Multiple Algorithms Research Articles

Connectivity and community structure of online and register-based social networks

The dominance of online social media data as a source for large-scale social network studies has recently been challenged by networks constructed from state-curated register data. In this paper focused on the cross-comparison of the network structures, we investigate the similarities and differences of the Dutch online social network (OSN) Hyves and a register-based social network (RSN) of the Netherlands. First and foremost, we find that node metrics and the connectivity of the two population-scale networks are similar, with more long-distance ties captured by the OSN, and with the OSN ties proving to be predictive of RSN ties. These results hold when correcting for population size and geographical distance, notwithstanding that these two factors appear to be the main drivers of connectivity. Second, we show using multiple algorithms that the community structure of the two networks is similar and that neither follows strict administrative geographical delineations (e.g., provinces). Instead, communities appear to either center around large metropolitan areas or, outside of the country’s most urbanized area, comprise large blocks of interdependent municipalities. Beyond population and distance-related patterns, communities also highlight the persistence of deeply rooted sociocultural communities such as the Dutch Bible belt. The findings presented in this work aid in interpreting results from future studies in which register-based social networks are used to obtain insights into the social network structure of an entire population.

Harnessing Unsupervised Ensemble Learning for Biomedical Applications: A Review of Methods and Advances

Advancements in data availability and computational techniques, including machine learning, have transformed the field of bioinformatics, enabling the robust analysis of complex, high-dimensional, and heterogeneous biomedical data. This paper explores how diverse bioinformatics tasks, including differential expression analysis, network inference, and somatic mutation calling, can be reframed as binary classification tasks, thereby providing a unifying framework for their analysis. Traditional single-method approaches often fail to generalize across datasets due to differences in data distributions, noise levels, and underlying biological contexts. Ensemble learning, particularly unsupervised ensemble approaches, emerges as a compelling solution by integrating predictions from multiple algorithms to leverage their strengths and mitigate weaknesses. This review focuses on the principles and recent advancements in ensemble learning, with a particular emphasis on unsupervised ensemble methods. These approaches demonstrate their ability to address critical challenges in bioinformatics, such as the lack of labeled data and the integration of predictions from algorithms operating on different scales. Overall, this paper highlights the transformative potential of ensemble learning in advancing predictive accuracy, robustness, and interpretability across diverse bioinformatics applications.

Unrolled deep learning for breast cancer detection using limited-view photoacoustic tomography data.

Photoacoustic tomography (PAT) has emerged as a promising imaging modality for breast cancer detection, offering unique advantages in visualizing tissue composition without ionizing radiation. However, limited-view scenarios in clinical settings present significant challenges for image reconstruction quality and computational efficiency. This paper introduces novel unrolled deep learning networks based on split Bregman total variation (SBTV) and relaxed basis pursuit alternating direction method of multipliers (rBP-ADMM) algorithms to address these challenges. Our approach combines transfer learning from full-view to limited-view scenarios with U-Net denoiser integration, achieving state-of-the-art reconstruction quality (MS-SSIM> 0.95) while reducing reconstruction time by 92% compared to traditional methods. The effectiveness of different sensor configurations is analyzed through restricted isometry property (RIP) analysis and coherence values, demonstrating that semicircular arrays achieve a RIP constant of 0.76 and coherence of 0.77, closely approximating full-view performance (RIP: 0.75, coherence: 0.78). These metrics validate the theoretical foundation for accurate sparse signal recovery in limited-view scenarios. Comprehensive evaluations across semicircular, concave, and convex sensor arrangements show that the proposed U-SBTV network consistently outperforms existing methods, particularly when combined with the U-Net denoiser. This advancement in limited-view PAT reconstruction brings the technology closer to practical clinical application, potentially improving early breast cancer detection capabilities.

Path planning and optimization for transmission line barrier operations in complex terrain based on multimachine collaborative control

Abstract To address the challenges of path planning for transmission line barrier operations in complex terrains, this paper proposes a solution based on multimachine collaborative control. A path-planning algorithm that integrates multiple classical algorithms and incorporates multiobjective optimization methods is constructed to handle the complexities of challenging terrains. Auction algorithm and genetic algorithm are adopted to achieve optimal load balance by dynamically adjusting task allocation. Simulation results show that the proposed method effectively enhances task execution efficiency, optimizes path planning, and reduces energy consumption in complex terrains and dynamic environments, while maintaining high-obstacle avoidance success rates and communication stability.

Health Insurance Premium Prediction Based on Multiple Machine Learning Algorithms

The purpose of this paper is to predict the health insurance premium through a variety of machine learning algorithms, and compare and analyze the prediction effect of different algorithms. An open source dataset was selected for the study, and the experiments involved three machine learning models: linear regression, decision trees, and random forests. By testing these models, we obtain their performance in health insurance premium forecasts. The results show that the prediction performance of the random forest regression model is better than other models, and its score reaches 0.8564, which is the best algorithm among the three models. Second, the linear regression model has a score of 0.7584, and although its performance is not as good as that of random forest, it still shows some predictive power. Finally, the prediction effect of decision tree model is relatively poor, and the score is only 0.7097. To sum up, the experiments in this paper prove that the random forest model is undoubtedly the best choice in the prediction of health insurance premiums, which not only has good prediction accuracy, but also shows strong data processing ability. In the context of the growing importance of health insurance premium collection and analysis, the use of advanced machine learning algorithms such as Random Forest for forecasting will, to some extent, help insurance companies better price and assess risk. Therefore, it can be concluded that random forest regression model has the best performance for health insurance premium prediction and is an effective tool to achieve accurate prediction.

Optimization of Distributed UAV Swarm Placement for Target Localization Using Multiple Heuristic Algorithms Based on Compressive Sensing

Collaborative UAV swarms are increasingly deployed as temporary base stations in emergency situations to relay communications. This paper designs intelligent optimization algorithms, which minimizing mutual coherence within a region of interest (ROI) aims to solving challenge in these scenarios is optimizing the UAVs' locations to avoid mutual signal interference and ensure high communication quality. Minimizing mutual coherence reduces the likelihood of signal interference between UAVs. We explore various optimize algorithum, including Heuristic Search (HS) and Ant Colony Optimization (ACO) and so on. The results provide insights into each algorithm's performance in dynamic environments, helping to identify the most suitable approaches for UAV deployment in emergency scenarios. This study contributes to the development of efficient UAV deployment strategies, enhancing the reliability of UAV-based communication systems during critical events.

Identification of potential biomarkers for 2022 Mpox virus infection: a transcriptomic network analysis and machine learning approach

Monkeypox virus (MPXV), a zoonotic pathogen, re-emerged in 2022 with the Clade IIb variant, raising global health concerns due to its unprecedented spread in non-endemic regions. Recent studies have shown that Clade IIb (2022 MPXV) is marked by unique genomic mutations and epidemiological behaviors, suggesting variations in host-virus interactions. This study aimed to identify the differentially expressed genes (DEGs) induced by the 2022 MPXV infection through comprehensive bioinformatics analyses of microarray and RNA-Seq datasets from post-infected cell types with different MPXV clades. Subsequently, gene expression network analyses pinpoint the key DEGs, followed by their candidate drug assessment using the Drug SIGnatures DataBase (DSigDB) and validation by multiple machine learning algorithms. Comparative differential gene expression (DGE) analysis revealed 798 DEGs exclusive to the 2022 MPXV invasion in the skin cell types (keratinocytes). Intriguingly, 13 key DEGs were identified across hubs and clusters, highlighting their aberrant expressions in cell cycle regulation, immune responses, and cancer pathways. Biomarker screening via Random Forest (RF) model (selected with PyCaret from multiple models) and validation through t-distributed stochastic neighbor embedding (t-SNE) algorithm, principal component analysis (PCA), and ROC curve analysis employing Logistic Regression and Random Forest, identified 6 key DEGs (TXNRD1, CCNB1, BUB1, CDC20, BUB1B, and CCNA2) as promising biomarkers (AUC > 0.7) for clade IIb infection. This study anticipates that further investigation and clinical trials will catalyze novel detection and therapeutic options to combat 2022 MPXV infection in humans.

Efficient Workflow Scheduling in Cloud Computing Using Hybrid Algorithm

In recent years, cloud computing has increasingly embraced a pay-per-use model, offering dynamic, virtualized resources via the internet. The central challenge in this domain is efficiently scheduling workflows, considering deadlines and budgets while optimizing task allocation to virtual machines (VMs). Our study hypothesizes that improved scheduling can reduce energy consumption, streamline process execution, and lower operational costs. To test this hypothesis, we conducted a comparative analysis of two optimization techniques: Genetic Algorithm with Multiple Particle Swarm Optimization (GA + MPSO) and Genetic Algorithm with Bat Algorithm (GA + BAT). The analysis reveals that the combination of Genetic Algorithm and Bat Algorithm (GA + BAT) excels in optimizing cloud computing workflow scheduling. GA + BAT demonstrated superior performance by significantly reducing energy consumption, shortening process execution times, and decreasing operational costs. These findings validate our hypothesis, underscoring that optimizing cloud computing workflow scheduling can deliver substantial benefits. Consequently, by adopting GA + BAT, cloud service providers can enhance efficiency, reduce costs, and foster a more sustainable and responsive cloud infrastructure.

Single-cell profiling of SLC family transporters: uncovering the role of SLC7A1 in osteosarcoma

BackgroundOsteosarcoma is the most common malignant bone tumor in children and adolescents, characterized by high disability and mortality rates. Over the past three decades, therapeutic outcomes have plateaued, underscoring the critical need for innovative therapeutic targets. Solute carrier (SLC) family transporters have been implicated in the malignant progression of a variety of tumors, however, their specific role in osteosarcoma remains poorly understood.MethodsThe single-cell sequencing data from GSE152048 and GSE162454, along with RNA-seq from the TARGET and GSE21257 cohorts, were utilized for the analysis in this study. LASSO regression analysis was conducted to identify prognostic genes and construct an SLC-related prognostic signature. Survival analysis and ROC analysis evaluated the validity of the prognostic signature. The ESTIMATE and CIBERSORT Packages were utilized to assess the immune infiltration status. Pseudotime and CellChat analyses were performed to investigate the relationship between SLC7A1, malignant phenotypes, and the immune microenvironment. CCK8 assays, EdU staining, colony formation assays, Transwell assays, and co-culture systems were used to assess the effects of SLC7A1 on cell proliferation, metastasis, and macrophage polarization. Finally, virtual docking identified potential drugs targeting SLC7A1.ResultsSLCs displayed distinct expression patterns across various cell types within the osteosarcoma microenvironment, with myeloid cells exhibiting a preference for amino acid uptake. A prognostic model comprising nine genes was constructed via LASSO regression, with SLC7A1 showing the highest hazard ratio. Multiple analytical algorithms indicated that SLCs were associated with immune cell infiltration and immune checkpoint gene expression. Single-cell analysis indicated that SLC7A1 was predominantly expressed in osteosarcoma cells and correlated with various malignant tumor characteristics. SLC7A1 also regulate interactions between tumor cells and macrophages, as well as modulate macrophage function through multiple pathways. In vitro assays and survival analysis demonstrated that inhibition of SLC7A1 suppressed the malignant phenotype of osteosarcoma cells, with SLC7A1 expression correlating with poor prognosis. Co-culture models confirmed the involvement of SLC7A1 in macrophage polarization. Finally, virtual screening and CETSA identified Cepharanthine as potential inhibitors of SLC7A1.ConclusionSLC-related prognostic signatures can be utilized for the prognostic evaluation of osteosarcoma. Pharmacological inhibition of SLC7A1 may be a feasible therapeutic approach for osteosarcoma.

PeptideForest: Semisupervised Machine Learning Integrating Multiple Search Engines for Peptide Identification.

The first step in bottom-up proteomics is the assignment of measured fragmentation mass spectra to peptide sequences, also known as peptide spectrum matches. In recent years novel algorithms have pushed the assignment to new heights; unfortunately, different algorithms come with different strengths and weaknesses and choosing the appropriate algorithm poses a challenge for the user. Here we introduce PeptideForest, a semisupervised machine learning approach that integrates the assignments of multiple algorithms to train a random forest classifier to alleviate that issue. Additionally, PeptideForest increases the number of peptide-to-spectrum matches that exhibit a q-value lower than 1% by 25.2 ± 1.6% compared to MS-GF+ data on samples containing mixed HEK and Escherichia coli proteomes. However, an increase in quantity does not necessarily reflect an increase in quality and this is why we devised a novel approach to determine the quality of the assigned spectra through TMT quantification of samples with known ground truths. Thereby, we could show that the increase in PSMs below 1% q-value does not come with a decrease in quantification quality and as such PeptideForest offers a possibility to gain deeper insights into bottom-up proteomics. PeptideForest has been integrated into our pipeline framework Ursgal and can therefore be combined with a wide array of algorithms.

Lipid Metabolism Related Gene ACSL3 as a Biomarker for Predicting Immunotherapy Outcomes in Lung Adenocarcinoma.

Investigate the role of lipid metabolism in the tumor immune microenvironment (TIME) of lung adenocarcinoma (LUAD) and identify vital lipid metabolism-related genes (LMRGs) that contribute to immunotherapy outcomes. 1130 LUAD patients were acquired utilizing public databases. Multiple algorithms were used to analyze the contribution of lipid metabolism in TIME. Importantly, cell lines, clinical samples (52 patients in surgery cohort and 36 in immunotherapy cohort), animal models, RNA-seq, experiments in protein and mRNA levels were conducted for identifying and validating key biomarker in LUAD immunotherapy. A prognostic signature comprising 33 LMRGs was developed and validated as an effective predictor of prognosis and TIME, with a C-index of 0.766 (95% CI: 0.729-0.804). Additionally, we identified Acyl-CoA Synthetase Long Chain Family Member 3 (ACSL3) as a potential biomarker for immunotherapy prognosis. The expression of ACSL3 was verified in 88 clinical tissues from LUAD patients, which indicated that elevated ACSL3 expression was correlated with worse progression-free survival (PFS) (p<0.001) and overall survival (OS) (p=0.008). Subsequent experiments revealed that knockdown of ACSL3 in vivo enhanced the efficacy of immunotherapy, potentially through increasing interferon α secretion, as indicated by Bulk RNA-seq and ELISA analysis, thereby promoting the infiltration of anti-tumor immune cells. The study established a model that accurately predicts immunotherapy response, prognosis, and TIME dynamics in LUAD patients. Notably, the pivotal role of ACSL3 in driving tumor progression and immune evasion was uncovered, offering novel insights into the optimization of immunotherapy strategies for LUAD.

Comparing innovative AI algorithms to assess echocardiographic videos for clinical modelling.

We conduct a comparative study that employs the use of multiple dynamic deep learning algorithms to develop predictive models with video-based echocardiographic images using sample size determination as a key variable to assess optimal performance metrics. Our study compares performance of 3D convolutional neural networks, video vision transformers, and hybrid convolutional neural networks and Long Short-Term Memory models within both supervised and semi-supervised domains using variable sample sizes. For supervised learning, the ResNet3D model achieved the lowest Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) across all training set sizes (200, 400, and 800-video datasets), with the best performance observed on the 800-video training set (MAE: 7.409, RMSE: 10.216). In the semi-supervised setting, both ResNet3D and ResNet+LSTM models benefited from the inclusion of unlabelled data, particularly at larger dataset sizes. Because semi-supervised models use both labelled and unlabelled data sets, our findings are significant in showing that performance of certain predictive models using mixtures of unlabelled and labelled data is comparable to those using only labelled data with similar sample sizes, thus obviating the need for large sample sizes of labelled data.

The potential of short-wave infrared hyperspectral imaging and deep learning for dietary assessment: a prototype on predicting closed sandwiches fillings

IntroductionAccurate measurement of dietary intake without interfering in natural eating habits is a long-standing problem in nutritional epidemiology. We explored the applicability of hyperspectral imaging and machine learning for dietary assessment of home-prepared meals, by building a proof-of-concept, which automatically detects food ingredients inside closed sandwiches.MethodsIndividual spectra were selected from 24 hyperspectral images of assembled closed sandwiches, measured in a spectral range of 1116.14 nm to 1670.62 nm over 108 bands, pre-processed with Standard Normal Variate filtering, derivatives, and subsampling, and fed into multiple algorithms, among which PLS-DA, multiple classifiers, and a simple neural network.ResultsThe resulting best performing models had an accuracy score of ~80% for predicting type of bread, ~60% for butter, and ~ 28% for filling type. We see that the main struggle in predicting the fillings lies with the spreadable fillings, meaning the model may be focusing on structural aspects and not nutritional composition.DiscussionFurther analysis on non-homogeneous mixed food items, using computer vision techniques, will contribute toward a generalizable system. While there are still significant technical challenges to overcome before such a system can be routinely implemented in studies of free-living subjects, we believe it holds promise as a future tool for nutrition research and population intake monitoring.

Synthesizing Local Capacities, Multi-Source Remote Sensing and Meta-Learning to Optimize Forest Carbon Assessment in Data-Poor Regions

As the climate emergency escalates, the role of forests in carbon sequestration is paramount. This paper proposes a framework that integrates local capacities, multi-source remote sensing data, and meta-learning to enhance forest carbon assessment methodologies in data-scarce regions. By integrating multi-source optical and radar remote sensing data alongside community forest inventories, we applied a meta-modelling approach using stacked generalization ensemble to estimate forest above-ground carbon (AGC). We also conducted a Kruskal–Wallis test to determine significant differences in AGC among different tree species. The Kruskal–Wallis test (p = 1.37 × 10−13) and Dunn post-hoc analysis revealed significant differences in carbon stock potential among tree species, with Afzelia quanzensis (x~ = 12 kg/ha, P-holm-adj. = 0.05) and the locally known species M’buta (x~ = 6 kg/ha, P-holm-adj. = 5.45 × 10−9) exhibiting a significantly higher median AGC. Our results further showed that combining optical and radar remote sensing data substantially improved prediction accuracy compared to single-source remote sensing data. To improve forest carbon assessment, we employed stacked generalization, combining multiple machine learning algorithms to leverage their complementary strengths and address individual limitations. This ensemble approach yielded more robust estimates than conventional methods. Notably, a stacking ensemble of support vector machines and random forest achieved the highest accuracy (R2 = 0.84, RMSE = 1.36), followed by an ensemble of all base learners (R2 = 0.83, RMSE = 1.39). Additionally, our results demonstrate that factors such as the diversity of base learners and the sensitivity of meta-leaners to optimization can influence stacking performance.

Identifying an Appropriate Extractant for Assessing the Presence of Available Potassium in Soils with High Mica Content: A Comprehensive Statistical Analysis

ABSTRACT The choice of extractant for estimating plant-available potassium (K) in mica-rich soil has not been universally established. In order to evaluate and monitor the complex chemical performance of K in K-rich soil, and subsequently, its availability in crops, a suitable extractant protocol that takes into consideration soil properties in relation to rice plant parts (root, shoot, and grain) uptake becomes necessary. Nine laboratory soil K extraction methods to estimate the K content using 63 soil samples collected from nearly 21 mica mines, encompassing a wide range of properties across different agricultural lands. While extraction using ammonium acetate is the widely used for soil K testing, other methods have also been proposed for estimating crop-available K. Among the various extractants tested, 0.5 M sodium bicarbonate at pH 8.5 (mean 132.24 mg/kg) gave the best results for estimating the K content in mica-rich soil from agricultural areas near mica mines. Additionally, the soil water-soluble K extractant showed promise for rice plant parts such as root, shoot, and grain K uptake, as indicated by a correlation coefficient of 0.94, 0.94, and 0.93 respectively. Based on the artificial neural network (ANN) and Sobol (variance-based method for determining how different input variables contribute to the output variance of a model) methods, water-soluble K and the Mehlich 3 extraction procedure showed the best results for suitable extractable forms of K and a positive impact on K uptake by rice grain. Taylor (a graphical tool used to compare various models) diagram used for comparing multiple machine learning algorithms, multivariate adaptive regression spline and extreme gradient boosting (XGBoost) emerged as the best-fit models for determining the contribution of each extractant to rice K uptake. Considering soils derived from weathered parent materials often have negative K balances, which could lead to negative of plant-available K.

Research on Advertising Prediction Based on Stacking Ensemble Algorithms

The purpose of advertising prediction is to analyze user behavior and features to forecast the effectiveness of ad placements, thereby achieving a higher return on investment. This paper employs a stacking-based ensemble learning method, integrating multiple machine learning algorithms to enhance predictive capabilities. To support this method, this article selected a dataset related to advertising click through rate (CTR). The paper constructs two models: Gradient Boosting Decision Tree (GBDT) and eXtreme Gradient Boosting (XGBoost). Then the paper utilizes Bayesian optimization to tune the model parameters to achieve optimal parameter combinations. The paper uses these two individual models as base learners in a stacking ensemble model, with a random forest model serving as the meta-learner for training. This paper evaluates the predictive performance of the models using two metrics: log loss value, and AUC value. After experimental evaluation, the AUC value of the Stack fusion model reached 0.9999, which is about 0.42%, 1.29%, and 10.99% higher than XGBoost (0.9956), Random Forest (0.9871), and GBDT (0.8904), respectively. In terms of logloss value, the Stack fusion model has a loss of 0.0101, significantly lower than XGBoost's 0.0595, a decrease of about 83.01%. Compared to the random forest's 0.1896 and GBDT's 0.1884, it decreased by approximately 94.67% and 94.64%, respectively. This result indicates that the ensemble model proposed in this paper demonstrates a significant improvement in click-through rate prediction when contrasted with individual models.

Analysis and Prediction of Grouting Reinforcement Performance of Broken Rock Considering Joint Morphology Characteristics

In tunnel engineering, joint shear slip caused by external disturbances is a key factor contributing to landslides, instability of surrounding rock masses, and related hazards. Therefore, accurately characterizing the macromechanical properties of joints is essential for ensuring engineering safety. Given the significant influence of rock joint morphology on mechanical behavior, this study employs the frequency spectrum fractal dimension (D) and the frequency domain amplitude integral (Rq) as quantitative descriptors of joint morphology. Using Fourier transform techniques, a reconstruction method is developed to model joints with arbitrary shape characteristics. The numerical model is calibrated through 3D printing and direct shear tests. Systematic parameter analysis validates the selected quantitative indices as effective descriptors of joint morphology. Furthermore, multiple machine learning algorithms are employed to construct a robust predictive model. Machine learning, recognized as a rapidly advancing field, plays a pivotal role in data-driven engineering applications due to its powerful analytical capabilities. In this study, six algorithms—Random Forest (RF), Support Vector Regression (SVR), BP Neural Network, GA-BP Neural Network, Genetic Programming (GP), and ANN-based MCD—are evaluated using 300 samples. The performance of each algorithm is assessed through comparative analysis of their predictive accuracy based on correlation coefficients. The results demonstrate that all six algorithms achieve satisfactory predictive performance. Notably, the Random Forest (RF) algorithm excels in rapid and accurate predictions when handling similar training data, while the ANN-based MCD algorithm consistently delivers stable and precise results across diverse datasets.

La floricultura como cultivo alternativo: Análisis descriptivo, modelación con inteligencia artificial, análisis de escenarios y análisis económico

Floriculture is a sector of growing global importance, contributing to employment generation, income creation, and the promotion of biodiversity and sustainability. This study aimed to identify the factors influencing the adoption of floriculture as an alternative crop in the province of Leoncio Prado, Peru, and to assess its economic viability. A total of 269 farmers were surveyed, analyzing attitudes, land suitability, and socioeconomic and environmental factors. Influential factors were identified using descriptive analysis, chi-square tests, and logistic regression (p &lt; 0.1). Additionally, multiple machine learning algorithms (Decision Trees, Logistic Regression, KNN, SVM, Ensemble, Neural Networks, Naive Bayes) with cross-validation (k = 5) and AUC metrics were employed to model adoption intentions. Scenarios were developed to increase the willingness to adopt floriculture, and an economic analysis of eight tropical species (Red Ginger, Anthurium, Emperor's Staff, Heliconia, Gardenia, Parrot's Beak, Golden Heliconias, Maracas) was conducted. The results reveal that willingness to change crops, participation in awareness campaigns, allocation of areas for conservation, and cost control are key factors. The neural network model achieved an AUC of 83.3%, and improved scenarios indicate that adoption could increase by up to 11.32%. Red Ginger demonstrated high profitability (NPV S/10428; IRR 51%; PBP 0.7 years). In conclusion, floriculture represents an economically and environmentally viable alternative that contributes to agricultural diversification and sustainability.

Construction of an anaplastic thyroid cancer stratification signature to guide immune therapy selection and validation of the pivotal gene HLF through in vitro experiments.

While most thyroid cancer patients have a favorable prognosis, anaplastic thyroid carcinoma (ATC) remains a particularly aggressive form with a median survival time of just five months. Conventional therapies offer limited benefits for this type of thyroid cancer. Our study aims to identify ATC patients who might bene t from immunotherapy. Our study uses multiple algorithms by R4.2.0, and gene expression and clinical data are collected from TCGA, GEO and local cohort. In vitro experiments, such as western blot and immunofluorescence staining, are performed. Using a set of five genes uniquely expressed across various types of thyroid cancer, we developed a machine-learning model to distinguish each type within the GEO dataset of thyroid cancer patients (GSE60542, GSE76039, GSE33630, GSE53157, GSE65144, GSE29265, GSE82208, GSE27155, GSE58545, GSE54958, and GSE32662). These genes allowed us to stratify ATC into three distinct groups, each exhibiting significantly different responses to anti-PD1 therapy as determined by consensus clustering. Through weighted gene co-expression network analysis (WGCNA), we identified 12 differentially expressed genes closely associated with immunotherapy outcomes. This led to the creation of a refined signature for predicting ATC's immune responsiveness to anti-PD1 therapy, which was further validated using thyroid cancer cohorts from TCGA and nine melanoma cohorts from clinical trials. Among the 12 genes, HLF stood out due to its strong association with various cancer hallmarks. Our study revealed that HLF impedes ATC progression by down-regulating the epithelial-to-mesenchymal transition (EMT) pathway, reducing T cell exhaustion, and increasing sensitivity to sorafenib, as demonstrated through our in-vitro experiments.

Secrecy-Constrained UAV-Mounted RIS-Assisted ISAC Networks: Position Optimization and Power Beamforming

This paper investigates secrecy solutions for integrated sensing and communication (ISAC) systems, leveraging the combination of a reflecting intelligent surface (RIS) and an unmanned aerial vehicle (UAV) to introduce new degrees of freedom for enhanced system performance. Specifically, we propose a secure ISAC system supported by a UAV-mounted RIS, where an ISAC base station (BS) facilitates secure multi-user communication while simultaneously detecting potentially malicious radar targets. Our goal is to improve parameter estimation performance, measured by the Cramér–Rao bound (CRB), by jointly optimizing the UAV position, transmit beamforming, and RIS beamforming, subject to constraints including the UAV flight area, communication users’ quality of service (QoS) requirements, secure transmission demands, power budget, and RIS reflecting coefficient limits. To address this non-convex, multivariate, and coupled problem, we decompose it into three subproblems, which are solved iteratively using particle swarm optimization (PSO), semi-definite relaxation (SDR), majorization–minimization (MM), and alternating direction method of multipliers (ADMM) algorithms. Our numerical results validate the effectiveness of the proposed scheme and demonstrate the potential of employing UAV-mounted RIS in ISAC systems to enhance radar sensing capabilities.