Search

Anterior cruciate ligament (ACL) tears are prevalent knee injures, particularly among active individuals. Accurate and timely diagnosis is essential for determining the optimal treatment strategy and assessing patient prognosis. Various previous studies have demonstrated the successful application of deep learning techniques in the field of medical image analysis. This study aimed to develop a deep learning model for detecting ACL tears in knee magnetic resonance Imaging (MRI) to enhance diagnostic accuracy and efficiency. The proposed model consists of three main modules: a Dual-Scale Data Augmentation module (DDA) to enrich the training data on both the spatial and layer scales; a selective group attention module (SG) to capture relationships across the layer, channel, and space scales; and a fusion module to explore the inter-relationships among various perspectives to achieve the final classification. To ensure a fair comparison, the study utilized a public dataset from MRNet, comprising knee MRI scans from 1250 exams, with a focus on three distinct views: axial, coronal, and sagittal. The experimental results demonstrate the superior performance of the proposed model, termed SGNET, in ACL tear detection compared with other comparison models, achieving an accuracy of 0.9250, a sensitivity of 0.9259, a specificity of 0.9242, and an AUC of 0.9747.

ABSTRACT: Milk-clotting enzyme (MCE) is the essential active agents in dairy processing. The traditional MCE is mainly obtained from animal sources, in which calf rennet is the most widely used in cheese industry. Traditional MCE substitute is becoming necessary due to its limited production and increased cheese consumption. A novel traditional MCE substitute was produced from Bacillus velezensis DB219 in this study. The DB219 MCE exhibited a notable specific activity of 6,110 Soxhlet units/mg and 3.16-fold purification yield with 28.87% recovery through ammonium sulfate fractionation and DEAE-Sepharose Fast Flow. The purified DB219 MCE was a metalloprotease with a molecular weight of 36 kDa. The DB219 MCE was weak acid resistance and stable at pH 6.0 to 10.0 and temperature

Limited resources for edge computing platforms in airborne and spaceborne imaging payloads prevent using complex image processing models. Model pruning can eliminate redundant parameters and reduce the computational load, enhancing processing efficiency on edge computing platforms. Current challenges in model pruning for remote-sensing object detection include the risk of losing target features, particularly during sparse training and pruning, and difficulties in maintaining channel correspondence for residual structures, often resulting in retaining redundant features that compromise the balance between model size and accuracy. To address these challenges, we propose the L1 reweighted regularization (L1RR) pruning method. Leveraging dynamic and self-adaptive sparse modules, we optimize L1 sparsity regularization, preserving the model’s target feature information using a feature attention loss mechanism to determine appropriate pruning ratios. Additionally, we propose a residual reconstruction procedure, which removes redundant feature channels from residual structures while maintaining the residual inference structure through output channel recombination and input channel recombination, achieving a balance between model size and accuracy. Validation on two remote-sensing datasets demonstrates significant reductions in parameters and floating point operations (FLOPs) of 77.54% and 65%, respectively, and a 48.5% increase in the inference speed on the Jetson TX2 platform. This framework optimally maintains target features and effectively distinguishes feature channel importance compared to other methods, significantly enhancing feature channel robustness for difficult targets and expanding pruning applicability to less difficult targets.

Quinolone antibiotic residues present a high risk for food safety and have gained widespread attention. Nanomaterials based sensors have the advantages of on-site application, sensitivity, and small sample volume requirement and has attracted significant interest in on-site detection technology research for quinolone antibiotics. This paper introduces the application of optical sensors including fluorescence, colorimetry, SERS, ICA and electrochemical sensors based on diverse nanomaterials in quinolone antibiotics detection and summarizes the characteristics of diverse sensors. The application of nanomaterials including quantum dots and upconversion in optical sensors, as well as those of carbon and metal nanomaterials and redox media in electrochemical sensors are reviewed and evaluated. The objective of this work is to provide new concepts for antibiotics detection in food and the development of sensors.

Target detection is a critical task in interpreting aerial images. Small target detection, such as vehicles, is challenging. Different lighting conditions affect the accuracy of vehicle detection. For example, vehicles are difficult to distinguish from the background in red, green, blue (RGB) images under low illumination conditions. In contrast, under high-illumination conditions, the color and texture of vehicles are not significantly different in thermal infrared (TIR) images. To improve the accuracy of vehicle detection under various illumination conditions, we propose an adaptive multimodal feature fusion and cross-modal vehicle index (AFFCM) model for vehicle detection. Based on the single-stage object detection model, AFFCM uses RGB and TIR images. It comprises three parts: 1) the softpooling channel attention (SCA) mechanism calculates the cross-modal feature weights of the RGB and TIR features using a fully connected layer during global weighted pooling; 2) we design a multimodal adaptive feature fusion (MAFF) module based on the cross-modal feature weights derived from the SCA mechanism; the MAFF selects features with high weight, compresses redundant features with low weight, and performs adaptive fusion using a multiscale feature pyramid; and 3) a cross-modal vehicle index is established to extract the target area, suppress complex background information, and minimize false alarms in vehicle detection. The mean average precision (mAP) on the Drone Vehicle dataset is 14.44% and 5.02% higher than that obtained using only RGB or TIR images. The mAP is 2.63% higher than that of state-of-the-art methods that utilize RGB and TIR images.

Abstract The milk-clotting enzyme (MCE) is a crucial ingredient in cheese manufacture. Due to the limits of traditional MCE, finding viable substitute is a pressing issue. This study aims to isolate and identify a wild strain with high milk-clotting activity (MCA) and low proteolytic activity (PA) and optimize the fermentation conditions for MCE production. A strain of Bacillus velezensis DB219 with high MCA/PA value (9.2) was isolated from dairy soil (Wuchang, Heilongjiang, China) and identified through 16S rRNA from 40 strains. The optimal wheat bran, carbon, nitrogen, inoculum size, volume and initial pH were 60 g/L, soluble starch 12.5 g/L, corn steep liquor 3 g/L, 5%, 40 mL and 6.15, respectively for improving DB219 MCE production through single factor experiment. The wheat bran concentration, corn steep liquor concentration and volume were the most critical factor and their changed range was determined through Plackett–Burman design and the steepest ascent/descent experiments. The response surface analysis experiment of three factors and three levels was conducted by Box–Behnken design. The theoretical optimal fermentation conditions for DB219 MCE were as follows: wheat bran concentration 60.14 g/L, soluble starch 12.5 g/L, corn steep liquor 3 g/L, inoculum size 5%, volume 40.08 mL and initial pH 6.15. DB219 MCE achieved the maximal MCA (3164.84 SU/mL) that was 101.9% of the predicted value (3104.49 SU/mL) and 4.3-fold higher than the control.

Vibrio parahaemolyticus is a halophilic and heat-labile gram-negative bacterium and is the most prevalent foodborne bacterium in seafood. In order to develop a rapid and sensitive method for detecting the foodborne pathogenic bacterium Vibrio parahaemolyticus, an aptamer-modified magnetic nanoparticle and an aptamer-modified upconversion nanoparticle were synthesised and used as a capture probe and a signal probe, respectively. The aptamer-modified magnetic nanoparticle, V. parahaemolyticus cell, and aptamer-modified upconversion nanoparticle formed a sandwich-like complex, which was rapidly separated from a complex matrix using a magnetic force, and the bacterial concentration was determined by fluorescence intensity analysis. The results showed that the fluorescence intensity signal correlated positively with the concentration of V. parahaemolyticus in the range of 3.2 × 102 to 3.2 × 105 CFU/mL, with a linear equation of y = 296.40x − 217.67 and a correlation coefficient of R2 = 0.9610. The detection limit of the developed method was 4.4 CFU/mL. There was no cross-reactivity with other tested foodborne pathogens. This method is highly specific and sensitive for the detection of V. parahaemolyticus, and can achieve the qualitative detection of this bacterium in a complex matrix.

The increasing demand for olefin polymers has prompted extensive research on catalysis, processes, and structural characterization to explore their economic potential. In this study, N,N'-bis(salicylidene)-1,2-ethylenediamine (salen) was synthesized, complexed with Fe(III) ion, and the complex was bridged by simple condensation reaction with five different dicarboxylic acids to produce dinuclear complexes. Different methods, such as 1H NMR, FT-IR, UV–Vis, and EDX elemental analysis, were used to analyze and identify the ligand and complexes. Spectroscopic results supported by molecular modeling studies, revealed that the complexes had an octahedral geometry. Structural optimization of the ligand and metal complexes was performed using Avogadro (version 1.2.0) and the Universal Force Field molecular mechanics level. Density functional theory was used to conduct computational calculations on the stability and reactivity of the zeolite-supported and unsupported complexes as catalysts for the polymerization of ethylene. Interestingly, the supported complexes were found to exhibit lower stability and higher reactivity than the neat metal complexes. As such, the synthesized complexes are promising catalysts for the production of polyethylene, and their use could significantly reduce the cost of industrial polyethylene production.

Deep convolution networks have been widely used in remote sensing target detection for various applications in recent years. Target detection models with many parameters provide better results but are not suitable for resource-constrained devices due to their high computational cost and storage requirements. Furthermore, current lightweight target detection models for remote sensing imagery rarely have the advantages of existing models. Knowledge distillation can improve the learning ability of a small student network from a large teacher network due to acceleration and compression. However, current knowledge distillation methods typically use mature backbones as teacher and student networks are unsuitable for target detection in remote sensing imagery. In this article, we propose a target detection model distillation (TDMD) framework using feature transition and label registration for remote sensing imagery. A lightweight attention network is designed by ranking the importance of the convolutional feature layers in the teacher network. Multiscale feature transition based on a feature pyramid is utilized to constrain the feature maps of the student network. A label registration procedure is proposed to improve the TDMD model's learning ability of the output distribution of the teacher network. The proposed method is evaluated on the DOTA and NWPU VHR-10 remote sensing image datasets. The results show that the TDMD achieves a mean Average Precision (mAP) of 75.47% and 93.81% on the DOTA and NWPU VHR-10 datasets, respectively. Moreover, the model size is 43% smaller than that of the predecessor model (11.8 MB and 11.6 MB for the two datasets).

Due to advances in remote sensing satellite imaging and image processing technologies and their wide applications, intelligent remote sensing satellites are facing an opportunity for rapid development. The key technologies, standards, and laws of intelligent remote sensing satellites are also experiencing a series of new challenges. Novel concepts and key technologies in the intelligent hyperspectral remote sensing satellite system have been proposed since 2011. The aim of these intelligent remote sensing satellites is to provide real-time, accurate, and personalized remote sensing information services. This article reviews the current developments in new-generation intelligent remote sensing satellite systems, with a focus on intelligent remote sensing satellite platforms, imaging payloads, onboard processing systems, and other key technological chains. The technological breakthroughs and current defects of intelligence-oriented designs are also analyzed. Intelligent remote sensing satellites collect personalized remote sensing data and information, with real-time data features and information interaction between remote sensing satellites or between satellites and the ground. Such developments will expand the use of remote sensing applications beyond government departments and industrial users to a massive number of individual users. However, this extension faces challenges regarding privacy protection, societal values, and laws regarding the sharing and distribution of data and information.

Abstract Cyanidin‐3‐O‐glucoside (C3G) is a kind of water‐soluble pigment widely existing in many plants. It has strong antioxidant and anti‐inflammatory activities. However, C3G cannot exist stably for a long time because of the phenolic hydroxyl groups in its structure. Liposome technology could improve the stability and bioavailability of compounds. Based on our previous studies, C3G liposomes prepared by ethanol injection method have a certain stability in two weeks of storage. In this study, THP‐1 macrophages treated with C3G and C3G liposomes can reduce the levels of inflammatory‐related factors, such as tumor necrosis factor‐a (TNF‐a), interleukin (IL)‐1β, IL‐6, and IL‐8, stimulated by lipopolysaccharide (LPS). Further studies showed that the LPS induction could increase the level of phosphorylated nuclear transcription factor NF‐κB and phosphorylated IkBa, while C3G and C3G liposomes could inhibit the expression of phosphorylated proteins. Moreover, C3G and C3G liposomes could protect macrophages from apoptosis. In conclusion, C3G prepared by liposome technology exhibits anti‐inflammatory activity, which provides a theoretical basis for the food industry to study functional food.

Background: Predicting outcomes of surgical aortic valve replacement (AVR) in patients with chronic severe aortic regurgitation (AR) and markedly reduced left ventricular (LV) function remains a challenge. This study aimed to explore the preoperative echocardiographic index that could predict the recovery of LV systolic function after surgery in patients with chronic severe AR and reduced left ventricular ejection fraction (LVEF). Methods: The study group consisted of 50 patients diagnosed with chronic severe AR (>6 months) and significantly reduced LVEF (18~35%, average 26.2 ± 5.3%). Low-dose dobutamine stress echocardiography (DSE) was performed before surgery. Only patients with an absolute increase in LVEF ≥8% during DSE were referred for surgical AVR. During following up (over six months to one year after surgery), the patients were divided into two groups by postoperative LVEF (> or ≤40%). DSE- and speckle tracking echocardiography (STE)-derived LV functional parameters were compared between groups to identify predictors of post-operative improvement in LVEF. Results: A total of 38 patients underwent AVR. One patient died before discharge. Post-surgical LV size and LVEF improved markedly after surgery in all patients (n = 37). Pre-surgical LV end-systolic diameter, baseline global longitudinal strain (GLS) and peak GLS were better in the group with LVEF >40% (n = 18; p < 0.05, t test). Baseline GLS and peak GLS correlated moderately with post-surgery LVEF (R = –0.581, p < 0.001; R = –0.596, p < 0.001; respectively). Logistic regression analysis demonstrated baseline GLS and peak GLS were the independent predictors of post-surgery improvement of LVEF. Peak GLS had the highest prediction value (area under the curve = 0.895, sensitivity and specificity: 89.5% and 77.8%, respectively), with a cutoff value of –9.4%. Conclusions: This study shows that STE combined with DSE can provide sensitive quantitative indices for predicting improvement of LV systolic function after AVR in patients with chronic severe AR and significantly decreased LVEF.

Abstract The oral infections were mainly caused by Streptococci and Staphylococcus aureus. Antibiotic therapies can eliminate these harmful bacteria. However, it can break beneficial microbes and lead to the persistence of resistant strains. The objective of our study was to select potential probiotic strains for the prevention of oral bacterial infections and evaluate their potential probiotic properties in oral cavity. AR113 (Lactobacillus plantarum) and AR340 (Lactobacillus paracasei) with significantly antimicrobial β-hemolytic streptococci and Staphylococcus aureus activity were isolated from Chinese pickle through agar well diffusion assay. Through the analyses of probiotic properties in antibiofilm, lysozyme and hydrogen peroxide tolerance, bacterial surface properties, adherence ability, tooth degradation and anti-inflammatory activity, the AR113 and AR340 showed anti-adhesion activity of 45.2–71.1% and 20.3–56.8% against β-hemolytic streptococci and 15.4–52.6% and 30.7–65.9% against Staphylococcus aureus, respectively, at different concentration. The two strains with high hydrophobicity, autoaggregation and survival rate adhered strongly to FaDu cells. AR113 and AR340 exhibited low calcium released from teeth (0.04 μg/mL and 0.03 μg/mL, respectively). ELISA analysis showed that AR113 and AR340 significantly inhibited the LPS-induced increase of NO and TNF-α expression. Strains-fermented skim milk inhibited the growth of β-hemolytic streptococci or Staphylococcus aureus. AR113 and AR340 were considered as probiotic candidates because of their higher antibacterial activity against some oral pathogenic bacteria, no potential of primitive cariogenicity. These candidates were expected as new probiotics with potential oral health benefits and no harmful effects.

Object detection is a focal point in remote sensing applications. Remote sensing images typically contain a large number of small objects and a wide range of orientations across objects. This results in great challenges to small object detection approaches based on remote sensing images. Methods directly employ channel relations with equal weights to construct information features leads to inadequate feature representation in complex image small object detection tasks. Multiscale detection methods improve the speed and accuracy of detection, while small objects themselves contain limited information, and the features are easily lost following down-sampling. During the detection, the feature images are independent across scales, resulting in a discontinuity at the detection scale. In this article, we propose the multiscale context and enhanced channel attention (MSCCA) model. MSCCA employs PeleeNet as the backbone network. In particular, the feature image channel attention is enhanced and the multiscale context information is fused with multiscale detection methods to improve the characterization ability of the convolutional neural network. The proposed MSCCA method is evaluated on two real datasets. Results show that for 512 × 512 input images, MSCCA was able to achieve 80.4% and 94.4% mAP on the DOTA and NWPU VHR-10, respectively. Meanwhile, the model size of MSCCA is 21% smaller than that of its predecessor. MSCCA can be considered as a practical lightweight oriented object detection model in remote sensing images.

The antioxidant activity of broccoli extracts and its purified compounds (polyphenols and purified sulforaphane (PSF)) was investigated. Broccoli extracts and SF protected B16 cells from H2O2-induced oxidative stress through increasing total antioxidant capacity (T-AOC) and activities of NADPH dehydrogenase quinone1 (NQO1) and catalase (CAT), and decreasing content of malondialdehyde (MDA). Western blot and qRT-PCR revealed that the expression of nuclear factor erythroid 2-related factor 2 (Nrf2), NQO1 and heme oxygenase 1 (HMOX-1) was increased. Broccoli extracts was better than SF in improving T-AOC, antioxidant enzyme activity (especially, NQO1) and reducing MDA content. Polyphenols and PSF were separated from broccoli extracts by high-speed countercurrent chromatography (HSCCC). They could significantly increase T-AOC and NQO1 activities (P

Object detection based on remote sensing imagery has become increasingly popular over the past few years. Unlike natural images taken by humans or surveillance cameras, the scale of remote sensing images is large, which requires the training and inference procedure to be on a cutting image. However, objects appearing in remote sensing imagery are often sparsely distributed and the labels for each class are imbalanced. This results in unstable training and inference. In this paper, we analyze the training characteristics of the remote sensing images and propose the fusion of the aggregated-mosaic training method, with the assigned-stitch augmentation and auto-target-duplication. In particular, based on the ground truth and mosaic image size, the assigned-stitch augmentation enhances each training sample with an appropriate account of objects, facilitating the smooth training procedure. Hard to detect objects, or those in classes with rare samples, are randomly selected and duplicated by the auto-target-duplication, which solves the sample imbalance or classes with insufficient results. Thus, the training process is able to focus on weak classes. We employ VEDAI and NWPU VHR-10, remote sensing datasets with sparse objects, to verify the proposed method. The YOLOv5 adopts the Mosaic as the augmentation method and is one of state-of-the-art detectors, so we choose Mosaic (YOLOv5) as the baseline. Results demonstrate that our method outperforms Mosaic (YOLOv5) by 2.72% and 5.44% on 512 × 512 and 1024 × 1024 resolution imagery, respectively. Moreover, the proposed method outperforms Mosaic (YOLOv5) by 5.48% under the NWPU VHR-10 dataset.

N,N-dimethylformamide (DMF) is a commonly-used solvent in industry and pharmaceutics for extracting acetylene and fabricating polyacrylonitrile fibers. It is also a starting material for a variety of intermediates such as esters, pyrimidines or chlordimeforms. However, after being used, DMF can be form 5–25% spent liquors (mass fraction) that are difficult to recycle with distillation. From the point of view of energy-efficiency and environment-friendliness, an emergent separation technology, pervaporation, is broadly applied in separation of azeotropic mixtures and organic–organic mixtures, dehydration of aqueous–organic mixtures and removal of trace volatile organic compounds from aqueous solutions. Since the advances in membrane technologies to separate N,N-dimethylformamide solutions have been rarely reviewed before, hence this review mainly discusses the research progress about various membranes in separating N,N-dimethylformamide aqueous solutions. The current state of available membranes in industry and academia, and their potential advantages, limitations and applications are also reviewed.

The nascent graph representation learning has shown superiority for resolving graph data. Compared to conventional convolutional neural networks, graph-based deep learning has the advantages of illustrating class boundaries and modeling feature relationships. Faced with hyperspectral image (HSI) classification, the priority problem might be how to convert hyperspectral data into irregular domains from regular grids. In this regard, we present a novel method that performs the localized graph convolutional filtering on HSIs based on spectral graph theory. First, we conducted principal component analysis (PCA) preprocessing to create localized hyperspectral data cubes with unsupervised feature reduction. These feature cubes combined with localized adjacent matrices were fed into the popular graph convolution network in a standard supervised learning paradigm. Finally, we succeeded in analyzing diversified land covers by considering local graph structure with graph convolutional filtering. Experiments on real hyperspectral datasets demonstrated that the presented method offers promising classification performance compared with other popular competitors.

This present work aimed to investigate the influence of CO2 on HCl removal using calcined Ca−Mg−Al layered double hydroxides (CaMgAl-LDHs) at medium−high temperature (400−800 °C) in a fixed-bed reactor. It was revealed that a moderate CO2 concentration (~6%) in the flue gas of the municipal solid-waste incinerators could reduce the HCl capacity of the CaMgAl-layered double oxides (CaMgAl-LDOs). The highest capacity for HCl removal was observed over the CaMgAl-LDOs at 600 °C. However, sintering was also detected when the reaction temperature was below the calcination temperature (600 °C). Moreover, the decreasing HCl adsorption capacity of CaMgAl-LDOs was attributed to the existence of CO2 in the flue gas, which could efficiently inhibit the decomposition of carbonates as well as the conversion into metal chloride during the HCl removal process.

Dietary broccoli is anti-inflammatory. Past studies have typically investigated raw broccoli, even though most consumers prefer cooked broccoli, where the plant myrosinase is inactivated by heat, resulting in failure of formation of the anti-inflammatory bioactive compound sulforaphane (SF). This study compareed efficacy of lightly cooked broccoli (CB) containing greatly diminished myrosinase activity, with raw broccoli (RB), in mitigating colitis in dextran sulfate sodium (DSS)-treated mice. Male C57BL/6 mice were fed for two weeks on a 10% RB, 10% CB or control diet, all based on the AIN-93M diet. Half (n = 9) of each group received drinking water, half received 2.5% DSS in water for one week, starting from Day 7 of the diet. Even with far less plant myrosinase activity, CB was essentially as effective as RB in lessening damage by DSS, evidenced by decreased disease activity index, attenuated colon length shrinkage, less endotoxin (lipopolysaccharide) leakage into blood, and less severe colon lesions as assessed by histopathology. mRNA expression of pro-inflammatory cytokines indicated that broccoli anti-inflammatory action may be through inhibition of the IL-6 trans-signaling pathway, as evidenced by reversal of the DSS-increased expression of IL-6, CCR2 and vascular cell adhesion molecule 1 (VCAM-1).

Macroporous TiO2 photocatalyst was synthesized by a facile nanocasting method using polystyrene (PS) spherical particles as the hard template. The synthesized photocatalyst was characterized by transmission electron microscope (TEM), scanning electron microscopy (SEM), thermogravimetry-differential thermogravimetry (TG-DTG), X-ray diffraction (XRD), and N2-sorption. TEM, SEM, and XRD characterizations confirmed that the macroporous TiO2 photocatalyst is composed of anatase phase. The high specific surface area of 87.85 m2/g can be achieved according to the N2-sorption analysis. Rhodamine B (RhB) was chosen as probe molecule to evaluate the photocatalytic activity of the TiO2 catalysts. Compared with the TiO2 materials synthesized in the absence of PS spherical template, the macroporous TiO2 photocatalyst sintered at 500°C exhibits much higher activity on the degradation of RhB under the UV irradiation, which can be assigned to the well-structured macroporosity. The macroporous TiO2 material presents great potential in the fields of environmental remediation and energy conversion and storage.

Genotoxicity assessment is of great significance in drug safety evaluation, and microarray is a useful tool widely used to identify genotoxic stress responsive genes. In the present work, by using oligonucleotide microarray in an in vivo model, we identified an unknown gene BC005512 (abbreviated as BC, official full name: cDNA sequence BC005512), whose expression in mouse liver was specifically induced by seven well-known genotoxins (GTXs), but not by non-genotoxins (NGTXs). Bioinformatics revealed that BC was a member of the GLN family of murine endogenous retrovirus (ERV). However, the relationship to genotoxicity and the cellular function of GLN are largely unknown. Using NIH/3T3 cells as an in vitro model system and quantitative real-time PCR, BC expression was specifically induced by another seven GTXs, covering diverse genotoxicity mechanisms. Additionally, dose-response and linear regression analysis showed that expression level of BC in NIH/3T3 cells strongly correlated with DNA damage, measured using the alkaline comet assay,. While in p53 deficient L5178Y cells, GTXs could not induce BC expression. Further functional studies using RNA interference revealed that down-regulation of BC expression induced G1/S phase arrest, inhibited cell proliferation and thus suppressed cell growth in NIH/3T3 cells. Together, our results provide the first evidence that BC005512, a member from GLN family of murine ERV, was responsive to DNA damage and involved in cell growth regulation. These findings could be of great value in genotoxicity predictions and contribute to a deeper understanding of GLN biological functions.