Visually impaired individuals were encountering serious issues in achieving safe and independent mobility both indoors and outdoors. All the assistive technologies available in the current state of the art, including white canes, guide dogs, smart canes, and advanced AI-based systems, were providing only partial assistance due to their respective limitations such as poor obstacle detection, high cost, low adaptability, and improper usage in continuous real-time applications. An assistive computer vision-based obstacle detection and navigation system for visually impaired individuals utilizing deep learning is proposed to address these problems. It employed an SSD-MobileNet model for accurate obstacle detection and identification of commonly found objects such as chairs, beds, humans, and vehicles. Audio and haptic feedback will be employed in real-time to assist in safe mobility. The proposed solution will be economical, adaptable, and efficient for real-time applications. The experimental outcomes revealed that the system improved the dependability of the navigation, user safety, and independence in terms of mobility of the visually impaired towards a better quality of life

Upconverting nanoparticles (UCNPs) offer significant potential for highly sensitive biosensing due to their background-free detection and excellent photostability. However, their intrinsically low upconversion efficiency limits their usage in practical applications. To reach ultra-high sensitivity with a simplified readout system, the excitation field, and consequently the upconversion luminescence, can be strengthened by using plasmon-based mechanisms. For qualitative detection, the optical field inhomogeneities in plasmon-enhanced upconversion do not generally cause practical limitations, but in quantitative detection, a homogeneous signal across the biosensor surface is required. With UCNPs this is particularly challenging since the upconversion efficiency is highly non-linear depending on the excitation field intensity. The motivation for this work is to compare the digital and analog readouts for quantitative biosensing applications. UCNPs are imaged via diverging surface plasmon polaritons, and the images are analyzed using both methods: digital detection counts individual UCNPs, whereas analog detection quantifies upconversion luminescence by integrating the pixel intensity across the image. Our results show that analog detection exhibits greater variability than digital detection. This indicates that digital detection is likely to provide better repeatability when employing plasmon-enhanced UCNPs for quantitative biosensing.

In a laser, the control of its spectral emission depends on the physical dimensions of the optical resonator, restricting it to a set of discrete cavity modes at specific frequencies1-4. Without modifying the optical cavity, this results in substantial gaps in the obtainable laser emission spectrum, as well as a fixed repetition rate, limiting the device's usability in various experiments and applications where a considerable degree of tunability is required in the spectral or temporal domain. Here we overcome this fundamental limit by demonstrating a monolithic semiconductor laser5-7 with a continuously tunable repetition rate from 4 GHz up to 16 GHz, by using a microwave driving signal that induces a spatiotemporal gain modulation along the entire laser cavity8,9, generating intracavity mode-locked pulses10-13 with a continuously tunable group velocity14. At the output, frequency combs15,16 with continuously tunable mode spacings are generated in the frequency domain, and coherent pulse trains with continuously tunable repetition rates are generated in the time domain17. Our results pave the way for fully tunable chip-scale lasers and frequency combs, which will be advantageous for use in a diverse variety of fields, from fundamental studies to applications such as high-resolution and dual-comb spectroscopy18,19.

Approximate computing is a rising technique aimed at developing arithmetic circuits that minimize power usage, resource consumption, and latency for error-tolerant applications, enabling faster and more efficient circuits suitable for resource-constrained environments. Square root computation is vital in hardware design for image and signal processing, but it’s resource-intensive and power-hungry. Optimizing it can boost power efficiency and performance. This work presents a hardware-efficient, fast and low-power Taylor series-based optimal unsigned square rooter (TSOSQR) designed for approximate square root computation of a 2<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i>-bit unsigned integer using addition and shift operations. Compared to a precise restoring array-based square rooter architecture, the suggested square rooter design uses 79% fewer resources, operates 56% faster, and provides a 81% increase in power savings. These designs are implemented on an Artix-7 FPGA using Verilog-HDL, verified through Xilinx Vivado simulation, and additionally synthesized at the ASIC level using the Cadence Genus Compiler targeting a standard-cell 45nm CMOS technology node. Extensive simulations and a detailed comparison of the proposed design against state-of-the-art approaches show that the TSOSQR effectively balances accuracy and hardware efficiency, significantly reducing power consumption and latency. The paper demonstrates the superior performance of the proposed square rooter in Sobel edge detection, image contrast enhancement, and K-means clustering, with an IoT-enabled FPGA edge vision node design flow, highlighting its effectiveness and advantage over existing approximate designs.

Optimized GPU usage in high energy physics applications

High-fidelity three-dimensional (3D) wave solvers accurately simulate acoustic wave propagation in complex vocal tract geometries but are computationally demanding, limiting their usage in real-time applications. In contrast, low-dimensional models are efficient but limited to cylindrical tracts, neglecting higher-order modes in their frequency responses. This paper introduces a lightweight lumped two-dimensional (2.5D) solver that combines the efficiency of low-dimensional models with the accuracy of 3D approaches to model straight tracts constrained to mid-sagittal symmetry. Like 3D, the 2.5D model captures transverse wave propagation and accounts for higher-order modes. We validate the model by comparing its transfer functions and pressure distributions against those of a conventional two-dimensional (2D) solver and a high-fidelity 3D finite element model for six straight tract geometries of varying complexity. This analysis demonstrates the abilities and limitations of the proposed method. The results show that the 2.5D solver closely matches the 3D model's transfer functions up to 12 kHz, with correlation coefficients exceeding 0.8 for symmetric tracts. For asymmetric geometries, it still performs significantly better than the 2D model. Additionally, the 2.5D solver achieves over two orders of magnitude computational speed-up compared to the 3D model, offering a better trade-off between accuracy and efficiency for vocal tract acoustic modeling.

Data-based evidence of normative gait in healthy adults (18-65 years): Demographics, methodology, and data availability-A scoping review.

Drug response prediction (DRP), accounting for the diverse biological characteristics of cancer types that affect sensitivity or resistance to treatment, is crucial for anticancer drug selection and discovery. Although numerous deep learning models for DRP have been developed, it has not been investigated whether these models can maintain reliable predictive power when applied to omics datasets not used for training, which is crucial for real-world applications. Moreover, they have not been rigorously examined through systematic benchmark tests to investigate the relationships between model architecture and performance. We present a new model, GCNPath, that exploits both graph convolution network (GCN) architectures and pathway-based feature reduction of gene expression data, which have previously been shown to improve DRP model performance. In comprehensive benchmark tests utilizing multiple cell omics platforms, GCNPath shows robust and competitive performances compared with state-of-the-art models including prediction of unseen drugs and the ability to overcome batch effects across various RNA datasets. This study demonstrates the validity of the pathway-level GCN model in DRP and suggests directions for developing DRP models with improved adaptability to diverse and heterogeneous datasets and enhanced usability for practical applications.

Native Staphylococcus aureus protein A exhibits strong affinity to the Fc and VH regions of human IgG1, IgG2, and IgG4, making it a valuable tool for monoclonal antibody (mAb) purification. However, its low stability under conditions such as increased alkaline concentrations during cleaning-in-place (CIP), protease exposure, thermal stress, and shear forces limits its usability for large-scale industrial applications. Recombinant Protein A (rProtein A) can be modified to improve key properties, including alkaline stability. In this study, we present targeted modifications to the C domain of native Protein A, evaluating multimeric variants for structural and functional improvements. The selected variant demonstrated extremely high stability after 60 h incubation at 0.5 M NaOH by maintaining more than &gt;90% initial dynamic binding capacity (DBC) and up to 80% DBC after 40 h in 1.0 M NaOH. However, the most impressive result obtained was the stability of the ligand in 1.5 M NaOH, retaining 80% DBC after 22 h and 60% DBC after 40 h. To the best of our knowledge, this is the first time that such high alkaline stability is reported for a rProtein A. To assess its application in monoclonal antibody purification, the optimized rProtein A ligand was immobilized on agarose resin and tested in chromatography processes. The resulting chromatography resin functionalized with the CmZmb ligand (now commercialized by Sunresin, China under the name of rProtein A Seplife Suno) exhibited a high dynamic binding capacity of 70 mg/mL, minimal ligand leaching under operational conditions (~15 ppm), and extended lifecycle performance (88% DBC retained after 120 purification cycles with 0.5 M NaOH CIP), making it well-suited for industrial-scale applications.

Artificial intelligence (AI) and machine learning concepts have long been explored in radiology, with deep learning accelerating progress in the past decade [1]. Improvements in graphics processing units (GPU), the core hardware component for AI and software developments, have led to significant innovation in this field, including pulmonary nodule detection, tumor segmentation, and pulmonary embolus detection to name a few [2, 3]. This technology has also allowed for improvements in workflow efficiency, decreasing scan times and improving integration of data from various imaging studies and the electronic medical records of patients to improve diagnostic accuracy and throughput [4]. This comes at a critical time for radiologists, where a marked increase in imaging compounded by fewer available radiologists and technologists has led to many practices being unable to keep up with the demand for imaging studies being performed in ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) [2]. AI usage in the improvement of image quality in MRI has shown promise, using end to end supervised, end to end unsupervised and generative modeling techniques for image optimization [5]. Through the use of convolutional neural networks (CNN) and U-net deep learning-based segmentation architecture, AI has demonstrated the capability to maintain high quality imaging with fewer data points compared to conventional reconstruction techniques. This can reduce scan time, an important factor when dealing with patient motion, and increase signal to noise, improving diagnostic accuracy. Deep learning architecture has also been used in MRI to generate high field strength imaging from low field strength inputs, potentially increasing access for advanced MRI imaging techniques to places they cannot afford 1.5 or 3 T strength MRI magnets [6]. Deep learning U-net techniques have also helped with segmentation of anatomy and tumors, with increased automation compared to other methods with training these models though still requiring human supervision [5]. AI–based post-processing continues to expand the capabilities of body MRI, offering different approaches to improve visualization of pathology and streamline image interpretation. Among body MRI applications, magnetic resonance cholangiopancreatography (MRCP) provides a particularly illustrative example. Emerging deep learning–driven segmentation techniques illustrate the potential of AI in MRCP; for example, to refine anatomic conspicuity by reducing signal overlap and enhancing the depiction of complex ductal structures, as quality of these sequences can be affected by surrounding intervening T2 hyperintense structures and other artifacts [7]. Comparing MRCP sequences in patients to assess for progression of disease processes such as primary sclerosing cholangitis (PSC) can be challenging as it is subjective with reader variability [8]. AI-driven semi-automated tools such as MRCP+ can help get objective quantitative measures, aiding in diagnosis of any potential changes in degree of biliary ductal dilation over time [4, 9]. Advances such as these reflect a broader shift toward algorithmic enhancement of existing datasets, with the goal of improving diagnostic confidence rather than fundamentally altering image acquisition. Several important limitations/pitfalls exist with AI usage in body MRI applications. Conservative labeling approaches favor sensitivity over specificity, which can leave residual signals from nearby structures. Very faint imaging abnormalities may fall below the model's detection threshold, leading to missing pathology. The quality of AI based algorithms depends on the data set from which the algorithms are derived. Limited data sets can lead to misdiagnoses of pathology secondary to anatomic variants, age-related changes, postoperative changes, medical devices, image artifacts, and satisfaction of search [10]. For these reasons, AI based post processed images should be considered a complement to, rather than a replacement for, conventional MRI, with careful review of source images and reconstructions to confirm findings [2]. The future integration of AI into abdominal imaging will depend not only on technical performance but also on thoughtful implementation, transparency of limitations, and continued emphasis on radiologist oversight. Ultimately, the success of these tools will be measured by their ability to support more accurate, patient-centered care while preserving the nuanced judgment that remains central to diagnostic radiology.

Humans are increasingly exposed to "eco-friendly" biodegradable microplastic pollution, whose usage in packaging and medical applications is growing exponentially. The bioplastic polylactic acid (PLA) has recently been demonstrated to release large quantities of oligomeric lactic acid (OLA) nanoplastics causing adverse health effects. No research has reported on intrauterine biodistribution of OLA, and how gestational exposure may impact on early development of the fetus. Here, we reveal that OLA plastics can readily breach the placental barrier and accumulate in various fetal organs in a mouse model. Gestational exposure to environmentally relevant dose of OLA impairs vasculature development, causing intrauterine growth restriction in the pups. Mechanistically, OLA causes blockage of the vascular endothelial growth factor pathway and abnormal physiological development of placenta, which is mediated by the obstruction of transcription factor GATA2 translocation into the nucleus. This study highlights the potential developmental health effect of oligomer nanoparticles released from biodegradable PLA plastic.

Abstract Copper's high electrical and thermal conductivity makes it appealing for use in various industrial products. However, challenges related to its weight and performance in extreme environments limit copper's usage in aerospace applications. Previous research has demonstrated that incorporating multilayer graphene (MLG) on a copper substrate via chemical vapor deposition (CVD) improves the performance of this conductor in high-temperature applications without incurring a weight penalty. Incorporating a high percentage of large-area graphene, needed for better performance, is difficult while fabricating wires. This work shows a method for making high-quality graphene-copper composite wires from a 25 μm and 50 μm copper foils, consolidated into a wire via repeated annealing and roller-drawing reductions. A copper foil wire without graphene is compared to the composite to highlight graphene's benefits. This research correlates the composite's resulting material properties to the microstructure and creation process. The final results suggest that graphene content aids in consolidation, removing one of the primary defects in manufacturing wires from foils. Reducing porosity through improved consolidation prevents early fracture under tensile loading. In addition, the specific conductivity at room temperature for bilayer graphene (BLG), few-layer graphene, and MLG samples was comparable to that of bare copper wire. Graphene content also improves the resulting high-temperature electrical properties by protecting the wire from further oxidation. Based on the data presented in this paper, recommendations are provided for further reducing void defects and enhancing the quality and performance of copper-graphene composite wires.

This study explores the development and evaluation of mycelium-based panels as a sustainable material for creative and installation applications. Grounded in sustainability theory and the increasing demand for eco-friendly alternatives, the research investigates how agricultural waste substrates such as rice straw and sawdust can be transformed into viable structural panels using fungal mycelium as a natural binding agent. The study utilized an exploratory-descriptive design, including substrate preparation, inoculation, molding, incubation, and drying, followed by systematic observation and respondent evaluations. Results revealed panels with natural beige-brown coloration, fibrous texture, moderate density, porosity, and mild earthy odor, indicating successful colonization and material stability. Evaluation through Likert-scale responses showed strong acceptance in terms of performance, reliability, adaptability, aesthetics, and sustainability, highlighting their usability for indoor installations and creative applications. The findings suggest that mycelium panels provide a functional, biodegradable, and visually appealing alternative to conventional panel materials, demonstrating promise for future eco-focused design innovations. The study concludes that mycelium panels offer practical benefits while supporting environmental responsibility, and recommends further refinement of substrate mixtures, mechanical testing, and full-scale installation trials to optimize their potential in sustainable design contexts.

Bigraphs provide a versatile modeling framework by combining spatial and connectivity relationships, making them valuable in graph transformations, model transformations, the design of programming languages, and the simulation of reactive systems. However, practical approaches to bigraph rewriting often rely on constraint satisfaction or SAT-solving techniques, which can be computationally intensive and limit usability in large-scale applications. This article addresses the challenge of efficient bigraph rewriting by leveraging the graph transformation engine GrGen.NET . We propose a novel approach to translate bigraphical rules into SPO-based graph transformation rules, enabling fast bigraph rewriting. Our solution is implemented in the tool BiGGer , which includes a command-line interface and a Java library for seamless integration. Through experimental evaluation, we demonstrate the correctness of our approach and its ability to significantly improve execution efficiency. Additionally, we implemented tracking rules that maintain semantic fidelity during the translation process and enable identity tracing across reactions, adding a feature to bigraphs that extends their applicability beyond static analysis.

The creation of unified, open, secure, reliable, and agile data spaces is essential for collecting, storing, and sharing data in a standardized and accessible manner, promoting data reuse and addressing current interoperability limitations. In this context, this research presents a proof of concept for a unified agronomic data space based on the structured integration of heterogeneous open data sources. The central hypothesis is that the automated acquisition, preprocessing, and harmonization of publicly available agronomic data can significantly improve accessibility, usability, and interoperability for agricultural decision support applications. To this end, a comprehensive analysis of relevant open data sources was conducted, followed by the design and implementation of configurable algorithms for automated data downloading, cleaning, validation, and integration. The proposed approach explicitly addresses key challenges such as heterogeneous data formats, inconsistent spatial and temporal resolutions, missing values, and outlier detection. As a result, a unified access point was developed, providing reliable agronomic information, including (i) preprocessed climatological time series, (ii) crop and phytosanitary data, (iii) high-resolution aerial orthophotography, (iv) remote-sensing imagery, (v) pest-related information, and (vi) time series of major vegetation indices. The proof of concept was implemented for olive groves in the Andalusian region of Spain; however, the methodology is fully transferable to other crops, regions, and institutional contexts where comparable open data sources are available. The results demonstrate the potential of shared agronomic data spaces to enhance data reuse, support scalable analytics, and facilitate interoperable, data-driven agricultural management beyond the specific regional case study.

Low-density materials, for example, aluminium honeycomb sandwich panels, are gaining extensive usage in structural applications due to their high strength-to-weight ratio and energy absorption capability. This paper presents the investigations on the low-velocity impact behavior of panels reinforced with nano-clay particles at different contents (0%, 2%, and 4% by weight), fabricated by the hand layup method. Panels were tested under impact energies of 10, 20, 30, and 40 J, and surface damage was quantified using MATLAB image processing to obtain numerical damage area (DA). Five machine learning (ML) models were developed to predict DA, with the Polynomial Regression (PR) model (R 2 = 0.94) provided a better balance between prediction accuracy and generalization, supported by Cρ (0.90) and LOOCV (R 2 ≈ 0.95) analyses. SHAP, Feature importance and Partial dependence analysis confirmed impact energy (IE) as the most influential factor, followed by nano-clay percentage (NCP). Experimentally, 2% NCP was identified as the optimal reinforcement level, achieving maximum damage reduction 41% at 10 J, decreasing to 23% at 40 J while 4% NCP yielded marginal additional benefits. Within the 2% NCP, the damage reduction decreased from 41% at 10 J to 23% at 40 J, which further confirms that impact energy is the most dominant factor, as derived from SHAP, feature importance, and Partial dependence analysis. The ANN model presented moderate success regarding the reconstruction of damage features with PSNR values in the range of 13.30–14.81 dB and SSIM scores between 0.62 and 0.64. Overall, integrating experimental analysis with ML modeling offers a robust route to predicting and understanding the impact-induced damage in composite structures.

The stability constant (logK1) and reactivity are ultimately the most crucial components to consider during the evaluation and selection of chelators to match with a specific radiometal ion for usage in radiopharmaceutical applications. These components evaluate the thermodynamic stability of the radiometal-chelator complex. Additionally, the effectiveness of chelator in binding with radiometal ions with relatively large atomic radii (e.g., 213Bi3+ and 225Ac3+) coupled with charge-diffuse properties result in weaker metal-ligand interactions, and this poses challenges in chelator development. The (2-[(carboxymethyl)]5-(4-nitrophenyl-1-[4,7,10-tris(carboxymethyl)-1,4,7,10-tPentan-2-yl) amino] acetic acid (3p-C-DEPA) is a new hybrid chelator designed for potential radio-complexation applications in radio-theranostics and preclinical data has shown great promise for this chelating ligand. Hence, this study investigates the stability constant and chemical reactivity descriptors of the complex generated between 3p-C-DEPA and the α-emitting radioisotopes 213Bi3+ and 225Ac3+ as well as the β-emitting particle 177Lu3+ for the first-time using density functional theory (DFT) calculations. The method employs two functional densities, MO6-HF and B3LYP, using the basis set 6-311G(d)/SDD, alongside the continuous solvation models SMD (solvation model density) and COSMO (conductor-like screening model). The interactions of all radiometals with the hybrid chelator 3p-C-DEPA are compared to the benchmark chelator, 1,4,7,10-tetrazacyclodecane-1,4,7,10-tetraacetic acid (DOTA), yielding comprehensive data on the stability constants and based structural features of radiometal-chelator complexes. DFT analysis has shown that the stability of the 3p-C-DEPA chelator complex formation is influenced by the atomic radius of the radiometal and the number of nitrogen and oxygen donors, proving to be effective for Ac3+ and Bi3+, in contrast to Lu3+, which shows lower stability constant values.

Cemented granular materials (CGMs) represent a transitional class of geomaterials where mechanical behavior is governed by the interplay between a discrete granular skeleton and a continuous cementitious matrix. While previous studies have focused on idealized spherical particles, this study aims to quantify the influence of the cement filling ratio (ranging from 10% to 100%) on the mechanical constitutive behavior of CGMs fabricated with large, irregular granitic aggregates (14–20 mm). Unconfined compressive tests and splitting tensile tests were conducted to evaluate the evolution of strength, stiffness, and failure modes. The results reveal a distinct mechanical transition governed by the cement filling ratio (ρm). The elastic modulus and splitting tensile strength exhibited a linear increase with ρm (R2 &gt; 0.95), indicating a direct dependence on the volume fraction of the binding phase. In contrast, the unconfined compressive strength (UCS) and peak strain displayed a bilinear growth pattern with a critical inflection point at ρm = 80%. For the specific irregular granitic aggregate skeleton investigated, this threshold marks the transition from contact-dominated stability to matrix-dominated continuum behavior. Below this threshold, strength gain is limited by the stability of discrete particle contacts; above 80%, the material behaves as a continuum, with UCS increasing rapidly to a maximum of 41.78 MPa at 100% filling. Furthermore, the dispersion of stress–strain responses significantly decreased as ρm exceeded 50%, attributed to the homogenization of stress distribution within the specimen. These findings provide a quantitative basis for optimizing cement usage in ground reinforcement applications, identifying 80% as a critical design threshold.

The generation of indistinguishable single photons is a fundamental requirement for photonic quantum technologies. However, spectral fluctuations, often induced by charge noise in epitaxial quantum dots (QDs), lead to exciton dephasing, thereby limiting their practical usage in quantum applications. We present a straightforward approach to mitigate charge noise-induced decoherence in droplet-etched GaAs QDs embedded in an n-i-p diode structure and integrated deterministically into an electrically contacted circular Bragg grating resonator for emission enhancement. The quantum device allows for the stabilization of the charge environment by applying an external electrical field while producing a photon extraction efficiency of (37±2)%. Hong-Ou-Mandel two-photon interference measurements reveal a strong dependence of the exciton dephasing time and interference visibility on the applied bias, in excellent agreement with our theoretical predictions. Notably, the reduction in visibility from a maximum, charge stabilized corrected value of 97% at the optimum bias point follows an inverse square dependence (∝1/I^{2}) with increasing diode current (I) in the forward direction. Under a quasi-resonant excitation scheme, we achieve a maximum exciton dephasing time (T_{2}^{*}) of approximately (6.8±0.5) ns, reaching nearly the Fourier limit (T_{2}=2T_{1}) without the need for complex echo schemes like Ramsey or Carr-Purcell-Meiboom-Gill sequences. These findings are consistent with theoretical predictions from rate equation modeling and quantum optical analysis, as well as voltage-dependent linewidth measurements, demonstrating optimized electrical control of exciton dephasing.

Background: Thiol compounds are made up of sulfur-containing amino acids and thiohistidine derivatives. It has generated attention for its diverse biological activities. Aim: The goal of this study is to isolate, characterize, and detect the biological activities, molecular docking, toxicity, and histopathological alterations of thiol compounds isolated from sea urchins. Methods: The chemical structure of thiol compounds extracted from the eggs and sperm of sea urchins was verified using elemental analysis, IR, MS, and NMR spectroscopy. In vitro tests include cytotoxicity, anticancer assay, antioxidant, cell migration, anti-inflammatory, and antimicrobial assays. The OECD Guideline 423 was followed when conducting the in vivo toxicity investigation. Results: The chemical formulas for Ovo thiol A and Sperm thiol are C7H11N3O2S and C6H8N3O2S. Ovo thiol A and Sperm thiol exhibited no cytotoxicity to normal HSF cells (IC50 &gt; 1000 µg/mL). However, against MCF-7 cancer cells, Ovo thiol A showed potent anticancer activity (IC50 = 17.35 µg/mL), and Sperm thiol exhibited significant activity (IC50 = 22.25 µg/mL). Ovo thiol A demonstrated higher antioxidant and anti-inflammatory activity than Sperm thiol. No significant antimicrobial or wound healing activities were observed. Docking results indicated favorable interactions of both Ovo thiol A and Sperm thiol with Bcl-2, Keap1, and glycogen synthase kinase-3β. Computational ADME studies predicted favorable oral passive absorption for Ovo thiol A in humans, while Sperm thiol showed low gastrointestinal absorption. In vivo toxicity assessment revealed non-significant changes in liver and kidney biomarkers. Also, the histology of the liver and kidney of mice treated with thiol compounds showed normal structure. Discussion: The biological activities of Ovo Thiol A and Sperm Thiol are related to the unique position of the thiol group in their structure. Conclusion: Ovo thiol A (1-N-Methyl-4-mercaptohistidine) and Sperm thiol (4- mercaptohistidine) are safe chemicals with biological activity such as anticancer, antioxidant, and anti-inflammatory properties, paving the way for their usage in medical applications.