The application of large language models in medicine: A scoping review.

This study systematically reviewed the application of large language models (LLMs) in medicine, analyzing 550 selected studies from a vast literature search. LLMs like ChatGPT transformed healthcare by enhancing diagnostics, medical writing, education, and project management. They assisted in drafting medical documents, creating training simulations, and streamlining research processes. Despite their growing utility in assisted diagnosis and improving doctor-patient communication, challenges persisted, including limitations in contextual understanding and the risk of over-reliance. The surge in LLM-related research indicated a focus on medical writing, diagnostics, and patient communication, but highlighted the need for careful integration, considering validation, ethical concerns, and the balance with traditional medical practice. Future research directions suggested a focus on multimodal LLMs, deeper algorithmic understanding, and ensuring responsible, effective use in healthcare.

A study on the catalytic domain of pork phospholipase A2: Enzymatic properties and hydrolysis characteristics of phosphatidylcholine and its hydroperoxide.

Endogenous phospholipase A2 (PLA2) plays an important role in phospholipids degradation during cured meat products manufacturing. The present study was undertaken to reveal more information about the endogenous PLA2 in muscles and its role in degradation of intramuscular phospholipids. With the catalytic domain of pork calcium-independent PLA2 (iPLA2cd), impacts of physic-chemical factors on the activity were investigated and substrate specificity of the enzyme were tested respectively. The optimum temperature and pH of pork iPLA2cd were 40 °C and 7.5, respectively. The iPLA2cd could be stimulated by adequate contents of NaCl and ATP, and inhibited by CaCl2 and NaNO2. For native phospholipids, the iPLA2cd was of a little higher affinity towards phosphatidylcholine (PC) than phosphatidylethanolamine (PE), phosphoserine (PS) and phosphatidylinositol (PI). The iPLA2cd could preferentially hydrolyze peroxidized PC over the native PC. The results would help better understand the degradation of phospholipids and the role played by endogenous enzymes during meat products manufacturing.

Highly accurate disease diagnosis and highly reproducible biomarker identification with PathFormer.

Biomarker identification is critical for precise disease diagnosis and understanding disease pathogenesis in omics data analysis, like using fold change and regression analysis. Graph neural networks (GNNs) have been the dominant deep learning model for analyzing graph-structured data. However, we found two major limitations of existing GNNs in omics data analysis, i.e., limited-prediction/diagnosis accuracy and limited-reproducible biomarker identification capacity across multiple datasets. The root of the challenges is the unique graph structure of biological signaling pathways, which consists of a large number of targets and intensive and complex signaling interactions among these targets. To resolve these two challenges, in this study, we presented a novel GNN model architecture, named PathFormer, which systematically integrate signaling network, priori knowledge and omics data to rank biomarkers and predict disease diagnosis. In the comparison results, PathFormer outperformed existing GNN models significantly in terms of highly accurate prediction capability (~30% accuracy improvement in disease diagnosis compared with existing GNN models) and high reproducibility of biomarker ranking across different datasets. The improvement was confirmed using two independent Alzheimer's Disease (AD) and cancer transcriptomic datasets. The PathFormer model can be directly applied to other omics data analysis studies.

Prediction and Verification of Potential Therapeutic Targets for Non-Responders to Infliximab in Ulcerative Colitis.

Background: Infliximab (IFX) has been widely used in ulcerative colitis (UC) patients. However, the subsequent effective treatment of IFX non-response in UC patients remains a challenge. This study aims to predict potential therapeutic targets for non-responders by performing a bioinformatic analysis of the data in the Gene Expression Omnibus (GEO) database and validation by biopsies. Methods: Colonic mucosal biopsies expression profiles of IFX-treated UC patients (GSE73661, GSE16879) were utilized to predict potential therapeutic targets. Bioinformatics analyses were used to explore potential biological mechanisms. CytoHubba was performed to screen hub genes. We used a validation dataset and colonic mucosal biopsies of UC patients to validate hub genes. Results: A total of 147 DEGs were identified (119 upregulated genes and 28 downregulated genes). GSEA showed that DEGs in GSE73661 were enriched in the pathways of the cytokine-cytokine receptor, the chemokine, and the adhesion molecules system. Based on the PPI network analysis, we identified four hub genes (and the transcription factor NF-κB). Then, we validate the expression of hub genes by reverse transcription-polymerase chain reaction (RT-PCR). We found higher expression of IL-6, IL1B, CXCL8, and CCL2 in non-responders compared to responders. Conclusion: In summary, four potential targets (IL-6, IL1B, CXCL8, and CCL2) were finally identified by performing a bioinformatics analysis of the datasets in the GEO database. Their expression was confirmed in colonic mucosal biopsies of patients with UC. These results can help to further explore the mechanism of non-responders to IFX in UC and to provide potential targets for their subsequent treatment.

Evaluation of the growth, adaption, and ecosystem services of two potentially-introduced urban tree species in Guangzhou under drought stress.

Under rapid urbanization and agglomeration of population, cities are facing various environmental challenges. As urban forests play a crucial role in mitigating native environmental problems and providing ecosystem services, cities might enhance their urban forest construction through multiple approaches, of which the introduction of exotic tree species could be an effective way. Under the background of constructing a high-quality forest city, Guangzhou was considering introducing a series of exotic tree species to improve the local urban greening, among which Tilia cordata Mill. and Tilia tomentosa Moench became the potential objects. As Guangzhou was reported to experience higher temperatures with less precipitation and face drought events with increasing frequency and intensity, whether the two tree species could survive in the dry environment required to be investigated profoundly. Thus, we launched a drought-simulation experiment and measured their above- and below-ground growth in 2020. In addition, their ecosystem services were also simulated and evaluated for their future adaption. Furthermore, a congeneric native tree species Tilia miqueliana Maxim was also measured in the same experiment as a comparison. Our results showed that Tilia miqueliana exhibited moderate patterns of growth and advantages in evapotranspiration and cooling. Besides, its investment in root development at horizontal level could account for its special strategy against drought stress. Tilia tomentosa's vigorous root growth could be the most positive behavior of coping with water deficit, which explained its maintenance of carbon fixation and implied a well adaption. Tilia cordata showed a complete decrease in above- and below-ground growth, especially for its fine root biomass. In addition, its ecosystem services were significantly reduced, reflecting a comprehensive failure when it faced a long-term scarcity of water. Therefore, it was necessary to supply sufficient water and under-ground space for their living in Guangzhou, especially for Tilia cordata. In the future, long-time observation of their growth under different stresses can be practical approaches to amplify their multiple ecosystem services.

Comparison of biochemical characteristics and gel properties of chicken myofibrillar protein affected by heme-iron and nonheme-iron oxidizing systems.

In this study, the effect of hemin and non-heme iron on the biochemical and gelling properties of chicken myofibrillar protein (MP) was compared. Results revealed that free radicals from hemin incubated MP were significantly higher than that in FeCl3 incubated samples (P < 0.05), and had higher ability to initiate protein oxidation. The carbonyl content, surface hydrophobicity, random coil increased with oxidant concentration, whereas the total sulfhydryl and α-helix content decreased in both oxidizing systems. The turbidity and particle size were increased after oxidant treatment, indicating oxidation promoted the cross-linking and aggregation of protein, and the degree of aggregation was higher in hemin treated MP compared with that incubated with FeCl3. The biochemical changes of MP resulted in an uneven and loose gel network structure, which significantly reduced the gel strength and water holding capacity (WHC) of the gel.

Lipidomic profile changes of yellow-feathered chicken meat during thermal processing based on UPLC-ESI-MS approach.

In this work, ultra-performance liquid chromatography-electrospray ionization mass spectrometry (UPLC-ESI-MS) was applied to compare the lipidomics profile of chicken breast meat heated to various temperatures. A total of 445 lipids including phospholipids, diacylglycerols, triacylglycerols, glycoglycerolipids, and sphingolipids were detected in chicken muscle. Significant decreases in a list of characteristic unsaturated phospholipid species including phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol and phosphatidylinositol were identified after heating, suggesting the occurrence of lipid oxidation, while significant increases in characteristic lysophospholipid species were found with increasing heating temperature, which were mainly derived from hydrolysis of phospholipids. Moreover, ether phospholipid and neutral lipid species exhibited remarkable changes during the heating process, and monoalk(en)yl diacylglycerol was first identified in chicken meat. The findings could contribute to the improved understanding of key lipids and biochemical reactions engaged in the heating of meat products.

Proteomic Analysis of Potential Targets for Non-Response to Infliximab in Patients With Ulcerative Colitis.

Background: Infliximab (IFX) is a potent therapeutic agent used for the treatment of conventional refractory ulcerative colitis (UC). However, the high non-response rate of IFX brings difficulties to clinical applications. In the context of proteomics research, our study of differentially expressed proteins (DEPs) is essential for non-response to IFX in UC patients and provides powerful insights into underlying drug resistance mechanisms. Methods: A total of 12 UC patients were divided into responders to IFX (UCinfG), non-responders to IFX (UCinfL), severe UC (UCsevere) without an IFX treatment history, and mild UC (UCmild) without an IFX treatment history. Subsequently, DEPs were identified from intestinal biopsy tissue between responders and non-responders to IFX by a label-free proteomic quantitative approach, and the general principle of functional protein screening was followed to deduce the potential drug targets and predictors for non-response to IFX in UC patients. Meanwhile, these targets excluded DEPs caused by the severity of inflammation for the first time. The differential expressions of candidate protein targets were validated at the gene sequence level using GEO2R analysis of the GEO database and qRT-PCR in some independent cohorts. Results: A total of 257 DEPs were screened out by mass spectrometry between UCinfG and UCinfL groups, excluding 22 DEPs caused by the severity of inflammation, and compared and verified at the gene sequence level in the Gene Expression Omnibus (GEO) database. Finally, five DEPs, including ACTBL2 (Q562R1), MBL2 (P11226), BPI (P17213), EIF3D (O15371), and CR1 (P17927), were identified as novel drug targets and predictive biomarkers for non-response to IFX. The drug targets were confirmed in the GEO database of the microarray results from three independent cohorts of 70 human intestinal biopsies and validated in qPCR data from 17 colonic mucosal biopsies. Among them, CR1 might affect the activation of the lectin pathway via complement-coated bacteria to play an opsonizing role in inflammation-related pathways closely associated with non-responders to IFX. Conclusion: This is the first report of proteomics analysis for the identification of novel drug targets based on intestinal biopsy tissue, which is significant for hypotheses for mechanistic investigation that are responsible for non-response to IFX and the development of clinical new pharmaceutical drugs.

Capsule endoscopy with retention of 4 years: A case report.

INTRODUCTION AND IMPORTANCE: Capsule endoscopy has been widely used in the diagnosis of small bowel diseases. Most CE can be smoothly excreted through the digestive tract. However, very few retention of CE may happen. CASE PRESENTATION: A 64-year-old man had been suffering from intermittent abdominal pain for 10 years. Capsule endoscopy was performed in local hospital 4 years ago. He was initially diagnosed with Crohn's disease and started on treatment. CTE and X-ray film of abdomen showed a suspected capsule endoscopy on the right side of pelvic cavity. Surgery was performed to remove the capsule. After the surgical treatment, no gastrointestinal symptoms relapsed for 9 months. CLINICAL DISCUSSION: It isn't uncommon for capsule endoscopy to be detained in Crohn's disease, because Strictures are the commonest complication of Crohn's disease. In order to prevent intestinal retention of capsule endoscopy, risk assessment should be carried out before capsule endoscopy. If detained CE isn't removed successfully by drug therapy and endoscopic therapy, surgery treatment has to be considered. CONCLUSION: In the present case, capsule endoscopy was found in the small intestine after 4 years, and the reason is worth pondering. We strongly recommend performing routine CTE, MRE and patency capsule examination before capsule endoscopy for patients suspected of stenosis. Routine abdominal X-ray film after examination is also useful for timely detection of capsule retention.

Recombinant Porcine 12-Lipoxygenase Catalytic Domain: Effect of Inhibitors, Selectivity of Substrates and Specificity of Oxidation Products of Linoleic Acid.

Lipoxygenase (LOX) is a major endogenous enzyme for the enzymatic oxidation of lipids during meat storage and meat product manufacturing. In the present work, some characteristics, i.e., effects of inhibitors, selectivity of substrates and specificity of oxidation products, were studied using recombinant porcine 12-lipoxygenase catalytic domain (12-LOXcd). Several familiar inhibitors were found inhibit the activity of recombinant porcine 12-LOXcd;nordihydroguaiaretic acid demonstrated the strongest inhibitory effect. The enzyme could oxygenate common polyunsaturated fatty acids, and showed the highest affinity to linoleic acid (LA), followed by arachidonic acid (AA), linolenic acid (LN) and docosahexaenoic acid (DHA). Under the action of porcine 12-LOXcd, LA was oxidized into four hydroxyoctadecadienoic acid (HODE) isomers, i.e., 13-Z,E-HODE, 13-E,E-HODE, 9-Z,E-HODE and 9-E,E-HODE. Variation of pH not only affected the yield of LA oxidation products, but also the distribution of HODE isomers. These results indicated that endogenous LOX activity and LOX-catalyzed lipid oxidation can be regulated during meat storage and meat product manufacturing.

Molecular Dynamics Simulation on the Effect of Self-Resistance Electric Heating on Carbon Fiber Surface Chemical Properties and Fiber/PP Interfacial Behavior.

Carbon fiber-reinforced thermoplastic (CFRT) composites have been dramatically employed in the automotive field on account of their superior performances, such as being light weight and high-strength. Self-resistance electric (SRE) heating provides a solution to the problem of high energy consumption in the conventional process of CFRT composites. The effect of SRE heating on the surface chemical properties of carbon fiber (CF) was investigated by X-ray photoelectron spectroscopy (XPS). XPS analysis suggests that the C-O-C epoxy group, the CF surface, would be degraded after SRE heating with strong current intensity, while there are weak changes in the content of -C-OH, -C-O-C-, -C-NH2 and -COOH groups with current intensity. The interfacial bonding properties and the radial distribution function (RDF) of CF-PP interfaces were carried out by molecular dynamics (MD) simulation. The simulation results show that the adhesion between the PP and the E44 sizing agent is weaker than that between CF and PP. There are no interaction modes between the PP and E44 sizing agent except van der Waals and electrostatic adsorption. The presence of the E44 sizing agent does not change the bonding mechanism at the interface of CF/PP.

Exploring the regulating mechanism of heat induced gelation of myosin by binding with Mb hemin prosthetic group.

The properties, structure and water holding capacity of myosin were analyzed after incubated with myoglobin (Mb) hemin prosthetic group. The results revealed moderate oxidation of hemin prosthetic group could improve the solubility of myosin. Besides, it could stretch the protein structure and cross-link the molecules to form the soluble polymer. Hence, moderate oxidation could improve the gel properties and the gel network structure. However, excessive oxidation would greatly reduce the physical and chemical properties of myosin, which was not conducive to the gel formation and would lead to a decrease in water retention. Moreover, fluorescence spectroscopy and cyclic voltammetry (CV) proved hemin prosthetic group had a high affinity for myosin. The interaction mechanism was further studied by molecular docking and molecular dynamics (MD) simulations. This study provides some fundamental prospects to be applied in the functional regulation of meat protein.

BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis.

Understanding which brain regions are related to a specific neurological disorder or cognitive stimuli has been an important area of neuroimaging research. We propose BrainGNN, a graph neural network (GNN) framework to analyze functional magnetic resonance images (fMRI) and discover neurological biomarkers. Considering the special property of brain graphs, we design novel ROI-aware graph convolutional (Ra-GConv) layers that leverage the topological and functional information of fMRI. Motivated by the need for transparency in medical image analysis, our BrainGNN contains ROI-selection pooling layers (R-pool) that highlight salient ROIs (nodes in the graph), so that we can infer which ROIs are important for prediction. Furthermore, we propose regularization terms-unit loss, topK pooling (TPK) loss and group-level consistency (GLC) loss-on pooling results to encourage reasonable ROI-selection and provide flexibility to encourage either fully individual- or patterns that agree with group-level data. We apply the BrainGNN framework on two independent fMRI datasets: an Autism Spectrum Disorder (ASD) fMRI dataset and data from the Human Connectome Project (HCP) 900 Subject Release. We investigate different choices of the hyper-parameters and show that BrainGNN outperforms the alternative fMRI image analysis methods in terms of four different evaluation metrics. The obtained community clustering and salient ROI detection results show a high correspondence with the previous neuroimaging-derived evidence of biomarkers for ASD and specific task states decoded for HCP. Our code is available at https://github.com/xxlya/BrainGNN_Pytorch.

Effects of combined ultrasound and low-temperature short-time heating pretreatment on proteases inactivation and textural quality of meat of yellow-feathered chickens.

This study aimed to investigate the effect of combined ultrasound and low-temperature short-time heating (ULTSTH) (40 kHz, 0.2 W/cm2 at 55 °C for 15 min) as pretreatment on proteases inactivation and textural quality of yellow-feathered chicken (YFC). Results showed ultrasound and low-temperature heating synergistically improved the inactivation of the most important meat proteases, calpain, cathepsin B and total proteases, with kinetics following the first order decay(s). Degradation of meat proteins was effectively reduced by ULTSTH compared to the pretreatment of chilling. Importantly, ULTSTH increased the firmness of breast meat and led to improved texture and microstructure. Lipid and protein oxidation of meat pretreated with ULTSTH were reduced during refrigerated storage period. Additionally, microorganisms in meat were inactivated by ULTSTH, which resulted in an obvious increase in the shelf life of meat. These findings suggested that ULTSTH is promising as an alternative pretreatment to obtain a favorable textural quality of YFC.

iTRAQ-based proteomic analysis of duck muscle related to lipid oxidation.

Lipid oxidation is the main cause of quality deterioration in meat and meat products. To facilitate the identification of candidate molecular biomarkers that are linked to lipid oxidation, we performed the proteomic analysis of duck muscle using isobaric tag for relative and absolute quantification (iTRAQ), followed by parallel reaction monitoring (PRM) to confirm the iTRAQ results. Pectoralis major muscles were divided into 2 groups in accordance with lipid oxidation, and iTRAQ-based analysis identified a total of 301 differentially expressed proteins, of which 15 proteins were examined by PRM assay. Proteins involved in lipid binding and metabolism, lipolysis, stress response, oxidative respiratory chain, and redox regulation were found to be differentially expressed between 2 groups and might affect lipid oxidation in muscles. The findings could contribute to the improved understanding of key proteins and processes engaged in lipid oxidation of meat.

Novel sample treatment method for the determination of free (E)-4-hydroxy-2-nonenal in meat products by liquid chromatography/tandem mass spectrometry using 4-hydroxy-2-nonenal-d3 as internal standard.

RATIONALE: (E)-4-Hydroxy-2-nonenal (HNE) is a reactive secondary product of lipid oxidation with biological significance. The analysis of HNE is a challenge due to its volatility and high activity. Developing sample preparation and analytical tools for the determination of free HNE is crucial for better understanding the actual level of free HNE in meat products. METHODS: Liquid nitrogen freezing, subzero-temperature extraction and derivatization were employed for meat sample treatment. Liquid chromatography/tandem mass spectrometry with electrospray ionization in negative ion mode was used for the determination of free HNE after isotope-coded derivatization. RESULTS: High repeatability and good recoveries with a limit of quantification as low as 0.25 pmol/g were found. Nineteen out of 24 samples, including chilled/processed meat products and meat-based instant foods, were found to contain free HNE with a range of 0.014-1.160 nmol/g. CONCLUSIONS: The proposed method showed satisfactory reliability, sensitivity and accuracy. We believe that such a sample preparation strategy will provide a powerful tool for better understanding the actual level of free HNE in meat products.

Evaluation of antioxidant property of heat shock protein 90 from duck muscle.

OBJECTIVE: The objectives of this study were to investigate the direct antioxidative effect of 90 Kda heat shock protein (Hsp90) obtained from duck muscle. METHODS: The interaction of Hsp90 with phospholipids and oxidized phospholipids was studied with surface plasmon resonance (SPR), and their further oxidation in the presence of Hsp90 was evaluated with thiobarbituric acid reactive substances (TBARS) assay. The scavenging effect on the 1,1-diphenyl-2-picrylhydrazyl (DPPH) and 2,2'-azinobis (3-ethylbenzthiazoline-6-sulfonic acid (ABTS) was measured, and the electron paramagnetic resonance (EPR) spectroscopy in combination with 5-tert-Butoxycarbonyl-5-methyl-1-pyrroline-N-oxide and 2-phenyl-4,4,5,5,-tetramethylimidazoline-1-oxyl-3-oxide (PTIO) was utilized to determine the abilities of Hsp90 in scavenging hydroxyl and PTIO radicals. RESULTS: SPR showed Hsp90 could bind with both phospholipids and oxidized phospholipids, and prevent their further oxidation by the TBARS assay. The DPPH and ABTS scavenging activity increased with Hsp90 concentration, and could reach 27% and 20% respectively at the protein concentration of 50 µM. The EPR spectra demonstrated Hsp90 could directly scavenge ·OH and PTIO· radicals. CONCLUSION: This suggests that Hsp90, a natural antioxidant in meat, may play an important role in cellular defense against oxidative stress, and may have potential use in meat products.

Effect of myoglobin, hemin, and ferric iron on quality of chicken breast meat.

OBJECTIVE: The objective was to evaluate the impact of different forms of iron including myoglobin, hemin, and ferric chloride on the quality of chicken breast meat. METHODS: Chicken breast muscles were subjected to 1, 2, 3 mg/mL of FeCl3, myoglobin and hemin treatment respectively, and the production of reactive oxygen species (ROS) and malondialdehyde, meat color, tenderness, water holding capacity and morphology of meat was evaluated. RESULTS: Hemin was found to produce more ROS and induce greater extent of lipid oxidation than myoglobin and ferric chloride. However, it showed that hemin could significantly increase the redness and decrease the lightness of the muscle. Hemin was also shown to be prominent in improving water holding capacity of meat, maintaining a relatively higher level of the immobilized water from low-field nuclear magnetic resonance measurements. Morphology observation by hematoxylin-eosin staining further confirmed the results that hemin preserved the integrity of the muscle. CONCLUSION: The results indicated that hemin may have economic benefit for the industry based on its advantage in improving water holding capacity and quality of meat.

Pooling Regularized Graph Neural Network for fMRI Biomarker Analysis.

Understanding how certain brain regions relate to a specific neurological disorder has been an important area of neuroimaging research. A promising approach to identify the salient regions is using Graph Neural Networks (GNNs), which can be used to analyze graph structured data, e.g. brain networks constructed by functional magnetic resonance imaging (fMRI). We propose an interpretable GNN framework with a novel salient region selection mechanism to determine neurological brain biomarkers associated with disorders. Specifically, we design novel regularized pooling layers that highlight salient regions of interests (ROIs) so that we can infer which ROIs are important to identify a certain disease based on the node pooling scores calculated by the pooling layers. Our proposed framework, Pooling Regularized-GNN (PR-GNN), encourages reasonable ROI-selection and provides flexibility to preserve either individual- or group-level patterns. We apply the PR-GNN framework on a Biopoint Autism Spectral Disorder (ASD) fMRI dataset. We investigate different choices of the hyperparameters and show that PR-GNN outperforms baseline methods in terms of classification accuracy. The salient ROI detection results show high correspondence with the previous neuroimaging-derived biomarkers for ASD.

Molecular Dynamics Simulation on the Interfacial Behavior of Over-molded Hybrid Fiber Reinforced Thermoplastic Composites.

Hybrid fiber reinforced thermoplastic composites are receiving important attention in lightweight applications. The fabrication process of hybrid thermoplastic composites is that discontinuous fiber reinforced thermoplastics are injected onto the continuous fiber reinforced thermoplastics by over-molding techniques. The key issue during this process is to get a reliable interfacial bonding strength. To understand the bonding mechanism at the heterogeneous interface of hybrid thermoplastic composites which is difficult to obtain through experimental investigations, a series of molecular dynamic (MD) simulations were conducted in this paper. The influence of processing parameters on the interfacial characteristics, i.e., the distribution of interfacial high-density enrichment areas, radius of gyration, diffusion coefficient and interfacial energy, were investigated during the forming process of a heterogeneous interface. Simulation results reveal that some of molecule chains get across the interface and tangle with the molecules from the other layer, resulting in the penetration phenomenon near the interface zone. In addition, the melting temperature and injection pressure exhibit positive effects on the interfacial properties of hybrid composites. To further investigate the interfacial bonding strength and fracture mechanism of the heterogeneous interface, the uniaxial tensile and sliding simulations were performed. Results show that the non-bonded interaction energy plays a crucial role during the fracture process of heterogeneous interface. Meanwhile, the failure mode of the heterogeneous interface was demonstrated to evolve with the processing parameters.