ABSTRACT
Microbial oxidation and the mechanism of Sb(III) are key governing elements in biogeochemical cycling. A novel Sb oxidizing bacterium, Klebsiella aerogenes HC10, was attracted early and revealed that extracellular metabolites were the main fractions driving Sb oxidation. However, linkages between the extracellular metabolite driven Sb oxidation process and mechanism remain elusive. Here, model phenolic and quinone compounds, i.e., anthraquinone-2,6-disulfonate (AQDS) and hydroquinone (HYD), representing extracellular oxidants secreted by K. aerogenes HC10, were chosen to further study the Sb(III) oxidation mechanism. N2 purging and free radical quenching showed that oxygen-induced oxidation accounted for 36.78% of Sb(III) in the metabolite reaction system, while hydroxyl free radicals (·OH) accounted for 15.52%. ·OH and H2O2 are the main driving factors for Sb oxidation. Radical quenching, methanol purification and electron paramagnetic resonance (EPR) analysis revealed that ·OH, superoxide radical (O2â¢-) and semiquinone (SQ-â¢) were reactive intermediates of the phenolic induced oxidation process. Phenolic-induced ROS are one of the main oxidants in metabolites. Cyclic voltammetry (CV) showed that electron transfer of quinone also mediated Sb(III) oxidation. Part of Sb(V) was scavenged by the formation of the secondary Sb(V)-bearing mineral mopungite [NaSb(OH)6] in the incubation system. Our study demonstrates the microbial role of oxidation detoxification and mineralization of Sb and provides scientific references for the biochemical remediation of Sb-contaminated soil.
Subject(s)
Antimony , Oxidation-Reduction , Reactive Oxygen Species , Electron Transport , Antimony/metabolism , Reactive Oxygen Species/metabolismABSTRACT
In present work, blue carbon dots (b-CDs) were derived from ammonium citrate and guanidine hydrochloride, and red carbon dots (r-CDs) were stemmed from malonate, ethylenediamine and mesotetra (4-carboxyphenyl) porphin based on facile hydrothermal method. Eco-friendly ratiometric fluorescence probe was innovatively constructed to effectively measure Hg2+ utilizing b-CDs and r-CDs. The developed probe displayed two typical emission peaks at 450 nm from b-CDs and 650 nm from r-CDs under the excitation at 360 nm. Mercury ion has strong quenching effect on the fluorescence intensity at 450 nm due to the electron transfer process and the fluorescence change at 450 nm was used as the response signal, whereas the fluorescence intensity at 650 nm kept unchangeable which resulted from the chemical inertness between Hg2+ and r-CDs, serving as the reference signal in the sensing system. Under optimal circumstances, this probe exhibited an excellent linearity between the fluorescence response values of ΔF450/F650 and Hg2+ concentrations over range of 0.01-10 µmol/L, and the limit of detection was down to 5.3 nmol/L. Furthermore, this probe was successfully employed for sensing Hg2+ in practical environmental water samples with satisfied recoveries of 98.5%-105.0%. The constructed ratiometric fluorescent probe provided a rapid, environmental-friendly, reliable, and efficient platform for measuring trace Hg2+ in environmental field.
Subject(s)
Carbon , Fluorescent Dyes , Mercury , Quantum Dots , Water Pollutants, Chemical , Mercury/analysis , Carbon/chemistry , Fluorescent Dyes/chemistry , Water Pollutants, Chemical/analysis , Quantum Dots/chemistry , Environmental Monitoring/methods , Spectrometry, Fluorescence/methods , Limit of Detection , FluorescenceABSTRACT
Ultrasonic humidifiers are commonly used in households to maintain indoor humidity and generate a large number of droplets or spray aerosols. However, there have been various health concerns associated with humidifier use, largely due to aerosols generated during operation. Here, we investigated the size distribution, chemical composition, and charged fraction of aerosol particles emitted from commercial ultrasonic humidifiers. Heavy metals in water used for humidifiers were found to be highly enriched in the ultrasonic humidifier aerosols (UHA), with the enrichment factors ranging from 102 to 107. This enrichment may pose health concerns for the building occupants, as UHA concentrations of up to 106 particles/cm3 or 3 mg/m3 were observed. Furthermore, approximately 90% of UHA were observed to be electrically charged, for the first time according to our knowledge. Based on this discovery, we proposed and tested a new method to remove UHA by using a simple electrical field. The designed electrical field in this work can efficiently remove 81.4% of UHA. Therefore, applying this electrical field could be an effective method to significantly reduce the health risks by UHA.
Subject(s)
Aerosols , Humidifiers , Metals, Heavy , Aerosols/analysis , Metals, Heavy/analysis , Air Pollution, Indoor/prevention & control , Air Pollution, Indoor/analysis , Air Pollutants/analysis , Ultrasonics , Environmental Monitoring/methodsABSTRACT
The reduction of nitrobenzene to aniline is very important for both pollution control and chemical synthesis. Nevertheless, difficulties still remain in developing a catalytic system having high efficiency and selectivity for the production of aniline. Herein, it was found that PdO nanoparticles highly dispersed on TiO2 support (PdO/TiO2) functioned as a highly efficient catalyst for the reduction of nitrobenzene in the presence of NaBH4. Under favorable conditions, 95% of the added nitrobenzene (1 mmol/L) was reduced within 1 min with an ultra-low apparent activation energy of 10.8 kJ/mol by using 0.5%PdO/TiO2 as catalysts and 2 mmol/L of NaBH4 as reductants, and the selectivity to aniline even reached up to 98%. The active hydrogen species were perceived as dominant species during the hydrogenation of nitrobenzene by the results of isotope labeling experiments and ESR spectroscopic. A mechanism was proposed as follows: PdO activates the nitro groups and leads to in-situ generation of Pd, and the generated Pd acts as the reduction sites to produce active hydrogen species. In this catalytic system, nitrobenzene prefers to be adsorbed on the PdO nanoparticles of the PdO/TiO2 composite. Subsequently, the addition of NaBH4 results in in-situ generation of a Pd/PdO/TiO2 composite from the PdO/TiO2 composite, and the Pd nanoclusters would activate NaBH4 to generate active hydrogen species to attack the adsorbed nitro groups. This work will open up a new approach for the catalytic transfer hydrogenation of nitrobenzene to aniline in green chemistry.
Subject(s)
Aniline Compounds , Nitrobenzenes , Palladium , Titanium , Nitrobenzenes/chemistry , Aniline Compounds/chemistry , Titanium/chemistry , Hydrogenation , Catalysis , Palladium/chemistry , Models, ChemicalABSTRACT
El programa de Instrucción de la Autorregulación Cognitiva (CSRI) se basa en la instrucción centrada en la estrategia e incluye tres componentes para la mejora del producto textual (enseñanza directa, modelado y práctica entre iguales). Se plantearon como objetivosanalizar si la instrucción con el programa CSRI en un género textual (texto de comparación) conducía a la transferencia espontánea en el producto textual (mejor coherencia, estructura y calidad) en un género no instruido (texto de opinión); y examinar si el orden de los componentes instruccionales presentaba algún efecto. Participaron 126 estudiantes de cuarto de educación primaria que fueron asignados a una de las dos condiciones experimentales en las que se aplicaba el programa CSRI (con diferente secuencia de los componentes instructivos), o a una condición de control en la que se seguía la enseñanza tradicional. Los resultados reflejaron que las dos secuencias de instrucción del programa CSRI mostraban beneficios en la estructura y coherencia del producto textual de los estudiantes en el posttest pero no 8 meses después. Para que los alumnos sean capaces de transferir adecuadamente lo aprendido a géneros textuales no instruidos, necesitan que los profesores les enseñen cómo hacerlo eficazmente.(AU)
Cognitive Self-Regulation Instruction (CSRI) program is a strategy-focused instruction with three instructional components for im-proving students writing product (direct teaching, modelling, and peer-practice). The present study aimed to explore whether the CSRI program leads to spontaneous transfer, improving the writing product (in terms of quality, structure, and text coherence) of an uninstructed genre (opinion text); and to examine whether the order in which the instructional compo-nents were implemented had an effect. A total of 126 students in their 4thyear of primary school participated in the study. They were randomly as-signed to one of two experimental conditions which received the CSRI but differed in the order the instructional components were delivered, or to a control condition which followed the traditional teaching approach. Our findings show that both CSRI sequences produced benefits in terms of greater structure and coherence of the writing product in the opinion text at post-test but not 8months after the intervention. In consequence, for students to be able to adequately transfer strategies to uninstructed text genres, they need teachers to teach them how to do it effectively.(AU)
Subject(s)
Humans , Male , Female , Education, Primary and Secondary , Writing , LearningABSTRACT
Plant leaf diseases are a significant concern in agriculture due to their detrimental impact on crop productivity and food security. Effective disease management depends on the early and accurate detection and diagnosis of these conditions, facilitating timely intervention and mitigation strategies. In this study, we address the pressing need for accurate and efficient methods for detecting leaf diseases by introducing a new architecture called DenseNet201Plus. DenseNet201 was modified by including superior data augmentation and pre-processing techniques, an attention-based transition mechanism, multiple attention modules, and dense blocks. These modifications enhance the robustness and accuracy of the proposed DenseNet201Plus model in diagnosing diseases related to plant leaves. The proposed architecture was trained using two distinct datasets: Banana Leaf Disease and Black Gram Leaf Disease. Through extensive experimentation, we evaluated the performance of DenseNet201Plus in terms of various classification metrics and achieved values of 0.9012, 0.9012, 0.9012, and 0.9716 for accuracy, precision, recall, and AUC for the banana leaf disease dataset, respectively. Similarly, the black gram leaf disease dataset model provides values of 0.9950, 0.9950, 0.9950, and 1.0 for accuracy, precision, recall, and AUC. Compared to other well-known pre-trained convolutional neural network (CNN) architectures, our proposed model demonstrates superior performance in both utilized datasets. Last but not least, we combined the strength of Grad-CAM++ with our proposed model to enhance the interpretability and localization of disease areas, providing valuable insights for agricultural practitioners and researchers to make informed decisions and optimize disease management strategies.
ABSTRACT
The use of solar greenhouses in China is increasing because they permit environmental conditions to be controlled. Studies of the heat transfer processes in the leaves of plants cultivated within solar greenhouses are needed. Here, we studied heat transfer processes in 'Shine Muscat' grapevine leaves under moderate deficit irrigation (MDI), severe deficit irrigation (SDI), and full irrigation (FI) treatments under varying weather conditions. The stomatal conductance, leaf temperature, and transpiration rate of both shade and sun grapevine leaves were measured, and the effects of ambient temperature and relative humidity on these variables were determined. A thermal physics model of the leaves was established to explore the heat dissipation process. On sunny days, the transpiration heat transfer of sun leaves in the MDI, SDI, and FI treatments was 2.62 MJ m-2·day-1, 2.44 MJ m-2·day-1, and 3.86 MJ m-2·day-1and 0.818 MJ m-2·day-1, 0.782 MJ m-2·day-1, and 1.185 MJ m-2·day-1 on rainy days, respectively. There was a significant difference in transpiration heat transfer under fully irrigated and deficit irrigation conditions under different weather conditions. Furthermore, transpiration heat transfer accounted for 41.49 % and 25.03 % of the total heat transfer of sun leaves in the FI treatment and 33.94 % and 29.43 % of the total heat transfer of shade leaves on rainy days, respectively, indicating that relative humidity plays a key role in determining transpiration heat transfer and leaf temperature and that its effect was greater on sun leaves than on shade leaves.
ABSTRACT
The elders in China's rural areas are facing challenges in maintaining agricultural production due to the outflow of rural laborers. The Transfer of land could alleviate the burden of land-based livelihoods for rural elders, but their decisions regarding land transfer are influenced by their social networks within the context of Chinese rural society. This study investigates how social networks impact the willingness of rural elders to transfer land. Using survey data from 782 rural elders in 32 villages across 11 provinces in China, this paper applies multilinear and binary logistic regression models. The results indicate that the willingness of rural elders to transfer land is affected by their social neteork: (1) Internal network scale, network heterogeneity, and frequency of external network relationships have a significantly positive influence on rural elders' willingness to transfer land, while frequency of internal network relationships has a significantly negative influence. (2) There are group differences in the above impacts, and these significant impacts occur only among male elderly individuals aged 60-69 years old or living in central and western regions. (3) Social networks primarily influence rural elders' willingness to transfer land through three mechanisms: information consultation, interpersonal trust, and material resource acquisition. A larger internal social network scale, higher heterogeneity within the network, and more frequent interactions with members of external networks lead to greater access to useful information, higher levels of trust in others, increased material resources availability, and an increased likelihood of transferring land. These findings can inform government policies aimed at improving practices related to land transfers and old age security for rural elders.
ABSTRACT
Facial expression recognition (FER) plays a crucial role in affective computing, enhancing human-computer interaction by enabling machines to understand and respond to human emotions. Despite advancements in deep learning, current FER systems often struggle with challenges such as occlusions, head pose variations, and motion blur in natural environments. These challenges highlight the need for more robust FER solutions. To address these issues, we propose the Attention-Enhanced Multi-Layer Transformer (AEMT) model, which integrates a dual-branch Convolutional Neural Network (CNN), an Attentional Selective Fusion (ASF) module, and a Multi-Layer Transformer Encoder (MTE) with transfer learning. The dual-branch CNN captures detailed texture and color information by processing RGB and Local Binary Pattern (LBP) features separately. The ASF module selectively enhances relevant features by applying global and local attention mechanisms to the extracted features. The MTE captures long-range dependencies and models the complex relationships between features, collectively improving feature representation and classification accuracy. Our model was evaluated on the RAF-DB and AffectNet datasets. Experimental results demonstrate that the AEMT model achieved an accuracy of 81.45% on RAF-DB and 71.23% on AffectNet, significantly outperforming existing state-of-the-art methods. These results indicate that our model effectively addresses the challenges of FER in natural environments, providing a more robust and accurate solution. The AEMT model significantly advances the field of FER by improving the robustness and accuracy of emotion recognition in complex real-world scenarios. This work not only enhances the capabilities of affective computing systems but also opens new avenues for future research in improving model efficiency and expanding multimodal data integration.
ABSTRACT
Photosystem I (PSI) light-harvesting antenna complexes LHCI contain spectral forms that absorb and emit photons of lower energy than that of its primary electron donor, P700. The most red-shifted fluorescence is associated with the Lhca4 complex. It has been suggested that this red emission is related to the inter-chlorophyll charge transfer (CT) states. In this work we present a systematic quantum-chemical study of the CT states in Lhca4, accounting for the influence of the protein environment by estimating the electrostatic interactions. We show that significant energy shifts result from these interactions and propose that the emission of the Lhca4 complex is related not only to the previously proposed a603+-a608- state, but also to the a602+-a603- state. We also investigate how different protonation patterns of protein amino acids affect the energetics of the CT states.
ABSTRACT
In this work, we explore a simplified model based on both analytical and computational methods for the study of film-boiling droplet motion on microscale ratchets. We consider a specific ratchet design with the length periods and depth of ratchets much smaller than the size of the droplet. We conclude based on our modeling that for the ratchet configuration considered in this paper, the conduction within the vapor film is the dominant means of heat transfer in comparison with convection and radiation. Furthermore, we demonstrate a more manageable two-dimensional model in which analytical approaches coupled with computational approaches yield reasonably accurate results in comparison to the actual experiments.
ABSTRACT
Introduction: The infertile patient's knowledge, attitude, and practice (KAP) toward embryo transfer may affect treatment outcomes and the mental health of women who underwent in vitro fertilization-embryo transfer (IVF-ET). This study aimed to investigate the KAP of embryo transfer among women who underwent IVF-ET. Methods: This cross-sectional study was conducted on women who underwent IVF-ET at our Hospital between May 2023 and November 2023, using a self-designed questionnaire. Results: A total of 614 valid questionnaires were finally included. The mean KAP scores were 19.46 ± 5.06 (possible range: 0 28), 39.41 ± 5.20 (possible range: 12-60), and 48.02 ± 6.75 (possible range: 0-60), respectively. The structural equation model demonstrated that knowledge has a direct effect on attitude (ß = 0.27, p < 0.001) and attitude has a direct effect on practice (ß = 0.55, p < 0.001) and anxiety (ß = 0.59, p < 0.001). Moreover, multivariable linear regression analysis showed that anxiety score [coefficient = 0.09, 95% confidence interval (CI): 0.03-0.16, p = 0.003], BMI (coefficient = 0.09, 95%CI: 0.03-0.16, p = 0.003), education (coefficient = 5.65-6.17, 95%CI: 1.09-10.7, p < 0.05), monthly per capita income (coefficient = 1.20-1.96, 95% CI: 0.21-3.07, p = 0.05), reasons for IVF (coefficient = -1.33-1.19, 95% CI: -2.49-0.09, p < 0.05), and more than 5 years of infertility (coefficient = -1.12, 95% CI: -2.11-0.13, p = 0.026) were independently associated with sufficient knowledge. Knowledge (coefficient = 0.19, 95% CI: 0.12-0.26, p < 0.001), anxiety (coefficient = 0.39, 95% CI: 0.34-0.45, p < 0.001), monthly per capita household income >10,000 (coefficient = 1.52, 95% CI: 0.61-2.43, p < 0.001), and three or more cycles of embryo transfer (coefficient = -2.69, 95% CI: -3.94-1.43, p < 0.001) were independently associated with active attitude. Furthermore, attitude (coefficient = 0.21, 95% CI: 0.11-0.30, p < 0.001) and anxiety (coefficient = 0.57, 95% CI: 0.49-0.65, p < 0.001) were independently associated with proactive practice. Discussion: Women who underwent IVF-ET had inadequate knowledge and negative attitudes but proactive practice toward embryo transfer, which were affected by anxiety, income, and reasons for IVF. It is necessary to strengthen the continuous improvement of patient education to improve the management of embryo transfer.
ABSTRACT
The repair of bone defects in the elderly individuals is significantly delayed due to cellular senescence and dysfunction, which presents a challenge in clinical settings. Furthermore, there are limited effective methods available to promote bone repair in older individuals. Herein, melatonin-loaded mesoporous bioactive glasses microspheres (MTBG) were successfully prepared based on their mesoporous properties. The repair of bone defects in aged rats was significantly accelerated by enhancing mitochondrial function through the sustained release of melatonin and bioactive ions. MTBG effectively rejuvenated senescent bone marrow mesenchymal stem cells (BMSCs) by scavenging excessive reactive oxygen species (ROS), stabilizing the mitochondrial membrane potential (ΔΨm), and increasing ATP synthesis. Analysis of the underlying mechanism revealed that the formation of tunneling nanotubes (TNTs) facilitated the intercellular transfer of mitochondria, thereby resulting in the recovery of mitochondrial function. This study provides critical insights into the design of new biomaterials for the elderly individuals and the biological mechanism involved in aged bone regeneration.
ABSTRACT
Oxygenic phototrophs use chlorophylls (Chls) as photosynthetically active pigments. A variety of Chl molecules have been found in photosynthetic eukaryotes including green plants, algae, and cyanobacteria. Here we review their molecular structures with stereochemistry, occurrence in light-harvesting antennas and reaction centers, biosyntheses in the late stage, chemical stabilities, and visible absorption maxima in diethyl ether. The observed maxima are comparable to those of semisynthetic Chl analogs, methyl pyropheophorbides, in dichloromethane. The effects of their peripheral substituents and core π-conjugation on the maxima of the monomeric states are discussed. Notably, the oxidation along the molecular x-axis in Chl-a produces its accessory pigments, Chls-b/c, and introduction of an electron-withdrawing formyl group along the y-axis perpendicular to the x-axis affords far-red light absorbing Chls-d/f.
ABSTRACT
Nanoplastics represent a global environmental concern due to their ubiquitous presence and potential adverse impacts on public and environmental health. There is a growing need to advance the mechanistic understanding of their reactivity as they interact with biological and environmental systems. Herein, for the first time, we report that polystyrene nanoplastics (PSNPs) have intrinsic peroxidase-like activity and are able to mediate oxidative stress. The peroxidase-like activity is dependent on temperature and pH, with a maximum at pH 4.5 and 40 °C. The catalytic activity exhibits saturation kinetics, as described by the Michaelis-Menten model. The peroxidase-like activity of PSNPs is attributed to their ability to mediate electron transfer from peroxidase substrates to H2O2. Ozone-induced PSNP aging can introduce oxygen-containing groups and disrupt aromatic structures on the nanoplastic surface. While ozonation initially enhances peroxidase-like activity by increasing oxygen-containing groups without degrading many aromatic structures, extended ozonation destroys aromatic structures, significantly reducing this activity. The peroxidase-like activity of PSNPs can mediate oxidative stress, which is generally positively correlated with their aromatic structures, as suggested by the ascorbic acid assay. These results help explain the reported oxidative stress exerted by nanoplastics and provide novel insights into their environmental and public health implications.
ABSTRACT
From microbes to humans, organisms perform numerous tasks for their survival, including food acquisition, migration, and reproduction. A complex biological task can be performed by either an autonomous organism or by cooperation among several specialized organisms. However, it remains unclear how autonomy and cooperation evolutionarily switch. Specifically, it remains unclear whether and how cooperative specialists can repair deleted genes through direct genetic exchange, thereby regaining metabolic autonomy. Here, we address this question by experimentally evolving a mutualistic microbial consortium composed of two specialists that cooperatively degrade naphthalene. We observed that autonomous genotypes capable of performing the entire naphthalene degradation pathway evolved from two cooperative specialists and dominated the community. This evolutionary transition was driven by the horizontal gene transfer (HGT) between the two specialists. However, this evolution was exclusively observed in the fluctuating environment alternately supplied with naphthalene and pyruvate, where mutualism and competition between the two specialists alternated. The naphthalene-supplied environment exerted selective pressure that favors the expansion of autonomous genotypes. The pyruvate-supplied environment promoted the coexistence and cell density of the cooperative specialists, thereby increasing the likelihood of HGT. Using a mathematical model, we quantitatively demonstrate that environmental fluctuations facilitate the evolution of autonomy through HGT when the relative growth rate and carrying capacity of the cooperative specialists allow enhanced coexistence and higher cell density in the competitive environment. Together, our results demonstrate that cooperative specialists can repair deleted genes through a direct genetic exchange under specific conditions, thereby regaining metabolic autonomy.
Subject(s)
Naphthalenes , Naphthalenes/metabolism , Gene Transfer, Horizontal , Biological Evolution , Symbiosis , Microbial Consortia/genetics , Microbial Consortia/physiology , GenotypeABSTRACT
Sodium salicylate (NaSA) molecule exists as salicylate anion in acetonitrile (ACN) and water solvents, and exhibits large Stokes shifted fluorescence due to excited state intramolecular proton transfer (ESIPT), with decay times of â¼5 ns in ACN and â¼3.9 ns in water by 300 nm (absorption maxima) excitation. Previous studies report both ground and excited state enol-keto tautomerization in ACN, but only excited state tautomerization in water at 10-4 M. However, the current work explores the effect of concentration and excitation wavelengths on the photoinduced dynamics of ESIPT in the salicylate anion. On increasing the concentration of NaSA, no change in absorption spectra appears in both the solvents, and emission spectra of NaSA in water remain unaffected by changes in concentration or excitation wavelength, whereas, a slight red shift and decrease in FWHM appears in ACN. Time-domain fluorescence measurements show predominantly single-exponential decay throughout the emission profile by 300 nm excitation above the 10-5 M concentration in both the solvents, while by 375 nm excitation, multi-exponential fluorescence decay is observed at low concentrations, and as the concentration of NaSA increases, this decay behaviour tends to converge towards a single exponential decay. These results suggest that solute-solvent interactions stabilize the ground-state intermolecular hydrogen-bonded species at low concentrations, while higher concentrations weaken these interactions, leading to emission solely from the salicylate anion. Peak fit analysis of excitation spectra confirms enol-keto tautomerization in both the solvents, with the keto form being more stabilized in ACN. These findings underscore that in ACN, behaviour of NaSA is influenced by both concentration and excitation wavelength and contrary to previous reports, the keto form of the molecule is also present in water, though at a very low concentration and an increase in non-radiative transitions in water cause fluorescence quenching.
ABSTRACT
The thermal-coupled levels (TCLs) of lanthanides have attracted great attention in the field of optical thermometer, offering an efficient method to achieve non-contect temperatuer feedback in complex environment. However, the iner 4f electrons are shielded, which becomes the core obstacle in improving the sensing performance. This issue is now circumvented by constructing an electron transfer pathway between Tm3+(1D2) and Eu3+(5D0) configurations. As a result, the electron transfer barrier is related to the relative temperature sensitivity, giving an insight into the modulation mechanism. Compared to the conventional TCLs systems, the relative temperature sensitivity of this strategy is highly concentration-responsive, increasing from 5.56 to 10.1 % K-1 as the Eu3+ molar concentration rises from 0.3 to 0.5 mol%. This work reveals the inner emission mechanism based on IVCT-supported emission mode, and presents the highly adjustability of sensing performance.
ABSTRACT
Pre-training deep learning models with large data sets of natural images, such as ImageNet, has become the standard for endoscopic image analysis. This approach is generally superior to training from scratch, due to the scarcity of high-quality medical imagery and labels. However, it is still unknown whether the learned features on natural imagery provide an optimal starting point for the downstream medical endoscopic imaging tasks. Intuitively, pre-training with imagery closer to the target domain could lead to better-suited feature representations. This study evaluates whether leveraging in-domain pre-training in gastrointestinal endoscopic image analysis has potential benefits compared to pre-training on natural images. To this end, we present a dataset comprising of 5,014,174 gastrointestinal endoscopic images from eight different medical centers (GastroNet-5M), and exploit self-supervised learning with SimCLRv2, MoCov2 and DINO to learn relevant features for in-domain downstream tasks. The learned features are compared to features learned on natural images derived with multiple methods, and variable amounts of data and/or labels (e.g. Billion-scale semi-weakly supervised learning and supervised learning on ImageNet-21k). The effects of the evaluation is performed on five downstream data sets, particularly designed for a variety of gastrointestinal tasks, for example, GIANA for angiodyplsia detection and Kvasir-SEG for polyp segmentation. The findings indicate that self-supervised domain-specific pre-training, specifically using the DINO framework, results into better performing models compared to any supervised pre-training on natural images. On the ResNet50 and Vision-Transformer-small architectures, utilizing self-supervised in-domain pre-training with DINO leads to an average performance boost of 1.63% and 4.62%, respectively, on the downstream datasets. This improvement is measured against the best performance achieved through pre-training on natural images within any of the evaluated frameworks. Moreover, the in-domain pre-trained models also exhibit increased robustness against distortion perturbations (noise, contrast, blur, etc.), where the in-domain pre-trained ResNet50 and Vision-Transformer-small with DINO achieved on average 1.28% and 3.55% higher on the performance metrics, compared to the best performance found for pre-trained models on natural images. Overall, this study highlights the importance of in-domain pre-training for improving the generic nature, scalability and performance of deep learning for medical image analysis. The GastroNet-5M pre-trained weights are made publicly available in our repository: huggingface.co/tgwboers/GastroNet-5M_Pretrained_Weights.
ABSTRACT
STUDY QUESTION: Can we develop a prediction model for the chance of a live birth following the transfer of an embryo created using donated oocytes? SUMMARY ANSWER: Three primary models that included patient, past treatment, and cycle characteristics were developed using Australian data to predict the chance of a live birth following the transfer of an embryo created using donated oocytes; these models were well-calibrated to the population studied, achieved reasonable predictive power and generalizability when tested on New Zealand data. WHAT IS KNOWN ALREADY: Nearly 9% of ART embryo transfer cycles performed globally use embryos created using donated oocytes. This percentage rises to one-quarter and one-half in same-sex couples and women aged over 45 years, respectively. STUDY DESIGN, SIZE, DURATION: This study uses population-based Australian clinical registry data comprising 9384 embryo transfer cycles that occurred between 2015 and 2021 for model development, with an external validation cohort of 1493 New Zealand embryo transfer cycles. PARTICIPANTS/MATERIALS, SETTING, METHODS: Three prediction models were compared that incorporated patient characteristics, but differed in whether they considered use of prior autologous treatment factors and current treatment parameters. We internally validated the models on Australian data using grouped cross-validation and reported several measures of model discrimination and calibration. Variable importance was measured through calculating the change in predictive performance that resulted from variable permutation. The best-performing model was externally validated on data from New Zealand. MAIN RESULTS AND THE ROLE OF CHANCE: The best-performing model had an internal validation AUC-ROC of 0.60 and Brier score of 0.20, and external validation AUC-ROC of 0.61 and Brier score of 0.23. While these results indicate â¼15% less discriminatory ability compared to models assessed on an autologous cohort from the same population the performance of the models was clearly statistically significantly better than random, demonstrated generalizability, and was well-calibrated to the population studied. The most important variables for predicting the chance of a live birth were the oocyte donor age, the number of prior oocyte recipient embryo transfer cycles, whether the transferred embryo was cleavage or blastocyst stage and oocyte recipient age. Of lesser importance were the oocyte-recipient parity, whether donor or partner sperm was used, the number of prior autologous embryo transfer cycles and the number of embryos transferred. LIMITATIONS, REASONS FOR CAUTION: The models had relatively weak discrimination suggesting further features need to be added to improve their predictive power. Variation in donor oocyte cohorts across countries due to differences such as whether anonymous and compensated donation are allowed may necessitate the models be recalibrated prior to application in non-Australian cohorts. WIDER IMPLICATIONS OF THE FINDINGS: These results confirm the well-established importance of oocyte age and ART treatment history as the key prognostic factors in predicting treatment outcomes. One of the developed models has been incorporated into a consumer-facing website (YourIVFSuccess.com.au/Estimator) to allow patients to obtain personalized estimates of their chance of success using donor oocytes. STUDY FUNDING/COMPETING INTEREST(S): This research was funded by the Australian government as part of the Medical Research Future Fund (MRFF) Emerging Priorities and Consumer Driven Research initiative: EPCD000007. L.R. declares personal consulting fees from Abbott and Merck, lecture fees from Abbott, receipt of an educational grant from Merck, past presidency of the Fertility Society of Australia & New Zealand and World Endometriosis Society and being a minor shareholder in Monash IVF Group (ASX:MVF). G.M.C. declares receipt of Australian government grant funding for the research study and the development and maintenance of the YourIVFSuccess website. O.F., J.N., and A.P. report no conflicts of interest. TRIAL REGISTRATION NUMBER: N/A.