Photoelectrochemical Sensor for H2S Based on a Lead-Free Perovskite/Metal-Organic Framework Composite.

Halide perovskites have emerged as a highly promising class of photoelectric materials. However, the application of lead-based perovskites has been hindered by their toxicity and relatively weak stability. In this work, a composite material comprising a lead-free perovskite cesium copper iodide (CsCu2I3) nanocrystal and a metal-organic framework (MOF-801) has been synthesized through an in situ growth approach. The resulting composite material, denoted as CsCu2I3/MOF-801, demonstrates outstanding stability and exceptional optoelectronic characteristics. MOF-801 may serve a dual role by acting as a protective barrier between CsCu2I3 nanocrystals and the external environment, as well as promoting the efficient transfer of photogenerated charge carriers, thereby mitigating their recombination. Consequently, CsCu2I3/MOF-801 demonstrates its utility by providing both stability and a notably high initial photocurrent. Leveraging the inherent reactivity between H2S and the composite material, which results in the formation of Cu2S and structural alteration, an exceptionally sensitive photoelectrochemical sensor for H2S detection has been designed. This sensor exhibits a linear detection range spanning from 0.005 to 100 µM with a remarkable detection limit of 1.67 nM, rendering it highly suitable for precise quantification of H2S in rat brains. This eco-friendly sensor significantly broadens the application horizon of perovskite materials and lays a robust foundation for their future commercialization.

Unimolecular Cascaded Multienzyme Conjugates Modulate the Microenvironment of Diabetic Wound to Promote Healing.

An abnormal microenvironment underlies poor healing in chronic diabetic chronic wounds. However, effectively modulating the microenvironment of the diabetic wound remains a great challenge due to sustained oxidative stress and chronic inflammation. Here, we present a unimolecular enzyme-polymer conjugate that demonstrates excellent multienzymatic cascade activities. The cascaded enzyme conjugates (CECs) were synthesized by grafting poly(N-acryloyl-lysine) (pLAAm) from the glycan moieties of glucose oxidase (GOx) via glycan-initiated polymerization. The resulting CECs exhibited multiple enzymatic properties of GOx, superoxide dismutase mimic, and catalase mimic activities simultaneously. The CECs facilitated the depletion of high blood glucose, ROS scavenging, bacteria-killing, anti-inflammatory effects, and sustained oxygen generation, which restored the microenvironment in diabetic wounds. In vivo results from a diabetic mouse model confirmed the capacity and efficiency of the cascade reaction for diabetic wound healing. Our findings demonstrate that the three-in-one enzyme-polymer conjugates alone can modulate the diabetic microenvironment for wound healing.

LINC00664/miR-411-5p/KLF9 feedback loop contributes to the human oral squamous cell carcinoma progression.

OBJECTIVES: Oral squamous cell carcinoma (OSCC) is one of the most aggressive head and neck cancers with high incidence. Multiple studies have revealed that long non-coding RNAs (lncRNAs) play pivotal roles in tumorigenesis. However, the role of long intergenic non-protein coding RNA 664 (LINC00664) on the progression of OSCC was still unclear. SUBJECTS AND METHODS: In this study, the expression of LINC00664 in OSCC tissues and cell lines was detected by quantitative real-time polymerase chain reaction (qRT-PCR). The functional role of LINC0664 was estimated by cell counting kit-8 (CCK-8), transwell assays, Western blot in vitro, and xenograft tumor model in vivo. The regulatory mechanism was investigated by RNA-binding protein immunoprecipitation (RIP), chromatin immunoprecipitation (ChIP), and luciferase reporter assays. RESULTS: LINC00664 was found to be upregulated in OSCC tissues and cell lines and was associated with poor prognosis of OSCC patients. LINC00664 knockdown suppressed OSCC cell proliferation, migration, invasion, and epithelial-mesenchymal transition (EMT). Moreover, Kruppel like factor 9 (KLF9) enhanced LINC00664 expression at transcription level. Interestingly, LINC00664 upregulated KLF9 expression by sponging miR-411-5p. In addition, knockdown of LINC00664 restrained tumor growth of OSCC in vivo. CONCLUSION: Our study identified the oncogenic roles of LINC00664 in OSCC tumorigenesis and EMT via KLF9/LINC00664/miR-411-5p/KLF9 feedback loop, which provides new perspectives of the potential therapeutic target for OSCC.

Aldehyde dehydrogenase 2 deficiency significantly exacerbates tert-butyl alcohol-induced toxicity in mice.

Owing to the use of ethyl tert-butyl ether (ETBE) as a fuel additive, the possible adverse effects of ETBE exposure have become a public concern. Our previous study showed that ETBE-induced toxicity in aldehyde dehydrogenase 2 (Aldh2) gene knockout (KO) mice was caused by its primary metabolite acetaldehyde, which was toxic. However, it is unclear whether tert-butyl alcohol (TBA), another main metabolite of ETBE, plays a role in ETBE-induced toxicity. To investigate this relationship, we analyzed the changes of TBA concentrations in tissues after ETBE exposure, and then evaluated the toxicity after direct TBA treatment in both KO and wild-type (WT) mice. An exposure to 500 ppm ETBE via inhalation resulted in the formation of its three metabolites, TBA, 2-methyl-1,2-propanediol and ethanol, whose concentrations in the liver, brain, fat and testis of male KO mice were significantly higher than the corresponding concentrations observed in male WT mice. Direct treatment to TBA (20 mg/mL of drinking water) caused significant changes in relative organ weights and histopathology, and increased levels of genetic damages in both types of mice. These toxic effects were also seen in KO mice exposed to a lower concentration of TBA (5 mg/mL), which was associated with increased oxidative stress in serum (reduced glutathione and reduced glutathione/oxidized glutathione ratio decreased). Our findings indicate that ALDH2 is involved in the metabolism of ETBE and TBA, and ALDH2 deficiency could greatly increase the sensitivity to TBA-induced toxicity.

Cytocompatible Modification of Thermoresponsive Polymers on Living Cells for Membrane Proteomic Isolation and Analysis.

Efficient strategies for enriching and separating proteins are important and challenging for membrane proteomics. Many existing methods are caught in the dilemma of preserving maximal membrane proteins while avoiding the contamination of cytoplasmic proteins and organelles. Here, we report a polymer anchoring strategy for the selective preparation of membrane proteins through cell surface-initiated atom transfer radical polymerization. The cytocompatible polymerization strategy enables thermoresponsive poly( N-isopropylacrylamide) (pNIPPAm) chains to be grown from a specific protein on the surface of living cells. The polymer tagged membrane protein could be easily separated and enriched by thermoprecipitation. This method led to the identification of 1825 proteins of which 1036 (71.7%) were specific membrane proteins in E. coli. The separated proteins were identified by 2-DE and mass spectrometry. Among the 12 protein spots from the gel slice, eight were identified as outer membrane proteins. The described strategy opens up a new avenue for membrane protein enrichment and separation and may expedite the future development of membrane proteomics.

Acute inhalation co-exposure to 1,2-dichloropropane and dichloromethane cause liver damage by inhibiting mitochondrial respiration and defense ability in mice.

1,2-Dichloropropane (1,2-DCP) is used as an industrial solvent, insecticide fumigant and household dry cleaning product. Carcinogenicity caused by long-term exposure to 1,2-DCP is well established. However, the possible liver damage and related toxic mechanisms associated with acute inhalation exposure to 1,2-DCP are rarely reported. In this study, we investigated the effects of individual and combined exposure to 1,2-DCP and dichloromethane (DCM) on mice liver. The results showed that 1,2-DCP significantly caused liver necrosis, possibly due to 1,2-DCP-induced inhibition of the mitochondrial respiratory chain complex I-IV activities, resulting in mitochondrial dysfunction and extreme ATP consumption. Moreover, 1,2-DCP also decreased mitochondrial defense ability by inhibiting the mitochondrial glutathione S-transferase 1 (MGST1) activity, further aggravating liver damage. Additionally, we found that DCM co-exposure potentially enhanced 1,2-DCP toxicity. Our findings suggest that inhibition of mitochondrial function and MGST1 activity play critical roles in modulating 1,2-DCP-induced liver damage. Furthermore, our results contribute to study the new mechanism of mitochondria-dominated signaling pathways underlying liver injury induced by 1,2-DCP and DCM.

Effects of occupational exposure to noise and dust on blood pressure in Chinese industrial workers.

Along with the rapid development of economy and urbanization, noise and air pollution are becoming major occupational health hazards in the process of industrial production. In this study, we collected data from 7293 industrial workers in China. The association between occupational exposure of noise and dust and blood pressure was investigated. Controlling for demographic variables, including sex, age, and length of service, a stepwise regression model with backward elimination was constructed. The results showed that both noise and dust decreased the level of systolic blood pressure (p < 0.001). This finding prompted the manufacturing industry to reduce noise and dust hazards and protect the occupational health of workers. Prospective studies in different populations are still required to verify the net contribution of noise and dust to the decrease in blood pressure.

Opportunities for Web-based Drug Repositioning: Searching for Potential Antihypertensive Agents with Hypotension Adverse Events.

BACKGROUND: Drug repositioning refers to the process of developing new indications for existing drugs. As a phenotypic indicator of drug response in humans, clinical side effects may provide straightforward signals and unique opportunities for drug repositioning. OBJECTIVE: We aimed to identify drugs frequently associated with hypotension adverse reactions (ie, the opposite condition of hypertension), which could be potential candidates as antihypertensive agents. METHODS: We systematically searched the electronic records of the US Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) through the openFDA platform to assess the association between hypotension incidence and antihypertensive therapeutic effect regarding a list of 683 drugs. RESULTS: Statistical analysis of FAERS data demonstrated that those drugs frequently co-occurring with hypotension events were more likely to have antihypertensive activity. Ranked by the statistical significance of frequent hypotension reporting, the well-known antihypertensive drugs were effectively distinguished from others (with an area under the receiver operating characteristic curve > 0.80 and a normalized discounted cumulative gain of 0.77). In addition, we found a series of antihypertensive agents (particularly drugs originally developed for treating nervous system diseases) among the drugs with top significant reporting, suggesting the good potential of Web-based and data-driven drug repositioning. CONCLUSIONS: We found several candidate agents among the hypotension-related drugs on our list that may be redirected for lowering blood pressure. More important, we showed that a pharmacovigilance system could alternatively be used to identify antihypertensive agents and sustainably create opportunities for drug repositioning.

Systematic drug safety evaluation based on public genomic expression (Connectivity Map) data: myocardial and infectious adverse reactions as application cases.

Adverse drug reaction (ADR) is of great importance to both regulatory agencies and the pharmaceutical industry. Various techniques, such as quantitative structure-activity relationship (QSAR) and animal toxicology, are widely used to identify potential risks during the preclinical stage of drug development. Despite these efforts, drugs with safety liabilities can still pass through safety checkpoints and enter the market. This situation raises the concern that conventional chemical structure analysis and phenotypic screening are not sufficient to avoid all clinical adverse events. Genomic expression data following in vitro drug treatments characterize drug actions and thus have become widely used in drug repositioning. In the present study, we explored prediction of ADRs based on the drug-induced gene-expression profiles from cultured human cells in the Connectivity Map (CMap) database. The results showed that drugs inducing comparable ADRs generally lead to similar CMap expression profiles. Based on such ADR-gene expression association, we established prediction models for various ADRs, including severe myocardial and infectious events. Drugs with FDA boxed warnings of safety liability were effectively identified. We therefore suggest that drug-induced gene expression change, in combination with effective computational methods, may provide a new dimension of information to facilitate systematic drug safety evaluation.

Green tea epigallocatechin gallate binds to and inhibits respiratory complexes in swelling but not normal rat hepatic mitochondria.

Epigallocatechin gallate (EGCG), the major flavonoid in green tea, is consumed via tea products and dietary supplements, and has been tested in clinical trials. However, EGCG can cause hepatotoxicity in humans and animals by unknown mechanisms. Here EGCG effects on rat liver mitochondria were examined. EGCG showed negligible effects on oxidative phosphorylation at 7.5-100µM in normal mitochondria. However, respiratory chain complexes (RCCs) were profoundly inhibited by EGCG in mitochondria undergoing Ca(2+) overload-induced mitochondrial permeability transition (MPT). As RCCs are located in mitochondrial inner membranes (IM) and matrix, it was reasoned that EGCG could not readily pass through IM to affect RCCs in normal mitochondria but may do so when IM integrity is compromised. This speculation was substantiated in three ways. (1) Purified EGCG-bound proteins were barely detectable in normal mitochondria and contained no RCCs as determined by Western blotting, but swelling mitochondria contained about 1.5-fold more EGCG-bound proteins which included four RCC subunits together with cyclophilin D that locates in mitochondrial matrix. (2) Swelling mitochondria consumed more EGCG than normal ones. (3) The MPT blocker cyclosporine A diminished the above-mentioned difference. Among four subunits of RCC II, only SDHA and SDHB which locate in mitochondrial matrix, but not SDHC or SDHD which insert into the IM, were found to be EGCG targets. Interestingly, EGCG promoted Ca(2+) overload-induced MPT only when moderate MPT already commenced. This study identified hepatic RCCs as targets for EGCG in swelling but not normal mitochondria, suggesting EGCG may trigger hepatotoxicity by worsening pre-existing mitochondria abnormalities.

Assessment of the reproductive toxicity of inhalation exposure to ethyl tertiary butyl ether in male mice with normal, low active and inactive ALDH2.

No data are available regarding aldehyde dehydrogenase 2 (ALDH2) polymorphisms related to the reproductive toxicity possibly caused by ethyl tertiary butyl ether (ETBE). In this study, two inhalation experiments were performed in Aldh2 knockout (KO), heterogeneous (HT) and wild type (WT) C57BL/6 male mice exposed to ETBE, and the data about general toxicity, testicular histopathology, sperm head numbers, sperm motility and sperm DNA damage were collected. The results showed that the 13-week exposure to 0, 500, 1,750 and 5,000 ppm ETBE significantly decreased sperm motility and increased levels of sperm DNA strand breaks and 8-hydroxy-deoxyguanosine in both WT and KO mice, the effects were found in 1,750 and 5,000 ppm groups of WT mice, and all of the three exposed groups of KO mice compared to the corresponding control; furthermore, ETBE also caused decrease in the relative weights of testes and epididymides, the slight atrophy of seminiferous tubules of testis and reduction in sperm numbers of KO mice exposed to ≥500 ppm. In the experiment of exposure to lower concentrations of ETBE (0, 50, 200 and 500 ppm) for 9 weeks, the remarkable effects of ETBE on sperm head numbers, sperm motility and sperm DNA damage were further observed in KO and HT mice exposed to 200 ppm ETBE, but not in WT mice. Our findings suggested that only exposure to high concentrations of ETBE might result in reproductive toxicity in mice with normal active ALDH2, while low active and inactive ALDH2 enzyme significantly enhanced the ETBE-induced reproductive toxicity in mice, even exposed to low concentrations of ETBE, mainly due to the accumulation of acetaldehyde as a primary metabolite of ETBE.

MMBERT: a unified framework for biomedical named entity recognition.

Named entity recognition (NER) is an important task in natural language processing (NLP). In recent years, NER has attracted much attention in the biomedical field. However, due to the lack of biomedical named entity identification datasets, the complexity and rarity of biomedical named entities and so on, biomedical NER is more difficult than general domain NER. So in this paper, we propose a framework (MMBERT) based on Transformer to solve the problems above. To address the scarcity of biomedical named entity recognition datasets, we introduce ERNIE-Health, a new Chinese language representation model pre-trained on large-scale biomedical text corpora. Because of the complexity and rarity of biomedical named entities, we use the Bert and CW-LSTM structures to get the joint feature vector of word pairs relations. In addition, we design multi-granularity 2D convolution to refine the relationship and representation between word pairs. Finally, we design a convolutional neural network (CNN) structure and a co-predictor to improve the model's generalization capability and prediction accuracy. We have conducted extensive experiments on three benchmark datasets, and the experimental results show that our model achieves the best results compared with several baseline models in the experiment.

Novel antimicrobial peptides against Cutibacterium acnes designed by deep learning.

The increasing prevalence of antibiotic resistance in Cutibacterium acnes (C. acnes) requires the search for alternative therapeutic strategies. Antimicrobial peptides (AMPs) offer a promising avenue for the development of new treatments targeting C. acnes. In this study, to design peptides with the specific inhibitory activity against C. acnes, we employed a deep learning pipeline with generators and classifiers, using transfer learning and pretrained protein embeddings, trained on publicly available data. To enhance the training data specific to C. acnes inhibition, we constructed a phylogenetic tree. A panel of 42 novel generated linear peptides was then synthesized and experimentally evaluated for their antimicrobial selectivity and activity. Five of them demonstrated their high potency and selectivity against C. acnes with MIC of 2-4 µg/mL. Our findings highlight the potential of these designed peptides as promising candidates for anti-acne therapeutics and demonstrate the power of computational approaches for the rational design of targeted antimicrobial peptides.

Colon-Targeted Oral Delivery of Hydroxyethyl Starch-Curcumin Microcapsules Loaded with Multiple Drugs Alleviates DSS-Induced Ulcerative Colitis in Mice.

Combination therapy through colon-targeted oral delivery of multiple drugs presents a promising approach for effectively treating ulcerative colitis (UC). However, the codelivery of drugs with diverse physicochemical properties in a single formulation remains a formidable challenge. Here, microcapsules are designed based on hydroxyethyl starch-curcumin (HES─CUR) conjugates to enable the simultaneous delivery of hydrophobic dexamethasone acetate (DA) and hydrophilic cefazolin sodium (CS), yielding multiple drug-loaded microcapsules (CS/DA-loaded HES─CUR microcapsules, CDHC-MCs) tailored for colon-targeted therapy of UC. Thorough characterization confirms the successful synthesis and exceptional biocompatibility of CDHC-MCs. Biodistribution studies demonstrate that the microcapsules exhibit an impressive inflammatory targeting effect, accumulating preferentially in inflamed colons. In vivo experiments employing a dextran-sulfate-sodium-induced UC mouse model reveal that CDHC-MCs not only arrest UC progression but also facilitate the restoration of colon length and alleviate inflammation-related splenomegaly. These findings highlight the potential of colon-targeted delivery of multiple drugs within a single formulation as a promising strategy to enhance UC treatment, and the CDHC-MCs developed in this study hold great potential in developing novel oral formulations for advanced UC therapy.

Injectable Hydrogels with Integrated Ph Probes and Ultrasound-Responsive Microcapsules as Smart Wound Dressings for Visual Monitoring and On-Demand Treatment of Chronic Wounds.

Hydrogel dressings capable of infection monitoring and precise treatment administration show promise for advanced wound care. Existing methods involve embedd ingorganic dyes or flexible electronics into preformed hydrogels, which raise safety issues and adaptability challenges. In this study, an injectable hydrogel based smart wound dressing is developed by integrating food-derived anthocyanidin as a visual pH probe for infection monitoring and poly(L-lactic acid) microcapsules as ultrasound-responsive delivery systems for antibiotics into a poly(ethylene glycol) hydrogel. This straightforwardly prepared hydrogel dressing maintains its favorable properties for wound repair, including porous morphology and excellent biocompatibility. In vitro experiments demonstrated that the hydrogel enabled visual assessment of pH within the range of 5 âˆ¼ 9.Meanwhile, the release of antibiotics could be triggered and controlled by ultrasound. In vivo evaluations using infected wounds and diabetic wounds revealed that the wound dressing effectively detected wound infection by monitoring pH levels and achieved antibacterial effects through ultrasound-triggered drug release. This led to significantly enhanced wound healing, as validated by histological analysis and the measurement of inflammatory cytokine levels. This injectable hydrogel-based smart wound dressing holds great potential for use in clinical settings to inform timely and precise clinical intervention and in community to improve wound care management.

Colon-targeted hydroxyethyl starch-curcumin microspheres with high loading capacity ameliorate ulcerative colitis via alleviating oxidative stress, regulating inflammation, and modulating gut microbiota.

Curcumin (CUR) is a natural polyphenol that holds promise for treating ulcerative colitis (UC), yet oral administration of CUR exhibits limited bioavailability and existing formulations for oral delivery of CUR often suffer from unsatisfactory loading capacity. This study presents hydroxyethyl starch-curcumin microspheres (HC-MSs) with excellent CUR loading capacity (54.52 %), and the HC-MSs can further encapsulate anti-inflammatory drugs dexamethasone (DEX) to obtain a combination formulation (DHC-MSs) with high DEX loading capacity (19.91 %), for combination therapy of UC. The microspheres were successfully engineered, retaining the anti-oxidative and anti-inflammatory activities of parental CUR and demonstrating excellent biocompatibility and controlled release properties, notably triggered by α-amylase, facilitating targeted drug delivery to inflamed sites. In a mouse UC model induced by dextran sulfate sodium, the microspheres effectively accumulated in inflamed colons and both HC-MSs and DHC-MSs exhibited superior therapeutic efficacy in alleviating UC symptoms compared to free DEX. Moreover, mechanistic exploration uncovered the multifaceted therapeutic mechanisms of these formulations, encompassing anti-inflammatory actions, mitigation of spleen enlargement, and modulation of gut microbiota composition. These findings underscore the potential of HC-MSs and DHC-MSs as promising formulations for UC, with implications for advancing treatment modalities for various inflammatory bowel disorders.

Recent Advances in Polymer-Based Biosensors for Food Safety Detection.

The excessive use of pesticides and drugs, coupled with environmental pollution, has resulted in the persistence of contaminants on food. These pollutants tend to accumulate in humans through the food chain, posing a significant threat to human health. Therefore, it is crucial to develop rapid, low-cost, portable, and on-site biosensors for detecting food contaminants. Among various biosensors, polymer-based biosensors have emerged as promising probes for detection of food contaminants in recent years, due to their various functions such as target binding, enrichment, and simple signal reading. This paper aims to discuss the characteristics of five types of food pollutants-heavy metals, pesticide residues, pathogenic bacteria, allergens, and antibiotics-and their adverse effects on human health. Additionally, this paper focuses on the principle of polymer-based biosensors and their latest applications in detecting these five types of food contaminants in actual food samples. Furthermore, this review briefly examines the future prospects and challenges of biosensors for food safety detection. The insights provided in this review will facilitate the development of biosensors for food safety detection.

[Correlation between blood pressure indexes and prognosis in sepsis patients: a cohort study based on MIMIC-III database].

OBJECTIVE: To investigate the correlation between early-stage blood pressure indexes and prognosis in sepsis patients. METHODS: A retrospective cohort study was conducted on the medical records of patients diagnosed with sepsis from 2001 to 2012 in the Medical Information Mart for Intensive Care-III (MIMIC-III) database. Patients were divided into survival group and death group according to the 28-day prognosis. General data of patients and heart rate (HR) and blood pressure at admission to ICU and within 24 hours after admission were collected. The blood pressure indexes including the maximum, median and mean value of systolic index, diastolic index and mean arterial pressure (MAP) index were calculated. The data were randomly divided into training set and validation set (4 : 1). Univariate Logistic regression analysis was used to screen covariates, and multivariate Logistic stepwise regression models were further developed. Model 1 (including HR, blood pressure, and blood pressure index related variables with P < 0.1 and other variables with P < 0.05) and Model 2 (including HR, blood pressure, and blood pressure index related variables with P < 0.1) were developed respectively. The receiver operator characteristic curve (ROC curve), precision recall curve (PRC) and decision curve analysis (DCA) curve were used to evaluate the quality of the two models, and the influencing factors of the prognosis of sepsis patients were analyzed. Finally, nomogram model was developed according to the better model and effectiveness of it was evaluated. RESULTS: A total of 11 559 sepsis patients were included in the study, with 10 012 patients in the survival group and 1 547 patients in the death group. There were significant differences in age, survival time, Elixhauser comorbidity score and other 46 variables between the two groups (all P < 0.05). Thirty-seven variables were preliminarily screened by univariate Logistic regression analysis. After multivariate Logistic stepwise regression model screening, among the indicators related to HR, blood pressure and blood pressure index, the HR at admission to ICU [odds ratio (OR) = 0.992, 95% confidence interval (95%CI) was 0.988-0.997] and the maximum HR (OR = 1.006, 95%CI was 1.001-1.011), maximum MAP index (OR = 1.620, 95%CI was 1.244-2.126), mean diastolic index (OR = 0.283, 95%CI was 0.091-0.856), median systolic index (OR = 2.149, 95%CI was 0.805-4.461), median diastolic index (OR = 3.986, 95%CI was 1.376-11.758) were selected (all P < 0.1). There were 14 other variables with P < 0.05, including age, Elixhauser comorbidity score, continuous renal replacement therapy (CRRT), use of ventilator, sedation and analgesia, norepinephrine, norepinephrine, highest serum creatinine (SCr), maximum blood urea nitrogen (BUN), highest prothrombin time (PT), highest activated partial thromboplastin time (APTT), lowest platelet count (PLT), highest white blood cell count (WBC), minimum hemoglobin (Hb). The ROC curve showed that the area under the curve (AUC) of Model 1 and Model 2 were 0.769 and 0.637, respectively, indicating that model 1 had higher prediction accuracy. The PRC curve showed that the AUC of Model 1 and Model 2 were 0.381 and 0.240, respectively, indicating that Model 1 had a better effect. The DCA curve showed that when the threshold was 0-0.8 (the probability of death was 0-80%), the net benefit rate of Model 1 was higher than that of Model 2. The calibration curve showed that the prediction effect of the nomogram model developed according to Model 1 was in good agreement with the actual outcome. The Bootstrap verification results showed that the nomogram model was consistent with the above results and had good prediction effects. CONCLUSIONS: The nomogram model constructed has good prediction effects on the 28-day prognosis in sepsis patients, and the blood pressure indexes are important predictors in the model.

Deep Learning-Based System for Disease Screening and Pathologic Region Detection From Optical Coherence Tomography Images.

Purpose: This study was designed to apply deep learning models in retinal disease screening and lesion detection based on optical coherence tomography (OCT) images. Methods: We collected 37,138 OCT images from 775 patients and labelled by ophthalmologists. Multiple deep learning models including ResNet50 and YOLOv3 were developed to identify the types and locations of diseases or lesions based on the images. Results: The model were evaluated using patient-based independent holdout set. For binary classification of OCT images with or without lesions, the performance accuracy was 98.5%, sensitivity was 98.7%, specificity was 98.4%, and the F1 score was 97.7%. For multiclass multilabel disease classification, the models was able to detect vitreomacular traction syndrome and age-related macular degeneration both with an accuracy of more than 99%, sensitivity of more than 98%, specificity of more than 98%, and an F1 score of more than 97%. For lesion location detection, the recalls for different lesion types ranged from 87.0% (epiretinal membrane) to 98.2% (macular pucker). Conclusions: Deep learning-based models have potentials to aid retinal disease screening, classification and diagnosis with excellent performance, which may serve as useful references for ophthalmologists. Translational Relevance: The deep learning-based models are capable of identifying and predicting different eye diseases and lesions from OCT images and may have potential clinical application to assist the ophthalmologists for fast and accuracy retinal disease screening.

DDI-GCN: Drug-drug interaction prediction via explainable graph convolutional networks.

Drug-drug interactions (DDI) may lead to unexpected side effects, which is a growing concern in both academia and industry. Many DDIs have been reported, but the underlying mechanisms are not well understood. Predicting and understanding DDIs can help researchers to improve drug safety and protect patient health. Here, we introduce DDI-GCN, a method that utilizes graph convolutional networks (GCN) to predict DDIs based on chemical structures. We demonstrate that this method achieves state-of-the-art prediction performance on the independent hold-out set. It can also provide visualization of structural features associated with DDIs, which can help us to study the underlying mechanisms. To make it easy and accessible to use, we developed a web server for DDI-GCN, which is freely available at http://wengzq-lab.cn/ddi/.