Feature selection is crucial for disease classification and prognosis in high-dimensional microarray data, as it reduces dimensionality, enhances model accuracy, and improves computational efficiency. However, most existing methods rely on a global set of features and overlook the relationships between features and sample subspaces. To address this limitation, this study proposes a filter-based local feature selection method based on an immune algorithm (IA-FLFS). This method dynamically assigns a unique feature subset to each sample neighborhood, replacing reliance on a global feature subset. It incorporates three key innovations: (1) a single- stage, filter-based approach that considers feature interactions and is independent of any learning model, ensuring high effectiveness and efficiency for high-dimensional datasets; (2) an enhanced clonal selection algorithm is utilized to identify feature subsets, enhancing search capabilities through filter-based initialization, adaptive differential evolution-based mutation, and symmetric uncertainty-based local search. (3) thread-level parallelism is applied to each feature subset, significantly reducing computation time. Experimental results on twelve datasets demonstrate IA-FLFS's superior accuracy, efficiency, and ability to produce smaller feature subsets, outperforming fourteen state-of-the-art feature selection methods on most datasets. Notably, compared to other local feature selection algorithms, it achieves significant accuracy improvements on over eight datasets, highlighting its potential as a powerful and efficient tool for high-dimensional microarray analysis.

AIMS:To develop and validate a machine learning-based risk prediction model for delirium in older inpatients. DESIGN:A prospective cohort study. METHODS:A prospective cohort study was conducted. Eighteen clinical features were prospectively collected from electronic medical records during hospitalisation to inform the model. Four machine learning algorithms were employed to develop and validate risk prediction models. The performance of all models in the training and test sets was evaluated using a combination of the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, Brier score, and other metrics before selecting the best model for SHAP interpretation. RESULTS:A total of 973 older inpatient data were utilised for model construction and validation. The AUC of four machine learning models in the training and test sets ranged from 0.869 to 0.992; the accuracy ranged from 0.931 to 0.962; and the sensitivity ranged from 0.564 to 0.997. Compared to other models, the Random Forest model exhibited the best overall performance with an AUC of 0.908 (95% CI, 0.848, 0.968), an accuracy of 0.935, a sensitivity of 0.992, and a Brier score of 0.053. CONCLUSION:The machine learning model we developed and validated for predicting delirium in older inpatients demonstrated excellent predictive performance. This model has the potential to assist healthcare professionals in early diagnosis and support informed clinical decision-making. IMPACT:By identifying patients at risk of delirium early, healthcare professionals can implement preventive measures and timely interventions, potentially reducing the incidence and severity of delirium. The model's ability to support informed clinical decision-making can lead to more personalised and effective care strategies, ultimately benefiting both patients and healthcare providers. REPORTING METHOD:This study was reported in accordance with the TRIPOD statement. PATIENT OR PUBLIC CONTRIBUTION:No patient or public contribution.

Microarray data, characterized by high dimensionality and small sample sizes, poses significant challenges in identifying genes relevant for disease classification and prognosis. Our study proposes a one-stage filter local feature selection based on an immune algorithm to efficiently select relevant features in microarray data. It embeds two filter-based feature selection methods into an improved discrete immune algorithm and allocates feature subsets for different regions with considering local sample behaviors. The experimental results demonstrate the superiority of our method compared with well-known global and local feature selection methods on five microarray datasets.

Myocardial ischemia-reperfusion injury (MIRI) significantly worsens the outcomes of patients with cardiovascular diseases. Dexmedetomidine (Dex) is recognized for its cardioprotective properties, but the related mechanisms, especially regarding metabolic reprogramming, have not been fully clarified. A total of 60 patients with heart valve disease are randomly assigned to Dex or control group. Blood samples are collected to analyze cardiac injury biomarkers and metabolomics. In vivo and vitro rat models of MIRI are utilized to assess the effects of Dex on cardiac function, lactate production, and mitochondrial function. It is found that postoperative CK-MB and cTNT levels are significantly lower in the Dex group. Metabolomics reveals that Dex regulates metabolic reprogramming and reduces lactate level. In Dex-treated rats, the myocardial infarction area is reduced, and myocardial contractility is improved. Dex inhibits glycolysis, reduces lactate, and improves mitochondrial function following MIRI. Lactylation proteomics identifies that Dex reduces the lactylation of Malate Dehydrogenase 2（MDH2), thus alleviating myocardial injury. Further studies reveal that MDH2 lactylation induces ferroptosis, leading to MIRI by impairing mitochondrial function. Mechanistic analyses reveal that Dex upregulates Nuclear Receptor Subfamily 3 Group C Member 1（NR3C1) phosphorylation, downregulates Pyruvate Dehydrogenase Kinase 4 (PDK4), and reduces lactate production and MDH2 lactylation. These findings provide new therapeutic targets and mechanisms for the treatment for MIRI.

BackgroundPatients with resectable esophageal squamous cell carcinoma (ESCC) receiving neoadjuvant immunotherapy (NIT) display variable treatment responses. The purpose of this study is to establish and validate a radiomics based on enhanced computed tomography (CT) and combined with clinical data to predict the major pathological response to NIT in ESCC patients.MethodsThis retrospective study included 82 ESCC patients who were randomly divided into the training group (n = 57) and the validation group (n = 25). Radiomic features were derived from the tumor region in enhanced CT images obtained before treatment. After feature reduction and screening, radiomics was established. Logistic regression analysis was conducted to select clinical variables. The predictive model integrating radiomics and clinical data was constructed and presented as a nomogram. Area under curve (AUC) was applied to evaluate the predictive ability of the models, and decision curve analysis (DCA) and calibration curves were performed to test the application of the models.ResultsOne clinical data (radiotherapy) and 10 radiomic features were identified and applied for the predictive model. The radiomics integrated with clinical data could achieve excellent predictive performance, with AUC values of 0.93 (95% CI 0.87–0.99) and 0.85 (95% CI 0.69–1.00) in the training group and the validation group, respectively. DCA and calibration curves demonstrated a good clinical feasibility and utility of this model.ConclusionEnhanced CT image-based radiomics could predict the response of ESCC patients to NIT with high accuracy and robustness. The developed predictive model offers a valuable tool for assessing treatment efficacy prior to initiating therapy, thus providing individualized treatment regimens for patients.

Sepsis-induced myocardial dysfunction (SIMD) is a prevalent and severe form of organ dysfunction with elusive underlying mechanisms and limited treatment options.In this study, the cecal ligation and puncture and lipopolysaccharide (LPS) were used to reproduce sepsis model in vitro and vivo.The level of voltage-dependent anion channel 2 (VDAC2) malonylation and myocardial malonyl-CoA were detected by mass spectrometry and LC-MS-based metabolomics.Role of VDAC2 malonylation on cardiomyocytes ferroptosis and treatment effect of mitochondrial targeting nano material TPP-AAV were observed.The results showed that VDAC2 lysine malonylation was significantly elevated after sepsis.In addition, the regulation of VDAC2 lysine 46 (K46) malonylation by K46E and K46Q mutation affected mitochondrial-related ferroptosis and myocardial injury.The molecular dynamic simulation and circular dichroism further demonstrated that VDAC2 malonylation altered the N-terminus structure of the VDAC2 channel, causing mitochondrial dysfunction, increasing mitochondrial ROS levels, and leading to ferroptosis.Malonyl-CoA was identified as the primary inducer of VDAC2 malonylation.Furthermore, the inhibition of malonyl-CoA using ND-630 or ACC2 knock-down significantly reduced the malonylation of VDAC2, decreased the occurrence of ferroptosis in cardiomyocytes, and alleviated SIMD.The study also found that the inhibition of VDAC2 malonylation by synthesizing mitochondria targeting nano material TPP-AAV could further alleviate ferroptosis and myocardial dysfunction following sepsis.In summary, our findings indicated that VDAC2 malonylation plays a crucial role in SIMD and that targeting VDAC2 malonylation could be a potential treatment strategy for SIMD.

BackgroundTo identify differentially expressed lipid metabolism-related genes (DE-LMRGs) responsible for immune dysfunction in sepsis. MethodsThe lipid metabolism-related hub genes were screened using machine learning algorithms, and the immune cell infiltration of these hub genes were assessed by CIBERSORT and Single-sample GSEA. Next, the immune function of these hub genes at the single-cell level were validated by comparing multiregional immune landscapes between septic patients (SP) and healthy control (HC). Then, the support vector machine-recursive feature elimination (SVM-RFE) algorithm was conducted to compare the significantly altered metabolites critical to hub genes between SP and HC. Furthermore, the role of the key hub gene was verified in sepsis rats and LPS-induced cardiomyocytes, respectively. ResultsA total of 508 DE-LMRGs were identified between SP and HC, and 5 hub genes relevant to lipid metabolism (MAPK14, EPHX2, BMX, FCER1A, and PAFAH2) were screened. Then, we found an immunosuppressive microenvironment in sepsis. The role of hub genes in immune cells was further confirmed by the single-cell RNA landscape. Moreover, significantly altered metabolites were mainly enriched in lipid metabolism-related signaling pathways and were associated with MAPK14. Finally, inhibiting MAPK14 decreased the levels of inflammatory cytokines and improved the survival and myocardial injury of sepsis. ConclusionThe lipid metabolism-related hub genes may have great potential in prognosis prediction and precise treatment for sepsis patients.

Recently, artificial immune algorithms have attracted great attention of researchers and been widely used in function optimization, pattern recognition and classification. Traditional artificial immune algorithms for classification are applied to supervised learning problems, which require completely labeled data for training models. However, in many real-world scenarios, it is difficult to obtain all labeled samples. To solve this problem, an efficient semi-supervised artificial immune algorithm for classification tasks is proposed. It employs a clonal selection algorithm to generate memory cells used for classification, which is achieved via selection, cloning, and mutation procedures. Moreover, it utilizes the ensemble learning technique to extend the co-training paradigm and improves classification performance by adding the most confident unlabeled samples into the labeled set. In addition, the theory of learning from noisy examples is adopted to decide whether there are enough newly labeled samples that are used to reduce the negative effects caused by noises. Experimental results show that the proposed method achieves better or comparable performance than well-known semi-supervised and supervised methods on four datasets.

Background: The precise diagnostic and prognostic biological markers were needed in immunotherapy for sepsis. Considering the role of necroptosis and immune cell infiltration in sepsis, differentially expressed necroptosis-related genes (DE-NRGs) were identified, and the relationship between DE-NRGs and the immune microenvironment in sepsis was analyzed.Methods: Machine learning algorithms were applied for screening hub genes related to necroptosis in the training cohort. CIBERSORT algorithms were employed for immune infiltration landscape analysis. Then, the diagnostic value of these hub genes was verified by the receiver operating characteristic (ROC) curve and nomogram. In addition, consensus clustering was applied to divide the septic patients into different subgroups, and quantitative real-time PCR was used to detect the mRNA levels of the hub genes between septic patients (SP) (n = 30) and healthy controls (HC) (n = 15). Finally, a multivariate prediction model based on heart rate, temperature, white blood count and 4 hub genes was established.Results: A total of 47 DE-NRGs were identified between SP and HC and 4 hub genes (BACH2, GATA3, LEF1, and BCL2) relevant to necroptosis were screened out via multiple machine learning algorithms. The high diagnostic value of these hub genes was validated by the ROC curve and Nomogram model. Besides, the immune scores, correlation analysis and immune cell infiltrations suggested an immunosuppressive microenvironment in sepsis. Septic patients were divided into 2 clusters based on the expressions of hub genes using consensus clustering, and the immune microenvironment landscapes and immune function between the 2 clusters were significantly different. The mRNA levels of the 4 hub genes significantly decreased in SP as compared with HC. The area under the curve (AUC) was better in the multivariate prediction model than in other indicators.Conclusion: This study indicated that these necroptosis hub genes might have great potential in prognosis prediction and personalized immunotherapy for sepsis.

The required navigation performance (RNP) procedure is one of the two basic navigation specifications for the performance-based navigation (PBN) procedure as proposed by the International Civil Aviation Organization (ICAO) through an integration of the global navigation infrastructures to improve the utilization efficiency of airspace and reduce flight delays and the dependence on ground navigation facilities. The approach stage is one of the most important and difficult stages in the whole flying. In this study, we proposed deep reinforcement learning (DRL)-based RNP procedure execution, DRL-RNP. By conducting an RNP approach procedure, the DRL algorithm was implemented, using a fixed-wing aircraft to explore a path of minimum fuel consumption with reward under windy conditions in compliance with the RNP safety specifications. The experimental results have demonstrated that the six degrees of freedom aircraft controlled by the DRL algorithm can successfully complete the RNP procedure whilst meeting the safety specifications for protection areas and obstruction clearance altitude in the whole procedure. In addition, the potential path with minimum fuel consumption can be explored effectively. Hence, the DRL method can be used not only to implement the RNP procedure with a simulated aircraft but also to help the verification and evaluation of the RNP procedure.

在全民健身上升为国家战略的背景下,体育设施的合理规划和优化配置是推动全民健身、建设体育强国的基础条件.以上海市徐家汇体育公园为例,基于居民多元化的体育需求,利用手机上网数据,通过提取内容关键词来定位居民体育偏好,并定位用户上网的空间位置,对用户进行空间加权后的偏好分析,从而得到居民对各项体育运动的需求指数.旨在量化体育设施规模和种类,为徐家汇体育公园城市设计的科学布局提供定量支撑.以期为体育设施规模测算提供一种全新的定量化研究方法.同时基于需求的偏好分析有助于更科学地规划体育设施及优化用地布局,并逐步落实到各类各级规划中.

Conventional feature selection algorithms select a global feature subset for the entire sample space. In contrast, in this paper we propose an efficient filter local feature selection algorithm based on artificial immune system, which assigns a locally relevant feature subset for each neighboring region of the sample space. This algorithm introduces a clonal selection algorithm to explore the search space for the optimal feature subsets, and adopts local clustering idea as an evaluation criterion that maximizes the inter-class distance and minimizes the intra-class distance in the small region of each sample. Experimental results on a wide variety of synthetic and UCI datasets demonstrates that our proposed method achieves better performance than both state-of-the-art global feature selection algorithms and local feature selection algorithms. In addition, a main parameter analysis of the proposed method is carried out.

The clinical manifestations of patients with schizophrenia and patients with depression not only have a certain similarity, but also change with the patient's mood, and thus lead to misdiagnosis in clinical diagnosis. Electroencephalogram (EEG) analysis provides an important reference and objective basis for accurate differentiation and diagnosis between patients with schizophrenia and patients with depression. In order to solve the problem of misdiagnosis between patients with schizophrenia and patients with depression, and to improve the accuracy of the classification and diagnosis of these two diseases, in this study we extracted the resting-state EEG features from 100 patients with depression and 100 patients with schizophrenia, including information entropy, sample entropy and approximate entropy, statistical properties feature and relative power spectral density (rPSD) of each EEG rhythm (δ, θ, α, β). Then feature vectors were formed to classify these two types of patients using the support vector machine (SVM) and the naive Bayes (NB) classifier. Experimental results indicate that: ① The rPSD feature vector P performs the best in classification, achieving an average accuracy of 84.2% and a highest accuracy of 86.3%; ② The accuracy of SVM is obviously better than that of NB; ③ For the rPSD of each rhythm, the β rhythm performs the best with the highest accuracy of 76%; ④ Electrodes with large feature weight are mainly concentrated in the frontal lobe and parietal lobe. The results of this study indicate that the rPSD feature vector P in conjunction with SVM can effectively distinguish depression and schizophrenia, and can also play an auxiliary role in the relevant clinical diagnosis.

Negative selection algorithm is an important algorithm in the artificial immune system, inspired by the biological immune system. Traditional negative selection algorithms lack adaptive learning ability in high-dimensional space due to data sparsity and meaningless distance measurement. To solve these problems, an improved negative selection algorithm called Negative Selection Algorithm with Complete Random Subspace Technique (RS-NSA), is proposed in this paper. It adopts a bootstrap method to reduce the rate of misclassification resulting from the anomalies covered by the regions of normal samples. By using the complete random subspace technology, it reduces dimensionality to alleviate the curse of dimensionality. In addition, the ensemble learning technique is introduced to improve accuracy, in which component classifiers can be replaced by any negative selection algorithm. Empirical evaluation on UCI datasets reveals that, compared with V-detector, our proposed method can not only achieve a higher detection rate and a lower false alarm rate, but also shorten the training time.

Semi-supervised learning, which uses a large amount of unlabeled data to improve the performance of a classifier when only a limited amount of labeled data is available, has become a hot topic in machine learning research recently. In this paper, we propose a semi-supervised ensemble of classifiers approach, for learning in time-varying data streams. This algorithm maintains all the desirable properties of the semi-supervised Co-trained random FOREST algorithm (Co-Forest) and extends it into evolving data streams. It assigns a weight to each example according to Poisson(1) to simulate the bootstrap sample method in data streams, which is used to keep the diversity of Random Forest. By utilizing incremental learning technology, it avoids unnecessary repetition training and improves the accuracy of base models. In addition, the ADaptive WINdowing (ADWIN2) is introduced to deal with concept drift, which makes it adapt to the varying environment. Empirical evaluation on both synthetic data and UCI data reveals that our proposed method outperforms state-of-the-art semi-supervised and supervised methods in time-varying data streams, and also achieves relatively high performance in stationary streams.

The financial crisis, which began in 2007, led to the economic recession and restructuring of the global economy. Different industrial sectors of the economy and different economies have been affected in varying degrees which have a major impact on the command and control function of cities. Based on the Forbes "The Global 2000" database, this paper analyses the command and control centers in the world and China and their changing positions in the period of 2006-2014. It is found that the status of cities in developed countries have fallen while Chinese cities have risen. Beijing is the primate command and control center in China, its status has risen because of the rapid growth of financial sectors. How the industry sectors impact on these changing positions is examined. The final result is a classification of cities into three groups based on the complexity of their industry profiles. The top three command and control cities, i.e., Beijing, Hong Kong and Shanghai are compared according to their sectors.