Gradient-flow adaptive importance sampling for Bayesian leave one out cross-validation for sigmoidal classification models.

We introduce a set of gradient-flow-guided adaptive importance sampling (IS) transformations to stabilize Monte-Carlo approximations of point-wise leave one out cross-validated (LOO) predictions for Bayesian classification models. One can leverage this methodology for assessing model generalizability by for instance computing a LOO analogue to the AIC or computing LOO ROC/PRC curves and derived metrics like the AUROC and AUPRC. By the calculus of variations and gradient flow, we derive two simple nonlinear single-step transformations that utilize gradient information to shift a model's pre-trained full-data posterior closer to the target LOO posterior predictive distributions. In doing so, the transformations stabilize importance weights. Because the transformations involve the gradient of the likelihood function, the resulting Monte Carlo integral depends on Jacobian determinants with respect to the model Hessian. We derive closed-form exact formulae for these Jacobian determinants in the cases of logistic regression and shallow ReLU-activated artificial neural networks, and provide a simple approximation that sidesteps the need to compute full Hessian matrices and their spectra. We test the methodology on an n≪p dataset that is known to produce unstable LOO IS weights.

Study of Impacts of Two Types of Cellular Aging on the Yeast Bud Morphogenesis.

Understanding the mechanisms of cellular aging processes is crucial for attempting to extend organismal lifespan and for studying age-related degenerative diseases. Yeast cells divide through budding, providing a classical biological model for studying cellular aging. With their powerful genetics, relatively short lifespan and well-established signaling pathways also found in animals, yeast cells offer valuable insights into the aging process. Recent experiments suggested the existence of two aging modes in yeast characterized by nucleolar and mitochondrial declines, respectively. In this study, by analyzing experimental data it was shown that cells evolving into those two aging modes behave differently when they are young. While buds grow linearly in both modes, cells that consistently generate spherical buds throughout their lifespan demonstrate greater efficacy in controlling bud size and growth rate at young ages. A three-dimensional chemical-mechanical model was developed and used to suggest and test hypothesized mechanisms of bud morphogenesis during aging. Experimentally calibrated simulations showed that tubular bud shape in one aging mode could be generated by locally inserting new materials at the bud tip guided by the polarized Cdc42 signal during the early stage of budding. Furthermore, the aspect ratio of the tubular bud could be stabilized during the late stage, as observed in experiments, through a reduction on the new cell surface material insertion or an expansion of the polarization site. Thus model simulations suggest the maintenance of new cell surface material insertion or chemical signal polarization could be weakened due to cellular aging in yeast and other cell types.

VSG-GAN: A high-fidelity image synthesis method with semantic manipulation in retinal fundus image.

In recent years, advancements in retinal image analysis, driven by machine learning and deep learning techniques, have enhanced disease detection and diagnosis through automated feature extraction. However, challenges persist, including limited data set diversity due to privacy concerns and imbalanced sample pairs, hindering effective model training. To address these issues, we introduce the vessel and style guided generative adversarial network (VSG-GAN), an innovative algorithm building upon the foundational concept of GAN. In VSG-GAN, a generator and discriminator engage in an adversarial process to produce realistic retinal images. Our approach decouples retinal image generation into distinct modules: the vascular skeleton and background style. Leveraging style transformation and GAN inversion, our proposed hierarchical variational autoencoder module generates retinal images with diverse morphological traits. In addition, the spatially adaptive denormalization module ensures consistency between input and generated images. We evaluate our model on MESSIDOR and RITE data sets using various metrics, including structural similarity index measure, inception score, Fréchet inception distance, and kernel inception distance. Our results demonstrate the superiority of VSG-GAN, outperforming existing methods across all evaluation assessments. This underscores its effectiveness in addressing data set limitations and imbalances. Our algorithm provides a novel solution to challenges in retinal image analysis by offering diverse and realistic retinal image generation. Implementing the VSG-GAN augmentation approach on downstream diabetic retinopathy classification tasks has shown enhanced disease diagnosis accuracy, further advancing the utility of machine learning in this domain.

Size-Related Pathway Flux Analysis of Ultrasmall Iron Oxide Nanoparticles in Macrophage Cell RAW264.7 for Safety Evaluation.

The endocytosis, intracellular transport, and exocytosis of different-sized nanoparticles were reported to greatly affect their efficacy and biosafety. The quantitation of endocytosis and exocytosis as well as subcellular distribution of nanoparticles might be an effective approach based on transport pathway flux analysis. Thus, the key parameters that could present the effects of three different-sized ultrasmall iron oxide nanoparticles (USIONPs) were systematically investigated in RAW264.7 cells. The endocytosis and exocytosis of USIONPs were related to their sizes; 15.4 nm of S2 could be quickly and more internalized and excreted in comparison to S1 (7.8 nm) and S3 (30.7 nm). In RAW264.7 cells, USIONPs were observed in endosomes, lysosomes, the Golgi apparatus, and autophagosomes via a transmission electron microscope. Based on flux analysis of intracellular transport pathways of USIONPs, it was found that 43% of S1, 40% of S2, and 44% of S3 were individually transported extracellularly through the Golgi apparatus-involved middle-fast pathway, while 24% of S1, 23% of S2, and 26% of S3 were transported through the fast recycling endosomal pathway, and the residues were transported through the slower speed lysosomal pathway. USIONPs might be transported via size-related endocytosis and exocytosis pathways. The pathway flux could be calculated on the basis of disturbance analysis of special transporters as well as their coding genes. Because there were rate differences among these transport pathways, this pathway flux could anticipate the intracellular remaining time and distribution of different-sized nanoparticles, the function exertion, and side effects of nanomaterials. The size of the nanomaterials could be optimized for improving functions and safety.

Astrocytes in ischemic stroke: Crosstalk in central nervous system and therapeutic potential.

In the central nervous system (CNS), a large group of glial cells called astrocytes play important roles in both physiological and disease conditions. Astrocytes participate in the formation of neurovascular units and interact closely with other cells of the CNS, such as microglia and neurons. Stroke is a global disease with high mortality and disability rate, most of which are ischemic stroke. Significant strides in understanding astrocytes have been made over the past few decades. Astrocytes respond strongly to ischemic stroke through a process known as activation or reactivity. Given the important role played by reactive astrocytes (RAs) in different spatial and temporal aspects of ischemic stroke, there is a growing interest in the potential therapeutic role of astrocytes. Currently, interventions targeting astrocytes, such as mediating astrocyte polarization, reducing edema, regulating glial scar formation, and reprogramming astrocytes, have been proven in modulating the progression of ischemic stroke. The aforementioned potential interventions on astrocytes and the crosstalk between astrocytes and other cells of the CNS will be summarized in this review.

Bone marrow mesenchymal stem cells therapy regulates sphingolipid and glycerophospholipid metabolism to promote neurological recovery in stroke rats: A metabolomics analysis.

Bone marrow mesenchymal stem cells (BMSCs) have therapeutic potential in the subacute/chronic phase of acute ischemic stroke (AIS), but the underlying mechanisms are not yet fully elucidated. There is a knowledge gap in understanding the metabolic mechanisms of BMSCs in stroke therapy. In this study, we administered BMSCs intravenously 24 h after reperfusion in rats with transient cerebral artery occlusion (MCAO). The treatment with BMSCs for 21 days significantly reduced the modified neurological severity score of MCAO rats (P < 0.01) and increased the number of surviving neurons in both the striatum and hippocampal dentate gyrus region (P < 0.01, respectively). Moreover, BMSCs treatment resulted in significant enhancements in various structural parameters of dendrites in layer V pyramidal neurons in the injured hemispheric motor cortex, including total length (P < 0.05), number of branches (P < 0.05), number of intersections (P < 0.01), and spine density (P < 0.05). Then, we performed plasma untargeted metabolomics analysis to study the metabolic changes of BMSCs on AIS. There were 65 differential metabolites identified in the BMSCs treatment group. Metabolic profiling analysis revealed that BMSCs modulate abnormal sphingolipid metabolism and glycerophospholipid metabolism, particularly affecting core members such as sphingomyelin (SM), ceramide (Cer) and sphingosine-1-phosphate (S1P). The metabolic network analysis and pathway-based compound-reaction-enzyme-gene network analysis showed that BMSCs inhibited the Cer-induced apoptotic pathway and promoted the S1P signaling pathway. These findings suggest that the enhanced effects of BMSCs on neuronal survival and synaptic plasticity after stroke may be mediated through these pathways. In conclusion, our study provides novel insight into the potential mechanisms of BMSCs treatment in stroke and sheds light on the possible clinical translation of BMSCs.

AIM2 inflammasome: A potential therapeutic target in ischemic stroke.

Ischemic stroke (IS) is a significant global public health issue with a high incidence, disability, and mortality rate. A robust inflammatory cascade with complex and wide-ranging mechanisms occurs following ischemic brain injury. Inflammasomes are multiprotein complexes in the cytoplasm that modulate the inflammatory response by releasing pro-inflammatory cytokines and inducing cellular pyroptosis. Among these inflammasomes, the Absent in Melanoma 2 (AIM2) inflammasome shows the ability to detect a wide range of pathogen DNAs, thereby triggering an inflammatory response. Recent studies have indicated that the aberrant expression of AIM2 inflammasome in various cells is closely associated with the pathological processes of ischemic brain injury. This paper summarizes the expression and regulatory role of AIM2 in CNS and peripheral immune cells and discusses current therapeutic approaches targeting AIM2 inflammasome. These findings aim to serve as a reference for future research in this field.

Cellular communication among smooth muscle cells: The role of membrane potential via connexins.

Communication via action potentials among neurons has been extensively studied. However, effective communication without action potentials is ubiquitous in biological systems, yet it has received much less attention in comparison. Multi-cellular communication among smooth muscles is crucial for regulating blood flow, for example. Understanding the mechanism of this non-action potential communication is critical in many cases, like synchronization of cellular activity, under normal and pathological conditions. In this paper, we employ a multi-scale asymptotic method to derive a macroscopic homogenized bidomain model from the microscopic electro-neutral (EN) model. This is achieved by considering different diffusion coefficients and incorporating nonlinear interface conditions. Subsequently, the homogenized macroscopic model is used to investigate communication in multi-cellular tissues. Our computational simulations reveal that the membrane potential of syncytia, formed by interconnected cells via connexins, plays a crucial role in propagating oscillations from one region to another, providing an effective means for fast cellular communication. Statement of Significance: In this study, we investigated cellular communication and ion transport in vascular smooth muscle cells, shedding light on their mechanisms under normal and abnormal conditions. Our research highlights the potential of mathematical models in understanding complex biological systems. We developed effective macroscale electro-neutral bi-domain ion transport models and examined their behavior in response to different stimuli. Our findings revealed the crucial role of connexinmediated membrane potential changes and demonstrated the effectiveness of cellular communication through syncytium membranes. Despite some limitations, our study provides valuable insights into these processes and emphasizes the importance of mathematical modeling in unraveling the complexities of cellular communication and ion transport.

Assessing the nonlinear association of environmental factors with antibiotic resistance genes (ARGs) in the Yangtze River Mouth, China.

The emergence of antibacterial resistance (ABR) is an urgent and complex public health challenge worldwide. Antibiotic resistant genes (ARGs) are considered as a new pollutant by the WHO because of their wide distribution and emerging prevalence. The role of environmental factors in developing ARGs in bacterial populations is still poorly understood. Therefore, the relationship between environmental factors and bacteria should be explored to combat ABR and propose more tailored solutions in a specific region. Here, we collected and analyzed surface water samples from Yangtze Delta, China during 2021, and assessed the nonlinear association of environmental factors with ARGs through a sigmoid model. A high abundance of ARGs was detected. Amoxicillin, phosphorus (P), chromium (Cr), manganese (Mn), calcium (Ca), and strontium (Sr) were found to be strongly associated with ARGs and identified as potential key contributors to ARG detection. Our findings suggest that the suppression of ARGs may be achieved by decreasing the concentration of phosphorus in surface water. Additionally, Group 2A light metals (e.g., magnesium and calcium) may be candidates for the development of eco-friendly reagents for controlling antibiotic resistance in the future.

The gut microbiome: an important role in neurodegenerative diseases and their therapeutic advances.

There are complex interactions between the gut and the brain. With increasing research on the relationship between gut microbiota and brain function, accumulated clinical and preclinical evidence suggests that gut microbiota is intimately involved in the pathogenesis of neurodegenerative diseases (NDs). Increasingly studies are beginning to focus on the association between gut microbiota and central nervous system (CNS) degenerative pathologies to find potential therapies for these refractory diseases. In this review, we summarize the changes in the gut microbiota in Alzheimer's disease, Parkinson's disease, multiple sclerosis, and amyotrophic lateral sclerosis and contribute to our understanding of the function of the gut microbiota in NDs and its possible involvement in the pathogenesis. We subsequently discuss therapeutic approaches targeting gut microbial abnormalities in these diseases, including antibiotics, diet, probiotics, and fecal microbiota transplantation (FMT). Furthermore, we summarize some completed and ongoing clinical trials of interventions with gut microbes for NDs, which may provide new ideas for studying NDs.

Mimicking effects of cholesterol in lipid bilayer membranes by self-assembled amphiphilic block copolymers.

The effect of cholesterol on biological membranes is important in biochemistry. In this study, a polymer system is used to simulate the consequences of varying cholesterol content in membranes. The system consists of an AB-diblock copolymer, a hydrophilic homopolymer hA, and a hydrophobic rigid homopolymer C, corresponding to phospholipid, water, and cholesterol, respectively. The effect of the C-polymer content on the membrane is studied within the framework of a self-consistent field model. The results show that the liquid-crystal behavior of B and C has a great influence on the chemical potential of cholesterol in bilayer membranes. The effects of the interaction strength between components, characterized by the Flory-Huggins parameters and the Maier-Saupe parameter, were studied. Some consequences of adding a coil headgroup to the C-rod are presented. Results of our model are compared to experimental findings for cholesterol-containing lipid bilayer membranes.

The interaction between intestinal microenvironment and stroke.

BACKGROUND: Stroke is not only a major cause of disability but also the third leading cause of death, following heart disease and cancer. It has been established that stroke causes permanent disability in 80% of survivors. However, current treatment options for this patient population are limited. Inflammation and immune response are major features that are well-recognized to occur after a stroke. The gastrointestinal tract hosts complex microbial communities, the largest pool of immune cells, and forms a bidirectional regulation brain-gut axis with the brain. Recent experimental and clinical studies have highlighted the importance of the relationship between the intestinal microenvironment and stroke. Over the years, the influence of the intestine on stroke has emerged as an important and dynamic research direction in biology and medicine. AIMS: In this review, we describe the structure and function of the intestinal microenvironment and highlight its cross-talk relationship with stroke. In addition, we discuss potential strategies aiming to target the intestinal microenvironment during stroke treatment. CONCLUSION: The structure and function of the intestinal environment can influence neurological function and cerebral ischemic outcome. Improving the intestinal microenvironment by targeting the gut microbiota may be a new direction in treating stroke.

Tetramethylpyrazine Attenuates Oxygen-glucose Deprivation-induced Neuronal Damage through Inhibition of the HIF-1α/BNIP3 Pathway: From Network Pharmacological Finding to Experimental Validation.

AIMS: A network pharmacological analysis combined with experimental validation was used to investigate the neuroprotective mechanism of the natural product Tetramethylpyrazine(TMP). BACKGROUND: Protecting neurons is critical for acute ischemic stroke treatment. Tetramethylpyrazine is a bioactive component extracted from Chuanxiong. The neuroprotective potential of TMP has been reported, but a systematic analysis of its mechanism has not been performed. OBJECTIVE: Based on the hints of network pharmacology and bioinformatics analysis, the mechanism by which TMP alleviates oxygen-glucose deprivation-induced neuronal damage through inhibition of the HIF-1α/BNIP3 pathway was verified. METHOD: In this study, we initially used network pharmacology and bioinformatics analyses to elucidate the mechanisms involved in TMP's predictive targets on a system level. The HIF-1α/BNIP3 pathway mediating the cellular response to hypoxia and apoptosis was considered worthy of focus in the bioinformatic analysis. An oxygen-glucose deprivation (OGD)-induced PC12 cell injury model was established for functional and mechanical validation. Cell viability, lactate dehydrogenase leakage, intracellular reactive oxygen species, percentage of apoptotic cells, and Caspase-3 activity were determined to assess the TMP's protective effects. Transfection with siRNA/HIF-1α or pcDNA/HIF-1α plasmids to silence or overexpress hypoxia-inducible factor 1α(HIF-1α). The role of HIF-1α in OGD-injured cells was observed first. After that, TMP's regulation of the HIF-1α/BNIP3 pathway was investigated. The pcDNA3.1/HIF-1α-positive plasmids were applied in rescue experiments. RESULT: The results showed that TMP dose-dependently attenuated OGD-induced cell injury. The expression levels of HIF-1α, BNIP3, and the Bax/Bcl-2 increased significantly with increasing OGD duration. Overexpression of HIF-1α decreased cell viability, increased BNIP3 expression, and Bax/Bcl-2 ratio; siRNA-HIF-1α showed the opposite effect. TMP treatment suppressed HIF-1α, BNIP3 expression, and the Bax/Bcl-2 ratio and was reversed by HIF-1α overexpression. CONCLUSION: Our study shows that TMP protects OGD-damaged PC12 cells by inhibiting the HIF-1α/BNIP3 pathway, which provides new insights into the mechanism of TMP and its neuroprotective potential.

Novel PLCL nanofibrous/keratin hydrogel bilayer wound dressing for skin wound repair.

In this study, a novel poly(L-lactate-caprolactone) copolymer (PLCL) nanofibrous/keratin hydrogel bilayer wound dressing loaded with fibroblast growth factor (FGF-2) was prepared by the low-pressure filtration-assisted method. The ability of the keratin hydrogel in the bilayer dressing to mimic the dermis and that of the nanofibrous PLCL to mimic the epidermis were discussed. Keratin hydrogel exhibited good porosity and maximum water absorption of 874.09%. Compared with that of the dressing prepared by the coating method, the interface of the bilayer dressing manufactured by the low-pressure filtration-assisted method (filtration time: 20 min) was tightly bonded, and its bilayer dressing interface could not be easily peeled off. The elastic modulus of hydrogel was about 44 kPa, which was similar to the elastic modulus of the dermis (2-80 kPa). Additionally, PLCL nanofibers had certain toughness and flexibility suitable for simulating the epidermal structures. In vitro studies showed that the bilayer dressing was biocompatible and biodegradable. In vivo studies indicated that PLCL/keratin-FGF-2 bilayer dressing could promote re-epithelialization, collagen deposition, skin appendages (hair follicles) regeneration, microangiogenesis construction, and adipose-derived stem cells (ADSCs) recruitment. The introduction of FGF-2 resulted in a better repair effect. The bilayer dressing also solved the problems of poor interface adhesion of hydrogel/electrospinning nanofibers. This paper also explored the preliminary role and mechanism of bilayer dressing in promoting skin healing, showing that its potential applications as a biomedical wound dressing in the field of skin tissue engineering.

Stem Cell-derived Extracellular Vesicles: A Promising Nano Delivery Platform to the Brain?

A very important cause of the frustration with drug therapy for central nervous system (CNS) diseases is the failure of drug delivery. The blood-brain barrier (BBB) prevents most therapeutic molecules from entering the brain while maintaining CNS homeostasis. Scientists are keen to develop new brain drug delivery systems to solve this dilemma. Extracellular vesicles (EVs), as a class of naturally derived nanoscale vesicles, have been extensively studied in drug delivery due to their superior properties. This review will briefly present current brain drug delivery strategies, including invasive and non-invasive techniques that target the brain, and the application of nanocarriers developed for brain drug delivery in recent years, especially EVs. The cellular origin of EVs affects the surface protein, size, yield, luminal composition, and other properties of EVs, which are also crucial in determining whether EVs are useful as drug carriers. Stem cell-derived EVs, which inherit the properties of parental cells and avoid the drawbacks of cell therapy, have always been favored by researchers. Thus, in this review, we will focus on the application of stem cell-derived EVs for drug delivery in the CNS. Various nucleic acids, proteins, and small-molecule drugs are loaded into EVs with or without modification and undergo targeted delivery to the brain to achieve their therapeutic effects. In addition, the challenges facing the clinical application of EVs as drug carriers will also be discussed. The directions of future efforts may be to improve drug loading efficiency and precise targeting.

Analysis of clinical predictors of kidney diseases in type 2 diabetes patients based on machine learning.

BACKGROUND: The heterogeneity of Type 2 Diabetes Mellitus (T2DM) complicated with renal diseases has not been fully understood in clinical practice. The purpose of the study was to propose potential predictive factors to identify diabetic kidney disease (DKD), nondiabetic kidney disease (NDKD), and DKD superimposed on NDKD (DKD + NDKD) in T2DM patients noninvasively and accurately. METHODS: Two hundred forty-one eligible patients confirmed by renal biopsy were enrolled in this retrospective, analytical study. The features composed of clinical and biochemical data prior to renal biopsy were extracted from patients' electronic medical records. Machine learning algorithms were used to distinguish among different kidney diseases pairwise. Feature variables selected in the developed model were evaluated. RESULTS: Logistic regression model achieved an accuracy of 0.8306 ± 0.0057 for DKD and NDKD classification. Hematocrit, diabetic retinopathy (DR), hematuria, platelet distribution width and history of hypertension were identified as important risk factors. Then SVM model allowed us to differentiate NDKD from DKD + NDKD with accuracy 0.8686 ± 0.052 where hematuria, diabetes duration, international normalized ratio (INR), D-Dimer, high-density lipoprotein cholesterol were the top risk factors. Finally, the logistic regression model indicated that DD-dimer, hematuria, INR, systolic pressure, DR were likely to be predictive factors to identify DKD with DKD + NDKD. CONCLUSION: Predictive factors were successfully identified among different renal diseases in type 2 diabetes patients via machine learning methods. More attention should be paid on the coagulation factors in the DKD + NDKD patients, which might indicate a hypercoagulable state and an increased risk of thrombosis.

Coupled chemical reactions: Effects of electric field, diffusion, and boundary control.

Chemical reactions involve the movement of charges, and this paper presents a mathematical model for describing chemical reactions in electrolytes. The model is developed using an energy variational method that aligns with classical thermodynamics principles. It encompasses both electrostatics and chemical reactions within consistently defined energetic and dissipative functionals. Furthermore, the energy variation method is extended to account for open systems that involve the input and output of charge and mass. Such open systems have the capability to convert one form of input energy into another form of output energy. In particular, a two-domain model is developed to study a reaction system with self-regulation and internal switching, which plays a vital role in the electron transport chain of mitochondria responsible for ATP generation-a crucial process for sustaining life. Simulations are conducted to explore the influence of electric potential on reaction rates and switching dynamics within the two-domain system. It shows that the electric potential inhibits the oxidation reaction while accelerating the reduction reaction.

Development of a Real-Time Recombinase-Aided Amplification Method to Rapidly Detect Methicillin-Resistant Staphylococcus aureus.

Methicillin-resistant staphylococcus aureus (MRSA) is a major pathogen responsible for human hospital and community-onset diseases and severe invasive livestock infections. Rapid detection of MRSA is essential to control the spread of MRSA. Conventional identification methods and antibacterial susceptibility tests of MRSA are time-consuming. The commonly used qPCR assay also has the disadvantages of being complicated and expensive, restricting its application in resource-limited clinical laboratories. Here, a real-time fluorescent recombinase-assisted amplification (RAA) assay targeting the most conserved regions within the mecA gene of MRSA was developed and evaluated to detect MRSA. The detection limit of this assay was determined to be 10 copies/reaction of positive plasmids. The established RAA assay showed high specificity for MRSA detection without cross-reactivities with other clinically relevant bacteria. The diagnostic performance of real-time RAA was evaluated using 67 clinical S. aureus isolates from dairy farms, which were detected in parallel using the TaqMan probe qPCR assay. The results showed that 56 and 54 samples tested positive for MRSA by RAA and qPCR, respectively. The overall agreement between both assays was 97.01% (65/67), with a kappa value of 0.9517 (p < 0.001). Further linear regression analysis demonstrated that the detection results between the two assays were significantly correlated (R2 = 0.9012, p < 0.0001), indicating that this RAA assay possesses similar detection performance to the qPCR assay. In conclusion, our newly established RAA assay is a time-saving and convenient diagnostic tool suitable for MRSA detection and screening.

Multi-objective learning and explanation for stroke risk assessment in Shanxi province.

Stroke is the leading cause of death in China (Zhou et al. in The Lancet, 2019). A dataset from Shanxi Province is analyzed to predict the risk of patients at four states (low/medium/high/attack) and to estimate transition probabilities between various states via a SHAP DeepExplainer. To handle the issues related to an imbalanced sample set, the quadratic interactive deep model (QIDeep) was first proposed by flexible selection and appending of quadratic interactive features. The experimental results showed that the QIDeep model with 3 interactive features achieved the state-of-the-art accuracy 83.33%(95% CI (83.14%; 83.52%)). Blood pressure, physical inactivity, smoking, weight, and total cholesterol are the top five most important features. For the sake of high recall in the attack state, stroke occurrence prediction is considered an auxiliary objective in multi-objective learning. The prediction accuracy was improved, while the recall of the attack state was increased by 17.79% (to 82.06%) compared to QIDeep (from 71.49%) with the same features. The prediction model and analysis tool in this paper provided not only a prediction method but also an attribution explanation of the risk states and transition direction of each patient, a valuable tool for doctors to analyze and diagnose the disease.