Concept-based Lesion Aware Transformer for Interpretable Retinal Disease Diagnosis.

Existing deep learning methods have achieved remarkable results in diagnosing retinal diseases, showcasing the potential of advanced AI in ophthalmology. However, the black-box nature of these methods obscures the decision-making process, compromising their trustworthiness and acceptability. Inspired by the concept-based approaches and recognizing the intrinsic correlation between retinal lesions and diseases, we regard retinal lesions as concepts and propose an inherently interpretable framework designed to enhance both the performance and explainability of diagnostic models. Leveraging the transformer architecture, known for its proficiency in capturing long-range dependencies, our model can effectively identify lesion features. By integrating with image-level annotations, it achieves the alignment of lesion concepts with human cognition under the guidance of a retinal foundation model. Furthermore, to attain interpretability without losing lesion-specific information, our method employs a classifier built on a cross-attention mechanism for disease diagnosis and explanation, where explanations are grounded in the contributions of human-understandable lesion concepts and their visual localization. Notably, due to the structure and inherent interpretability of our model, clinicians can implement concept-level interventions to correct the diagnostic errors by simply adjusting erroneous lesion predictions. Experiments conducted on four fundus image datasets demonstrate that our method achieves favorable performance against state-of-the-art methods while providing faithful explanations and enabling conceptlevel interventions. Our code is publicly available at https://github.com/Sorades/CLAT.

Sorption of Iodine on Biochar Derived from the Processing of Urban Sludge and Garden Waste at Different Pyrolysis Temperatures.

The United Nations proposed the Sustainable Development Goals with the aim to make human settlements in cities resilient and sustainable. The excessive discharge of urban waste including sludge and garden waste can pollute groundwater and lead to the emission of greenhouse gases (e.g., CH4). The proper recycling of urban waste is essential for responsible consumption and production, reducing environmental pollution and addressing climate change issues. This study aimed to prepare biochar with high adsorption amounts of iodine using urban sludge and peach wood from garden waste. The study was conducted to examine the variations in the mass ratio between urban sludge and peach wood (2/1, 1/1, and 1/2) as well as pyrolysis temperatures (300 °C, 500 °C, and 700 °C) on the carbon yield and adsorption capacities of biochar. Scanning electron microscopy, Brunauer-Emmett-Teller analysis, Fourier transform infrared spectrometry, powder X-ray diffraction, and elemental analysis were used to characterize the biochar produced at different pyrolysis temperatures and mass ratios. The results indicate that the carbon yield of biochar was found to be the highest (>60%) at a pyrolysis temperature of 300 °C across different pyrolysis temperatures. The absorbed amounts of iodine in the aqueous solution ranged from 86 to 223 mg g-1 at a mass ratio of 1:1 between urban sludge and peach wood, which were comparably higher than those observed in other mass ratios. This study advances water treatment by offering a cost-effective method by using biochar derived from the processing of urban sludge and garden waste.

Personalized starting age of gastric cancer screening based on individuals' risk profiles: a population-based prospective study.

BACKGROUND: The current recommended starting age for gastric cancer (GC) lacks unified guideline and individualized criteria. We aimed to determine risk-stratified starting age for GC screening in China based on individuals' risk profiles, and develop an online calculator for clinical application. METHODS: In this multi-center population-based prospective study, we divided participants enrolled during 2015-2017 (n = 59,771, aged 40-69) into screened and unscreened groups and observed them for primary endpoints-GC occurrence, all-cause and GC-specific deaths. The median follow-up was 6.07 years. To determine the reference starting age, the effectiveness of GC screening was assessed by age-groups after propensity-score-matching. Further, we categorized the calculated individual risk scores (using well-established risk factors) by quantiles. Subsequently, we used age-specific 10-year cumulative risk curves to estimate the risk-stratified starting age-when the individual's risk level matches reference starting age risk threshold. RESULTS: During follow-up, 475 GC cases, 182 GC deaths and 1,860 all-cause deaths occurred. All-cause and GC-specific mortality decreased among screened individuals aged ≥45 and 50-59 years, respectively. Thus, the average population (reference) starting age was set as 50 years. The 10-year cumulative risk of GC in average population aged 50 was 1.147%. We stratified the starting age using eight risk factors, and categorized participants as low-, medium-, and high-risk individuals, whose risk-stratified starting age was 58, 50, and 46, respectively. CONCLUSION: While high-risk individuals warrant 3-5 years earlier GC screening than average population (age 50), low-risk individuals can tolerate delayed screening. Our online, personalized starting-age calculator will help risk-adapted GC screening (https://web.consultech.com.cn/gastric/#/).

Facile, green and scalable synthesis of single-layer manganese dioxide nanosheets and its application for GSH and cTnI colorimetric detection.

Manganese dioxide (MnO2) nanosheets possess unique physical and chemical properties, making them widely applicable in various fields, such as chemistry and biomedicine. Although MnO2 nanosheets are produced using bottom-up wet chemistry synthesis methods, their scale is below the gram level and requires a long processing time, restricting their effective scale-up from laboratory to market. We report a facile, green and scalable synthesis of MnO2 nanosheets by mixing Shiranui mandarin orange juice and KMnO4 for 30 minutes. We produced more than one gram (1.095) of MnO2 nanosheets with a 0.65 nm mean thickness and a 50 nm mean lateral size. Furthermore, we established a visual colorimetric biosensing strategy based on MnO2 nanosheets for the assay of glutathione (GSH) and cardiac troponin I (cTnI), offering high sensitivity and feasibility in clinical samples. For GSH, the limit of detection was 0.08 nM, and for cTnI, it was 0.70 pg mL-1. Meanwhile, the strategy can be used for real-time analysis by applying a smartphone-enabled biosensing strategy, which can provide point-of-care testing in remote areas.

Research progress of Ganoderma lucidum polysaccharide in prevention and treatment of Atherosclerosis.

Atherosclerosis (AS) is a chronic inflammatory disease resulting from dysregulated lipid metabolism, constituting the pathophysiological foundation of cardiovascular and cerebrovascular diseases. AS has a high incidence rate and mortality rate worldwide. As such, traditional Chinese medicine (TCM) has been widely used recently due to its stable therapeutic effect and high safety. Ganoderma lucidum polysaccharides (GLP) are the main active ingredients of Ganoderma lucidum, a Chinese herbal medicine. Research has also shown that GLP has anti-inflammatory and antioxidant properties, regulates gut microbiota, improves blood glucose and lipid levels, and inhibits obesity. Most of the current research on GLP anti-AS is focused on animal models. Thus, its clinical application remains to be discovered. In this review, we combine relevant research results and start with the pathogenesis and risk factors of GLP on AS, proving that GLP can prevent and treat AS, providing a scientific basis and reference for the future prevention and treatment of AS with GLP.

Plant-based fascia tissues: Exploring materials and techniques for realistic simulation.

The growing demand for sustainable and ethical food options has led to significant advancements in plant-based meat substitutes (PBMS). PBMS have made considerable progress in simulating the taste, texture, and sensory properties of animal meat. Connective tissue is a fundamental component of animal meat that significantly influences tenderness, texture, and sensory properties. However, the imitation of realistic connective tissues has received relatively less attention in the PBMS industry. The current work focuses on exploring materials and techniques for the replication of plant-based connective tissues (PBCT). By understanding the structural and functional characteristics of animal connective tissues (ACT), it is possible to replicate these characteristics in PBCT. Hydrogels, with their ability to simulate certain properties of ACT, present a viable material for the creation of PBCT. To achieve the desired simulation, their mechanical and structural properties need to be enhanced by using several materials and several physical techniques.

Preparation of Perilla frutescens L. essential oil hydrogel beads and preservation application research in strawberry.

Perilla frutescens L. essential oil (PLEO) has antibacterial and antioxidant properties, which can effectively maintain the quality of fruits and extend their shelf life. In this study, sodium alginate and chitosan were used as wall materials, and PLEO microcapsule powder was used as the core materials to prepare PLEO hydrogel beads. The best results were obtained by using 2%w/v sodium alginate and 1.5%w/v chitosan as wall materials, with a core-to-wall ratio of 2:1 and homogenized for 15 min producing PLEO hydrogel beads with encapsulation efficiency of 82.61 %. For strawberries preservation, PLEO hydrogel beads preservation group had a better effect after 5 d of storage, showing a lower decay rate (15.71 %), better maintaining the hardness of 1.75 kg/cm2, and a weight loss of 3.29 %. Furthermore, organic acids and total phenols were retained more in this group, the number of microorganisms was significantly reduced, and sensory qualities were improved, especially taste and color. This study provides important insights into the application of natural preservatives in the food industry and promotes sustainable practices in food preservation.

A single-bridge interleaved three-level LLC resonant converter with current sharing capability for fuel cell system.

This paper propose a wide gain single-bridge interleaved three-level LLC resonant converter with current sharing capability. This novel converter offers numerous advantages, including low cost, low current ripple, reduced voltage and current stress, widely gain range, high efficiency, good current sharing capability and versatile application scalability. Utilizing the three-level inverter + full-wave rectifier LLC converter as a representative case, the paper conducts an in-depth analysis and research on the proposed method. Finality, a 600 W experimental prototype was constructed and tested. Experimental results reveal that the proposed converter exhibits lower current ripple and a broader gain range. Moreover, the converter shows good current sharing capability (with a resonant element tolerance of 10%, the current error between the two phases does not exceed 12%) and high efficiency (peaking at 95.8%).

Risk factors for the recurrence of left atrioventricular valvular regurgitation after surgical repair of partial and transitional atrioventricular septal defect.

Background: Left atrioventricular valvular regurgitation (LAVVR) recurrence after partial and transitional atrioventricular septal defect (AVSD) repair is the main risk factor associated with reoperation or mortality. The purpose of this study was to identify risk factors associated with the recurrence of LAVVR after surgical repair of transitional and partial AVSD at a single institution. Methods: A hundred and fifty-seven patients who underwent anatomical repair for partial and transitional AVSD from January 2013 to December 2021 were included in our institutional database. Demographic characteristics, operative information, comorbidities, complications, and outcomes were retrieved from electronic medical records. Echocardiographic evaluations included cardiac dimensions, the degree of LAVVR, and the anatomy of the atrioventricular valve. Results: After a median follow-up period of 5.8 years, 40 patients had recurrent moderate or even more severe LAVVR. Compared with patients without recurrent LAVVR, those experiencing LAVVR recurrence were more likely to have larger preoperative left atrial (LA) size and larger left ventricular (LV) size after standardization, larger left atrioventricular valve (LAVV) cleft width, higher proportions of preoperative moderate or even more severe LAVVR, and immediately postoperative mild to moderate or even more severe LAVVR. Univariate Cox regression analysis showed that age at first repair, height, LA size after standardization, LV size after standardization, the severity of preoperative LAVVR, immediately postoperative LAVVR, and the LAVV cleft width more than 1cm were risk factors for recurrent LAVVR (P<0.05 for all). Multivariable Cox regression analysis showed that mild to moderate or even more severe LAVVR postoperatively [hazard ratio (HR) 9.53, 95% confidence interval (CI): 3.78-24.01; P<0.001], the width of LAVV cleft more than 1 cm (HR: 3.90, 95% CI: 1.80-8.48; P<0.001) and age at first repair (HR: 0.45, 95% CI: 0.31-0.66; P<0.001) were independently associated with the recurrence of LAVVR. Conclusions: The width of LAVV cleft, mild to moderate or even more severe LAVVR immediately after surgery, and age at initial surgery are risk factors for recurrent LAVVR. The presence of recurrent LAVVR necessitates proactive surveillance to facilitate timely reintervention.

Species delimitation 4.0: integrative taxonomy meets artificial intelligence.

Although species are central units for biological research, recent findings in genomics are raising awareness that what we call species can be ill-founded entities due to solely morphology-based, regional species descriptions. This particularly applies to groups characterized by intricate evolutionary processes such as hybridization, polyploidy, or asexuality. Here, challenges of current integrative taxonomy (genetics/genomics + morphology + ecology, etc.) become apparent: different favored species concepts, lack of universal characters/markers, missing appropriate analytical tools for intricate evolutionary processes, and highly subjective ranking and fusion of datasets. Now, integrative taxonomy combined with artificial intelligence under a unified species concept can enable automated feature learning and data integration, and thus reduce subjectivity in species delimitation. This approach will likely accelerate revising and unraveling eukaryotic biodiversity.

Cycloastragenol induces apoptosis and protective autophagy through AMPK/ULK1/mTOR axis in human non-small cell lung cancer cell lines.

OBJECTIVE: Studies have demonstrated that cycloastragenol induces antitumor effects in prostate, colorectal and gastric cancers; however, its efficacy for inhibiting the proliferation of lung cancer cells is largely unexplored. This study explores the efficacy of cycloastragenol for inhibiting non-small cell lung cancer (NSCLC) and elucidates the underlying molecular mechanisms. METHODS: The effects of cycloastragenol on lung cancer cell proliferation were assessed using an adenosine triphosphate monitoring system based on firefly luciferase and clonogenic formation assays. Cycloastragenol-induced apoptosis in lung cancer cells was evaluated using dual staining flow cytometry with an annexin V-fluorescein isothiocyanate/propidium iodide kit. To elucidate the role of cycloastragenol in the induction of apoptosis, apoptosis-related proteins were examined using Western blots. Immunofluorescence and Western blotting were used to determine whether cycloastragenol could induce autophagy in lung cancer cells. Genetic techniques, including small interfering RNA technology, were used to investigate the underlying mechanisms. The effects against lung cancer and biosafety of cycloastragenol were evaluated using a mouse subcutaneous tumor model. RESULTS: Cycloastragenol triggered both autophagy and apoptosis. Specifically, cycloastragenol promoted apoptosis by facilitating the accumulation of phorbol-12-myristate-13-acetate-induced protein 1 (NOXA), a critical apoptosis-related protein. Moreover, cycloastragenol induced a protective autophagy response through modulation of the adenosine 5'-monophosphate-activated protein kinase (AMPK)/unc-51-like autophagy-activating kinase (ULK1)/mammalian target of rapamycin (mTOR) pathway. CONCLUSION: Our study sheds new light on the antitumor efficacy and mechanism of action of cycloastragenol in NSCLC. This insight provides a scientific basis for exploring combination therapies that use cycloastragenol and inhibiting the AMPK/ULK1/mTOR pathway as a promising approach to combating lung cancer. Please cite this article as follows: Zhu LH, Liang YP, Yang L, Zhu F, Jia LJ, Li HG. Cycloastragenolinduces apoptosis and protective autophagy through AMPK/ULK1/mTOR axis in human non-small celllung cancer cell lines. J Integr Med. 2024: Epub ahead of print.

Salt ions improve soybean protein isolate/curdlan complex fat substitutes: Effect of molecular interactions on freeze-thaw stability.

Although emulsion gels show significant potential as fat substitutes, they are vulnerable to degreasing, delamination, and other undesirable processes during freezing, storage, and thawing, leading to commercial value loss in terms of juiciness, flavor, and texture. This study investigated the gel strength and freeze-thaw stability of soybean protein isolate (SPI)/curdlan (CL) composite emulsion gels after adding sodium chloride (NaCl). Analysis revealed that adding low salt ion concentrations promoted the hardness and water-holding capacity (WHC) of fat substitutes, while high levels displayed an inhibitory effect. With 40 mM NaCl as the optimum concentration, the hardness increased from 259.33 g (0 mM) to 418.67 g, the WHC increased from 90.59 % to 93.18 %, exhibiting good freeze-thaw stability. Confocal laser scanning microscopy (CLSM) and particle size distribution were used to examine the impact of salt ion concentrations on protein particle aggregation and the damaging effect of freezing and thawing on the proteoglycan complex network structure. Fourier-transform infrared spectroscopy (FTIR) and protein solubility evaluation indicated that the composite gel network structure consisted of covalent contacts between the proteoglycan molecules and hydrogen bonds, playing a predominant role in non-covalent interaction. This study showed that the salt ion concentration in the emulsion gel affected its molecular interactions.

Rapid reconstitution of ubiquitinated nucleosome using a non-denatured histone octamer ubiquitylation approach.

BACKGROUND: Histone ubiquitination modification is emerging as a critical epigenetic mechanism involved in a range of biological processes. In vitro reconstitution of ubiquitinated nucleosomes is pivotal for elucidating the influence of histone ubiquitination on chromatin dynamics. RESULTS: In this study, we introduce a Non-Denatured Histone Octamer Ubiquitylation (NDHOU) approach for generating ubiquitin or ubiquitin-like modified histone octamers. The method entails the co-expression and purification of histone octamers, followed by their chemical cross-linking to ubiquitin using 1,3-dibromoacetone. We demonstrate that nucleosomes reconstituted with these octamers display a high degree of homogeneity, rendering them highly compatible with in vitro biochemical assays. These ubiquitinated nucleosomes mimic physiological substrates in function and structure. Additionally, we have extended this method to cross-linking various histone octamers and three types of ubiquitin-like proteins. CONCLUSIONS: Overall, our findings offer an efficient strategy for producing ubiquitinated nucleosomes, advancing biochemical and biophysical studies in the field of chromatin biology.

Effect of soybean proteins and peptides on the growth and adhesive ability of Limosilactobacillus Reuteri DSM17938.

Limosilactobacillus reuteri DSM17938 is one of the most pivotal probiotics, whose general beneficial effects on the intestinal microbiota are well recognized. Enhancing their growth and metabolic activity can effectively regulate the equilibrium of intestinal microbiota, leading to improved physical health. A common method to promote the growth of Lactobacillus is the addition of prebiotics. Current research suggests that proteins and their hydrolysates from different sources with potential prebiotic activity can also promote the growth of probiotics. In this study, soybean proteins and peptides were effective in promoting the growth, organic acid secretion, and adhesive properties of Limosilactobacillus reuteri DSM17938 to Caco-2 cells. These results illustrate the feasibility of soybean proteins and peptides as prebiotics, providing theoretical and practical advantages for their application.

Federated Learning for Generalization, Robustness, Fairness: A Survey and Benchmark.

Federated learning has emerged as a promising paradigm for privacy-preserving collaboration among different parties. Recently, with the popularity of federated learning, an influx of approaches have delivered towards different realistic challenges. In this survey, we provide a systematic overview of the important and recent developments of research on federated learning. Firstly, we introduce the study history and terminology definition of this area. Then, we comprehensively review three basic lines of research: generalization, robustness, and fairness, by introducing their respective background concepts, task settings, and main challenges. We also offer a detailed overview of representative literature on both methods and datasets. We further benchmark the reviewed methods on several well-known datasets. Finally, we point out several open issues in this field and suggest opportunities for further research. We also provide a public website to continuously track developments in this fast advancing field: https://github.com/WenkeHuang/MarsFL.

Causal relationship between gut microbiota and functional outcomes after ischemic stroke: A comprehensive Mendelian randomization study.

AIMS: To investigate the association of the genetic predisposition of specific gut microbiotas with the clinical outcome of ischemic stroke. METHODS: We leveraged publicly available genome-wide association study (GWAS) data to perform Mendelian randomization (MR) analysis. The gut microbiota-related GWAS data from 18,340 individuals from the international consortium MiBioGen was used. The summary data for functional outcomes after ischemic stroke was obtained from the Genetics of Ischemic Stroke Functional Outcome (GISCOME) network meta-analysis. The primary outcomes were judged by the modified Rankin Scale (mRS). The principal analyses were conducted using the inverse-variance weighted (IVW) MR method. The Cochran's Q test, weighted median, MR-Egger regression, leave-one-SNP-out analysis, MR-Pleiotropy Residual Sum, and Outlier methods were adopted as sensitivity analyses. Furthermore, we performed bi-directional MR analysis and the MR Steiger directionality test to examine the direction of the causal relations. RESULTS: The results demonstrated that the genetic predisposition of genus Lactococcus, genus Ruminococcaceae NK4A214 group, family Peptostreptococcaceae, and genus Odoribacter was positively associated with favorable functional outcome after ischemic stroke. Genus Collinsella, genus Ruminococcaceae UCG005, genus Akkermansia, genus Eubacterium oxidoreducens group, and family Verrucomicrobiaceae were identified to be associated with worse functional outcomes after ischemic stroke. Our results showed no evidence of heterogeneity, directional pleiotropic effects, or collider bias, and the sensitivity of our analysis was acceptable. CONCLUSION: The genetic predisposition of different gut microbiotas was associated with the clinical outcome of ischemic stroke. Microbiota adjustment was a promising method to improve the clinical outcome of ischemic stroke.

Lethal pulmonary thromboembolism in mice induced by intravenous human umbilical cord mesenchymal stem cell-derived large extracellular vesicles in a dose- and tissue factor-dependent manner.

Mesenchymal stem cell-derived extracellular vesicles (MSC-EVs) have obvious advantages over MSC therapy. But the strong procoagulant properties of MSC-EVs pose a potential risk of thromboembolism, an issue that remains insufficiently explored. In this study, we systematically investigated the procoagulant activity of large EVs derived from human umbilical cord MSCs (UC-EVs) both in vitro and in vivo. UC-EVs were isolated from cell culture supernatants. Mice were injected with UC-EVs (0.125, 0.25, 0.5, 1, 2, 4 µg/g body weight) in 100 µL PBS via the tail vein. Behavior and mortality were monitored for 30 min after injection. We showed that these UC-EVs activated coagulation in a dose- and tissue factor-dependent manner. UC-EVs-induced coagulation in vitro could be inhibited by addition of tissue factor pathway inhibitor. Notably, intravenous administration of high doses of the UC-EVs (1 µg/g body weight or higher) led to rapid mortality due to multiple thrombus formations in lung tissue, platelets, and fibrinogen depletion, and prolonged prothrombin and activated partial thromboplastin times. Importantly, we demonstrated that pulmonary thromboembolism induced by the UC-EVs could be prevented by either reducing the infusion rate or by pre-injection of heparin, a known anticoagulant. In conclusion, this study elucidates the procoagulant characteristics and mechanisms of large UC-EVs, details the associated coagulation risk during intravenous delivery, sets a safe upper limit for intravenous dose, and offers effective strategies to prevent such mortal risks when high doses of large UC-EVs are needed for optimal therapeutic effects, with implications for the development and application of large UC-EV-based as well as other MSC-EV-based therapies.

Universal materials model of deep-learning density functional theory Hamiltonian.

Realizing large materials models has emerged as a critical endeavor for materials research in the new era of artificial intelligence, but how to achieve this fantastic and challenging objective remains elusive. Here, we propose a feasible pathway to address this paramount pursuit by developing universal materials models of deep-learning density functional theory Hamiltonian (DeepH), enabling computational modeling of the complicated structure-property relationship of materials in general. By constructing a large materials database and substantially improving the DeepH method, we obtain a universal materials model of DeepH capable of handling diverse elemental compositions and material structures, achieving remarkable accuracy in predicting material properties. We further showcase a promising application of fine-tuning universal materials models for enhancing specific materials models. This work not only demonstrates the concept of DeepH's universal materials model but also lays the groundwork for developing large materials models, opening up significant opportunities for advancing artificial intelligence-driven materials discovery.