including secondary, recycled and rebedded pollen. Therefore, a robust seismic and chronological control of the sedimentary facies is essential. In addition, studies of modern pollen sedimentation and its relationship to contemporaneous vegetation are valuable for obtaining a more realistic interpretation of the sedimentary evidence. To explore the significance of the experimental evidence obtained and to support the interpretation of sedimentary records from the same basin, we analysed a large set of modern pollen data from the Ría de Vigo (NW Iberia). The pollen samples derived from different sedimentary environments were compared with the local and regional vegetation cover. Pollen evidence from the various limnetic systems studied allows the identification of major vegetation types in the basin. However, in all the cases, the reconstructed relative pollen contributions of each vegetation unit are often distorted by the overrepresentation of certain anemophilous pollen types, the underrepresentation of some entomophilous species, and the specific taphonomy of each site of sedimentation. The ability of the seabed pollen evidence to represent the modern deciduous and alluvial forests, as well as the saltmarsh vegetation onshore, increases in the shallowest points of the ria (shallower than −10 m). Conversely, pastures and crops are better represented at intermediate depths (shallower than −30 m), while scrubland vegetation is better represented in samples at more than 20 m below modern sea level. It is concluded that shallow seabed pollen can provide information on the main elements of the modern vegetation cover of the emerged basin, including the main elements of the vegetation cover. However, the selection of the most suitable subtidal sites for coring, combined with pollen data from several environmental contexts, is critical for achieving an accurate reconstruction of the changing conditions of the emerged basin over time. Full article

The rhizosphere microbiota of arid plants plays a crucial role in adaptation to environmental stress. However, few studies have characterized microorganisms associated with Agave species and their contribution to resilience against salinity and drought. This study aimed to isolate and characterize halotolerant bacteria from the rhizosphere of Agave potatorum Zucc from two different sites and evaluate their in vitro Na⁺ sequestration, desiccation resistance, and phytohormone production. These traits were compared with those of halotolerant bacteria isolated from a highly saline soil at a third site. Bacteria were obtained through serial dilutions and cultured on R2A plates supplemented with varying NaCl concentrations. The most efficient Na⁺-sequestering isolates underwent an 18-day desiccation assay, and their production of indole-3-acetic acid (IAA) and gibberellic acid (GA₃) was quantified. Among the 48 halotolerant isolates obtained, 7 (SM1, SM10, SPM5, SM7, SM19, VZ9, and SPM1) exhibited the highest Na⁺ sequestration efficiency. Among these isolates, SM1 exhibited the highest in vitro Na⁺ sequestration capacity (10.74 μg L⁻¹, p < 0.05). SM1 and SPM1 demonstrated the greatest desiccation resistance, at 88.39% and 83.05%, respectively. Additionally, SM7 produced the highest levels of IAA (13.69 μg mL⁻¹, p < 0.05), while SM1 exhibited the highest GA₃ production (1285.38 μg mL⁻¹, p < 0.05). Based on these characteristics, isolates SPM1 and SM1 exhibited the highest efficiency in tolerating drought and salinity stress. However, isolate SPM1 may colonize the rhizosphere of A. potatorum more effectively, likely due to its adaptation as a native isolate to the edaphic and environmental conditions in which this agave thrives. Molecular identification confirmed that the isolates belong to the genera Kosakonia, Priestia, Streptomyces, Bacillus, Stutzerimonas, Pseudomonas, and Exiguobacterium. This study highlights the diversity of halotolerant bacteria in the rhizosphere of A. potatorum and their potential as bioinoculants for enhancing soil fertility and restoring degraded soils. Full article