Elastic Net Model Research Articles

Predicting Prefecture-Level Well-Being Indicators in Japan Using Search Volumes in Internet Search Engines: Infodemiology Study.

In recent years, the adoption of well-being indicators by national governments and international organizations has emerged as an important tool for evaluating state governance and societal progress. Traditionally, well-being has been gauged primarily through economic metrics such as gross domestic product, which fall short of capturing multifaceted well-being, including socioeconomic inequalities, life satisfaction, and health status. Current well-being indicators, including both subjective and objective measures, offer a broader evaluation but face challenges such as high survey costs and difficulties in evaluating at regional levels within countries. The emergence of web log data as an alternative source of well-being indicators offers the potential for more cost-effective, timely, and less biased assessments. This study aimed to develop a model using internet search data to predict well-being indicators at the regional level in Japan, providing policy makers with a more accessible and cost-effective tool for assessing public well-being and making informed decisions. This study used the Regional Well-Being Index (RWI) for Japan, which evaluates prefectural well-being across 47 prefectures for the years 2010, 2013, 2016, and 2019, as the outcome variable. The RWI includes a comprehensive approach integrating both subjective and objective indicators across 11 domains, including income, job, and life satisfaction. Predictor variables included z score-normalized relative search volume (RSV) data from Google Trends for words relevant to each domain. Unrelated words were excluded from the analysis to ensure relevance. The Elastic Net methodology was applied to predict RWI using RSVs, with α balancing ridge and lasso effects and λ regulating their strengths. The model was optimized by cross-validation, determining the best mix and strength of regularization parameters to minimize prediction error. Root mean square errors (RMSE) and coefficients of determination (R2) were used to assess the model's predictive accuracy and fit. An analysis of Google Trends data yielded 275 words related to the RWI domains, and RSVs were collected for 211 words after filtering out irrelevant terms. The mean search frequencies for these words during 2010, 2013, 2016, and 2019 ranged from -1.587 to 3.902, with SDs between 3.025 and 0.053. The best Elastic Net model (α=0.1, λ=0.906, RMSE=1.290, and R2=0.904) was built using 2010-2016 training data and 2-13 variables per domain. Applied to 2019 test data, it yielded an RMSE of 2.328 and R2 of 0.665. This study demonstrates the effectiveness of using internet search log data through the Elastic Net machine learning method to predict the RWI in Japanese prefectures with high accuracy, offering a rapid and cost-efficient alternative to traditional survey approaches. This study highlights the potential of this methodology to provide foundational data for evidence-based policy making aimed at enhancing well-being.

Evaluating the effectiveness of elastic net model in predicting stock closing price

Abstract. Accurate stock price prediction is important for financial investors, as it directly influences investment strategies and decision-making. In this study, the author focuses on evaluating the effectiveness of the Elastic Net model in forecasting the closing prices of Apple Inc.'s stock. The dataset, spanning from 2014 to 2023, is sourced from Kaggle and includes both historical prices and a range of technical indicators, thereby ensuring the accuracy and reliability of the research data. To enhance the predictive accuracy of the model, this paper carefully selects the 14-day moving average of the closing prices along with a set of relevant technical indicators as the key input variables. This selection aims to capture the underlying trends and patterns in the stock's price. The predictive performance of the Elastic Net model is evaluated using multiple evaluation metrics, including Mean Squared Error (MSE) and R-squared values, which quantify the model's accuracy and explanatory power. Additionally, this paper utilizes the Autocorrelation Function (ACF) plot to analyze the residuals and further validate the model's effectiveness. The results demonstrate that the Elastic Net model is well-suited for handling high-dimensional financial data, exhibiting strong performance in predicting Apple Inc.'s stock prices. However, it also underscores the model's potential problem as a valuable tool for investors seeking to make proper decisions in real-life conditions.

Social Connection and Suicidal Thoughts and Behaviors in the Million Veteran Program Cohort

People with lower levels of social connection are at increased risk for suicidal thoughts and behaviors. This extends to populations at greater risk of death by suicide, including U.S. military veterans. Despite this well-established association, it is unclear which measures of social connection are most useful in identifying veterans who could benefit from intervention to prevent suicide. To this end, we used data from the Million Veteran Program (MVP) to investigate the measures of social connection most strongly associated with suicidal thoughts and behaviors. Our sample included 264,626 veterans who reported on three measures of social connection—marital status, household size, and perceived social support—and were assessed for suicidal thoughts and behaviors using electronic health records. Veterans who were partnered (OR = 0.78, 95% CI [0.76-0.80], p < .001), living with others (OR = 0.71, 95% CI [0.70-0.73], p < .001), or reported higher levels of social support (OR = 0.67, 95% CI [0.65-0.69], p < .001), were less likely to have suicidal thoughts and behaviors. These associations varied by veterans’ age, sex, and era of military service. When combined into a single risk score, lower levels of social connection were associated with greater likelihood of suicidal thoughts and behaviors (β = 1.42, 95% CI [1.40-1.43], p < .001). Social support, particularly positive social interactions, showed the strongest associations with suicidal thoughts and behaviors in elastic net regression models. Common measures of social connection, particularly social support, could be useful in assessing suicide risk and treatment needs for veterans.

Diagnostic MRI Score to Differentiate Susac Syndrome from Primary Angiitis of the Central Nervous System and Multiple Sclerosis.

Susac syndrome (SuS), multiple sclerosis (MS), and primary angiitis of the central nervous system (PACNS) present diagnostic challenges due to overlapping clinical features. We aimed to enhance diagnostic precision by developing the SPAMS (SuS, PACNS, MS) score, a practical radiological tool. This multicenter study included 99 patients (43 SuS, 37 MS, 19 PACNS) from South American countries. Relevant MRI features were identified through an elastic-net model determined key variables. The SPAMS score assigned 2 points for snowball lesions, 1 point for spokes-like lesions, or if there are more than 4 lesions in the corpus callosum, corpus callosum involvement, or cerebellar involvement. It subtracted 1 point if gadolinium-enhancing lesions or 4 points if Dawson's fingers are present. Bootstrapping validated the optimal cutoff at 2 points, exhibiting a diagnostic performance of area under the curve = 0.931, sensitivity = 88%, specificity = 89%, positive predictive value = 88%, negative predictive value = 89%, and accuracy = 88%. When specific MRI findings coexisted, the SPAMS score differentiated SuS from MS and PACNS. Access to MRI and standard protocol sequences makes it a valuable tool for timely diagnosis and treatment, potentially preventing disability progression and severe clinical outcomes. ANN NEUROL 2024;96:846-854.

Metabolic Transcriptional Activation in Ulcerative Colitis Identified Through scRNA-seq Analysis

Background: Ulcerative colitis is a chronic inflammatory disease affecting the colon. During chronic inflammation of epithelial cells, lipid metabolism via pro-inflammatory eicosanoids is known to modify the immune response. Methods: Starting from the Mammalian Metabolic Database, the expression of metabolic enzymes was investigated in two independent cohorts from transcriptome datasets GSE38713 and GSE11223, which analyzed ulcerative colitis tissue samples from the digestive tract. Results: In the first cohort, 145 differentially expressed enzymes were identified as significantly regulated between ulcerative colitis tissues and normal controls. Overexpressed enzymes were selected to tune an Elastic Net model in the second cohort. Using the best parameters, the model achieved a prediction accuracy for ulcerative colitis with an area under the curve (AUC) of 0.79. Twenty-two metabolic enzymes were found to be commonly overexpressed in both independent cohorts, with decreasing Elastic Net predictive coefficients as follows: LIPG (3.98), PSAT1 (3.69), PGM3 (2.74), CD38 (2.28), BLVRA (1.99), CBR3 (1.94), NT5DC2 (1.76), PHGDH (1.71), GPX7 (1.58), CASP1 (1.56), ASRGL1 (1.4), SOD3 (1.25), CHST2 (0.965), CHST11 (0.95), KYNU (0.94), PLAG2G7 (0.92), SRM (0.87), PTGS2 (0.80), LPIN1 (0.47), ME1 (0.31), PTGDS (0.14), and ADA (0.13). Functional enrichment analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database highlighted the main implications of these enzymes in cysteine and methionine metabolism (adjusted p-value = 0.01), arachidonic acid and prostaglandin metabolism (adjusted p-value = 0.01), and carbon metabolism (adjusted p-value = 0.04). A metabolic score based on the transcriptional activation of the validated twenty-two enzymes was found to be significantly greater in Ulcerative colitis samples compared to healthy donor samples (p-value = 1.52 × 10−8). Conclusions: A metabolic expression score was established and reflects the implications of heterogeneous metabolic pathway deregulations in the digestive tract of patients with ulcerative colitis.

Using Four Machine Learning Methods to Analyze the Association Between Polycyclic Aromatic Hydrocarbons and Visual Impairment in American Adults: Evidence from NHANES

The causes of visual impairment are complex and may be influenced by exposure to environmental pollutants. Using data from the 2003–2004 National Health and Nutrition Examination Survey (NHANES), we examined the association between exposure to ten polycyclic aromatic hydrocarbons (PAHs) and vision problems in 1149 U.S. adults. We employed various supervised learning methods, including variable selection techniques such as Lasso and elastic net, weighted quantile sum regression (WQS), and Bayesian kernel machine regression (BKMR), to assess the association between PAHs and the occurrence of visual impairments. The mediation effects between urinary 2-fluorene and inflammation were evaluated using mediation analysis. Both the lasso and elastic net models consistently identified two specific PAH congeners, 2-fluorene and 1-phenanthrene, as significant predictors. The WQS regression revealed a positive relationship between the PAH mixture and visual impairment, with notable contributions from urinary 2-fluorene (weight = 0.39) and 9-fluorene (weight = 0.21). BKMR analysis indicated that the likelihood of visual impairment increases with higher PAH exposure, showing a general upward trend. This trend also revealed a positive association between visual impairment and exposure to four specific PAH metabolites, including 2-fluorene. A significant mediation effect was observed for alkaline phosphatase (p = 0.03), with a proportion mediated of 10.48%. Our findings suggest a significant association between PAHs and visual impairment, with multiple statistical models consistently emphasizing the crucial role of 2-fluorene exposure. This study highlights the importance of considering environmental pollutants as significant contributors to visual health outcomes, providing insights for preventing visual impairment.

Using machine learning methods to identify trajectories of change and predict responders and non-responders to short-term dynamic therapy

Objectives Predicting therapy responders can significantly improve clinical outcomes. This study aims to identify predictors of response to short-term dynamic therapy. Methods Data from 95 patients who underwent 16-session therapy were analyzed using machine learning. Weekly progress was monitored with the Outcome Questionnaire (OQ45) and Target Complaints (TC). A machine learning model identified change trajectories for responders and non-responders, with a random forest algorithm and elastic net modeling predicting trajectory group membership using pre-treatment data. Results A weak positive relationship was found between the trajectories of the two outcome variables. The results of the different analysis methods were compared and discussed. Important predictors of OQ45 trajectories, based on random forest modeling, included initial symptom severity, difficulties in emotion regulation, coldness, avoidant attachment, conscientiousness, interpersonal problems, non-acceptance of negative emotion, neuroticism, emotional clarity, impulsivity, and emotion awareness (72.8% accuracy). Initial problem severity, self-scarifying extraversion, and non-assertiveness were the most dominant predictors for TC trajectories (62.8% accuracy). Conclusions These findings offer data-driven insights for selecting short-term dynamic therapy. Predicting response for the OQ45, a nomothetic measure, does not extend to the TC, an idiographic measure, and vice versa, highlighting the importance of multidimensional outcome evaluations for personalized treatment.

Priority-Elastic net for binary disease outcome prediction based on multi-omics data

BackgroundHigh-dimensional omics data integration has emerged as a prominent avenue within the healthcare industry, presenting substantial potential to improve predictive models. However, the data integration process faces several challenges, including data heterogeneity, priority sequence in which data blocks are prioritized for rendering predictive information contained in multiple blocks, assessing the flow of information from one omics level to the other and multicollinearity.MethodsWe propose the Priority-Elastic net algorithm, a hierarchical regression method extending Priority-Lasso for the binary logistic regression model by incorporating a priority order for blocks of variables while fitting Elastic-net models sequentially for each block. The fitted values from each step are then used as an offset in the subsequent step. Additionally, we considered the adaptive elastic-net penalty within our priority framework to compare the results.ResultsThe Priority-Elastic net and Priority-Adaptive Elastic net algorithms were evaluated on a brain tumor dataset available from The Cancer Genome Atlas (TCGA), accounting for transcriptomics, proteomics, and clinical information measured over two glioma types: Lower-grade glioma (LGG) and glioblastoma (GBM).ConclusionOur findings suggest that the Priority-Elastic net is a highly advantageous choice for a wide range of applications. It offers moderate computational complexity, flexibility in integrating prior knowledge while introducing a hierarchical modeling perspective, and, importantly, improved stability and accuracy in predictions, making it superior to the other methods discussed. This evolution marks a significant step forward in predictive modeling, offering a sophisticated tool for navigating the complexities of multi-omics datasets in pursuit of precision medicine’s ultimate goal: personalized treatment optimization based on a comprehensive array of patient-specific data. This framework can be generalized to time-to-event, Cox proportional hazards regression and multicategorical outcomes. A practical implementation of this method is available upon request in R script, complete with an example to facilitate its application.

Serum metabolomics improve risk stratification for incident heart failure

Abstract Background The prediction and early detection of heart failure (HF) is vital to reduce its substantial impact on quality of life, survival and healthcare expenditure. Current incident HF risk stratification strategies achieve only modest performance. Moreover, state-of-the-art risk scores focus on commonly acknowledged clinical risk factors, thus necessitating the integration of heterogenous data sources and prioritizing individuals with obvious risk constellations. Purpose Here, we explored the predictive value of serum metabolomics (168 metabolites detected by proton nuclear magnetic resonance (1H-NMR) spectroscopy) for incident HF. Methods Leveraging data of n = 68,311 individuals and &amp;gt; 0.8 million person-years of follow-up from the UK Biobank (UKB) cohort, we (I) evaluated the association of individual metabolites with incident HF via fitting of per-metabolite COX proportional hazards models and (II) trained and validated elastic net (EN) models to predict incident HF using the serum metabolome. Discriminative performance was benchmarked against a comprehensive, well-validated clinical risk score (Pooled Cohort Equations to Prevent HF, PCP-HF). External validation was conducted in the independent FINRISK study. Results Several metabolites were independently associated with incident HF (90/168 adjusting for age and sex, 48/168 adjusting for PCP-HF). Performance-optimized risk models effectively retained key predictors representing highly correlated clusters (≈ 80 % feature reduction). The addition of metabolomics to PCP-HF improved predictive performance (Harrel’s C: 0.768 vs. 0.755, ΔC = 0.013 (95% confidence interval (CI) 0.004 – 0.022), continuous net reclassification improvement (NRI) = 0.287 (95% CI 0.200 – 0.367), relative integrated discrimination improvement (IDI): 17.47 % (95% CI 9.463 - 27.825)). Simplified models only including age, sex and metabolomics performed almost as well as the PCP-HF model (Harrel’s C: 0.745 vs. 0.755, ΔC = 0.010 (95% CI -0.004 - 0.027), continuous NRI: 0.097 (95% CI -0.025 - 0.217), relative IDI: 13.445 % (95% CI -10.608 - 41.454)). Risk and survival stratification was improved by integrating metabolomics. External validation within FINRISK revealed similar patterns using the original model coefficients. Conclusions Serum metabolomics improve the prediction of incident HF risk. Scores obtained from the combination of age, sex and metabolomics exhibit similar predictive power as clinical risk models. Offering serum metabolomics to a middle-aged cohort might significantly improve HF risk stratification, thus facilitating preventive efforts. Via a single blood draw, serum metabolomics displays a highly cost- and time-effective, standardizable, and scalable alternative to clinical risk scores involving physical measurements and history taking in addition to clinical chemistry.Relative performance (test split)ROC &amp; DCA plots (test split)

Remote Sensing–Based UAV Imaging in Heat Pattern Analysis Impact on Climate Change Detection Using Fuzzy Stacked Lasso Elastic-Net Model

Remote Sensing–Based UAV Imaging in Heat Pattern Analysis Impact on Climate Change Detection Using Fuzzy Stacked Lasso Elastic-Net Model

Archetypal Analysis of Kidney Allograft Biopsies Using Next-generation Sequencing Technology.

In kidney transplantation, molecular diagnostics may be a valuable approach to improve the precision of the diagnosis. Using next-generation sequencing (NGS), we aimed to identify clinically relevant archetypes. We conducted an Illumina bulk RNA sequencing on 770 kidney biopsies (540 kidney recipients) collected between 2006 and 2021 from 11 European centers. Differentially expressed genes were determined for 11 Banff lesions. An ElasticNet model was used for feature selection, and 4 machine learning classifiers were trained to predict the probability of presence of the lesions. NGS-based classifiers were used in an unsupervised archetypal analysis to different archetypes. The association of the archetypes with allograft survival was assessed using the iBox risk prediction score. The ElasticNet feature selection reduced the number of the genes from a range of 859-10 830 to a range of 52-867 genes. NGS-based classifiers demonstrated robust performances (precision-recall area under the curves 0.708-0.980) in predicting the Banff lesions. Archetypal analysis revealed 8 distinct phenotypes, each characterized by distinct clinical, immunological, and histological features. Although the archetypes confirmed the well-defined Banff rejection phenotypes for T cell-mediated rejection and antibody-mediated rejection, equivocal histologic antibody-mediated rejection, and borderline diagnoses were reclassified into different archetypes based on their molecular signatures. The 8 NGS-based archetypes displayed distinct allograft survival profiles with incremental graft loss rates between archetypes, ranging from 90% to 56% rates 7 y after evaluation (P < 0.0001). Using molecular phenotyping, 8 archetypes were identified. These NGS-based archetypes might improve disease characterization, reclassify ambiguous Banff diagnoses, and enable patient-specific risk stratification.

Proteomic Analysis Uncovers Multi-Protein Signatures Associated with Early Diabetic Kidney Disease in Youth with Type 2 Diabetes Mellitus.

The onset of diabetic kidney disease (DKD) in youth with type 2 diabetes mellitus often occurs early, leading to complications in young adulthood. Risk biomarkers associated with the early onset of DKD are urgently needed in youth with type 2 diabetes. We conducted an in-depth analysis of 6596 proteins (SomaScan 7K) in 374 baseline plasma samples from the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) study to identify multi-protein signatures associated with the onset of albuminuria (urine albumin-to-creatinine ratio [UACR] ≥30 mg/g), a rapid decline in estimated glomerular filtration rate (eGFR) [annual eGFR decline >3 mL/min/1.73m2 and/or ≥3.3% at two consecutive visits], and hyperfiltration (≥135 mL/min/1.73m2 at two consecutive visits). Elastic net Cox regression with 10-fold cross-validation was applied to the top 100 proteins (ranked by p-value) to identify multi-protein signatures of time to development of DKD outcomes. Participants in the TODAY study (14±2 years old, 63% female, 7±6 months diabetes duration) experienced high rates of early DKD: 43% developed albuminuria, 48% hyperfiltration, and 16% rapid eGFR decline. Increased levels of seven and three proteins were predictive of shorter time to develop albuminuria and rapid eGFR decline, respectively; 118 proteins predicted time to development of hyperfiltration. Elastic net Cox proportional hazards model identified multi-protein signatures of time to incident early DKD with concordance for models with clinical covariates and selected proteins between 0.81 and 0.96, while the concordance for models with clinical covariates only was between 0.56 and 0.63. Our research sheds new light on proteomic changes early in the course of youth-onset type 2 diabetes that associate with DKD. Proteomic analyses identified promising risk factors that predict DKD risk in youth with type 2 diabetes and could deepen our understanding of DKD mechanisms and potential interventions.

Characterizing metabolomic signatures related to coffee and tea consumption and their association with incidence and dynamic progression of type 2 diabetes: A multi-state analysis.

Our study aimed to investigate the impact of tea and coffee consumption and related metabolomic signatures on dynamic transitions from diabetes-free status to incident type 2 diabetes (T2D), and subsequently to T2D-related complications and death. We included 438,970 participants in the UK Biobank who were free of diabetes and diabetes complications at baseline. Of these, 212,146 individuals had information on all metabolic biomarkers. We identified tea- and coffee-related metabolomic signatures using elastic net regression models. We examined associations of tea and coffee intake and related metabolomic signatures with the onset and progression of T2D using multi-state regression models. We observed that tea and coffee consumption and related metabolomic signatures were inversely associated with the risk of five T2D transitions. For example, HRs (95% CIs) per SD increase of the tea-related metabolomic signature were 0.87 (0.85, 0.89), 0.97 (0.95, 0.99), 0.91 (0.90, 0.92), 0.92 (0.91, 0.94), and 0.91 (0.90, 0.92) for transitions from diabetes-free state to incident T2D, from diabetes-free state to total death, from incident T2D to T2D complications, from incident T2D to death, and from T2D complications to death. These findings highlight the benefit of tea and coffee intake in reducing the risk of occurrence and progression of T2D.

Performance of the Classical Model in Feature Selection Across Varying Database Sizes of Healthcare Data

Machine learning is increasingly being applied to medical research, particularly in selecting predictive modelling variables. By identifying relevant variables, researchers can improve model accuracy and reliability, leading to better clinical decisions and reduced overfitting. Efficient utilization of resources and the validity of medical research findings depend on selecting the right variables. However, few studies compare the performance of classical and modern methods for selecting characteristics in health datasets, highlighting the need for a critical evaluation to choose the most suitable approach. We analysed the performance of six different variable selection methods, which includes stepwise, forward and backward selection using p-value and AIC, LASSO, and Elastic Net. Health-related surveillance data on behaviors, health status, and medical service usage were used across ten databases, with sizes ranging from 10% to 100%, maintaining consistent outcome proportions. Varying database sizes were utilized to assess their impact on prediction models, as they can significantly influence accuracy, overfitting, generalizability, statistical power, parameter estimation reliability, computational complexity, and variable selection. The stepwise and backward AIC model showed the highest accuracy with an Area under the ROC Curve (AUC) of 0.889. Despite its sparsity, the Lasso and Elastic Net model also performed well. The study also found that binary variables were considered more crucial by the Lasso and Elastic Net model. Importantly, the significance of variables remained consistent across different database sizes. The study shows that no major variations in results between the fitness metric of the model and the number of variables in stepwise and backward p-value models, irrespective of the database's size. LASSO and Elastic Net models surpassed other models throughout various database sizes, and with fewer variables.

Prediction of Temperature Variability on Power Transmission Line Parameters using Intelligent Approaches

Due to changes in meteorological factors, the instability in the power at the end of the transmission system demands considerable attention. The temperature of the transmission line varies, which has a significant impact on the line parameters. An accurate prediction of line parameters behaviour is necessary to ensure system reliability. The present study is a step towards predicting variations in line parameters with respect to temperature variation. In addition, power loss and voltage drop due to variations in resistance are also predicted. Support Vector Machine (SVM) and ElasticNet, a machine learning algorithm, predict line parameters such as resistance, inductance, capacitance, voltage drop, and power losses. Furthermore, different seasons-based SVM and ElasticNet models for these parameters are considered. Seasons-based models are divided into two types, namely, summer and winter. 220-Kilovolt transmission data and weather information are used as model inputs. Predicted results of transmission line parameters are described in the form of RMSE and MRE. Moreover, the performance results of SVM and ElasticNet are also compared to show better prediction results. The result shows that the minimum prediction error of line parameters are 0.0511, 0.301, 0.426, 0.913, and 0.1501 in RMSE and 4.212, 0.518, 2.888, 0.097, and 0.615 percentages in MRE. This research work may provide technical guidance to transmission line engineers on enhancing the performance of transmission systems.

Machine learning and Bayesian network analyses identifies associations with insomnia in a national sample of 31,285 treatment-seeking college students

BackgroundA better understanding of the relationships between insomnia and anxiety, mood, eating, and alcohol-use disorders is needed given its prevalence among young adults. Supervised machine learning provides the ability to evaluate which mental disorder is most associated with heightened insomnia among U.S. college students. Combined with Bayesian network analysis, probable directional relationships between insomnia and interacting symptoms may be illuminated.MethodsThe current exploratory analyses utilized a national sample of college students across 26 U.S. colleges and universities collected during population-level screening before entering a randomized controlled trial. We used a 4-step statistical approach: (1) at the disorder level, an elastic net regularization model examined the relative importance of the association between insomnia and 7 mental disorders (major depressive disorder, generalized anxiety disorder, social anxiety disorder, panic disorder, post-traumatic stress disorder, anorexia nervosa, and alcohol use disorder); (2) This model was evaluated within a hold-out sample. (3) at the symptom level, a completed partially directed acyclic graph (CPDAG) was computed via a Bayesian hill-climbing algorithm to estimate potential directionality among insomnia and its most associated disorder [based on SHAP (SHapley Additive exPlanations) values)]; (4) the CPDAG was then tested for generalizability by assessing (in)equality within a hold-out sample using structural hamming distance (SHD).ResultsOf 31,285 participants, 20,597 were women (65.8%); mean (standard deviation) age was 22.96 (4.52) years. The elastic net model demonstrated clinical significance in predicting insomnia severity in the training sample [R2 = .44 (.01); RMSE = 5.00 (0.08)], with comparable performance in the hold-out sample (R2 = .33; RMSE = 5.47). SHAP values indicated that the presence of any mental disorder was associated with higher insomnia scores, with major depressive disorder as the most important disorder associated with heightened insomnia (mean |SHAP|= 3.18). The training CPDAG and hold-out CPDAG (SHD = 7) suggested depression symptoms presupposed insomnia with depressed mood, fatigue, and self-esteem as key parent nodes.ConclusionThese findings provide insights into the associations between insomnia and mental disorders among college students and warrant further investigation into the potential direction of causality between insomnia and depression.Trial registrationTrial was registered on the National Institute of Health RePORTER website (R01MH115128 || 23/08/2018).

Whole-blood RNA biomarkers for predicting survival in non-human primates following thoracic radiation

Radiation injury, either from radiotherapy or a mass-casualty event requires a health care system that can efficiently allocate resources to patients. We conducted a comprehensive transcriptome analysis of whole blood from a nonhuman primate model that received upper thoracic radiation (9.8–10.7 Gy). Blood samples were collected at multiple time points, extending up to 270 days post-irradiation with a minimum n = 6 for initial time points (Day 3-Day 40) and a total number of n = 28 primates. No males receiving the higher dose survived to Day 270. Using the Elastic Net model in R we found that pooling biomarkers from Day 3–21 increased our accuracy in discerning survival time, pleural effusion or dose compared to using biomarkers specific to a single day. For survival data, in predicting short term (less than 90 day), medium term (Day 91–269) or long-term survival (Day 270), prediction accuracy using only Day 3 data was 0.14 (95% Confidence Interval (CI) 0.1, 0.19) while pooled data for Male and Female was 0.76 (CI 0.69, 0.82). When pooled data was divided by biological sex, accuracy was 0.7 (CI 0.58, 0.8) for pooled data from Males and 0.84 (CI 0.76, 0.91) for Females. The development of RNA biomarkers as a tool to aid in clinical decision-making could significantly improve patient care in cases of radiation injury, whether from radiotherapy or mass-casualty events. Further validation and clinical translation of these findings could lead to improved patient care and management strategies in cases of radiation exposure.

Predicting Methanol Space-Time Yield from CO? Hydrogenation Using Machine Learning: Statistical Evaluation of Penalized Regression Techniques

This study investigates the effectiveness of machine learning techniques, specifically penalized regression models Ridge Regression, Lasso Regression, and Elastic Net Regression in predicting methanol space-time yield (STY) from CO? hydrogenation data. Using a dataset derived from Cu-based catalyst research, the study implemented a comprehensive preprocessing approach, including data cleaning, imputation, outlier removal, and normalization. The models were rigorously evaluated through 10-fold cross-validation and tested on unseen data. Ridge Regression outperformed the other models, achieving the lowest Root Mean Squared Error (RMSE) of 0.7706, Mean Absolute Error (MAE) of 0.5627, and Mean Squared Error (MSE) of 0.5938. In comparison, Lasso and Elastic Net Regression models exhibited higher error metrics. Feature importance analysis revealed that Gas Hourly Space Velocity (GHSV) and Molar Masses of Support significantly influence catalytic activity. These findings suggest that Ridge Regression is a promising tool for accurately predicting methanol production, providing valuable insights for optimizing catalytic processes and advancing sustainable practices in chemical engineering.

Data-driven nexus between malaria incidence and World Bank indicators in the Mekong River during 2000-2022.

The increase in hydro dams in the Mekong River amidst the prevalence of multidrug-resistant malaria in Cambodia has raised concerns about global public health. Political conflicts during Covid-19 pandemic led cross-border movements of malaria cases from Myanmar and caused health care burden in Thailand. While previous publications used climatic indicators for predicting mosquito-borne diseases, this research used globally recognizable World Bank indicators to find the most impactful indicators related with malaria and shed light on the predictability of mosquito-borne diseases. The World Bank datasets of the World Development Indicators and Climate Change Knowledge Portal contain 1494 time series indicators. They were stepwise screened by Pearson and Distance correlation. The sets of five and four contain respectively 19 and 149 indicators highly correlated with malaria incidence which were found similarly among five and four GMS countries. Living areas, ages, career, income, technology accessibility, infrastructural facilities, unclean fuel use, tobacco smoking, and health care deficiency have affected malaria incidence. Tonle Sap Lake, the largest freshwater lake in Southeast Asia, could contribute to the larval habitat. Seven groups of indicator topics containing 92 indicators with not-null datapoints were analyzed by regression models, including Multiple Linear, Ridge, Lasso, and Elastic Net models to choose 7 crucial features for malaria prediction via Long Short Time Memory network. The indicator of people using at least basic sanitation services and people practicing open defecation were health factors had most impacts on regression models. Malaria incidence could be predicted by one indicator to reach the optimal mean absolute error which was lower than 10 malaria cases (per 1,000 population at risk) in the Long Short Time Memory model. However, public health crises caused by political problems should be analyzed by political indexes for more precise predictions.

Lifestyle score is associated with cellular immune profiles in healthy Tanzanian adults

Immune system and vaccine responses vary across geographical locations worldwide, not only between high and low-middle income countries (LMICs), but also between rural and urban populations within the same country. Lifestyle factors such as housing conditions, exposure to microorganisms and parasites and diet are associated with rural-and urban-living. However, the relationships between these lifestyle factors and immune profiles have not been mapped in detail. Here, we profiled the immune system of 100 healthy Tanzanians living across four rural/urban areas using mass cytometry. We developed a lifestyle score based on an individual's household assets, housing condition and recent dietary history and studied the association with cellular immune profiles. Seventeen out of 80 immune cell clusters were associated with living location or lifestyle score, with eight identifiable only using lifestyle score. Individuals with low lifestyle score, most of whom live in rural settings, showed higher frequencies of NK cells, plasmablasts, atypical memory B cells, T helper 2 cells, regulatory T cells and activated CD4+ T effector memory cells expressing CD38, HLA-DR and CTLA-4. In contrast, those with high lifestyle score, most of whom live in urban areas, showed a less activated state of the immune system illustrated by higher frequencies of naïve CD8+ T cells. Using an elastic net machine learning model, we identified cellular immune signatures most associated with lifestyle score. Assuming a link between these immune profiles and vaccine responses, these signatures may inform us on the cellular mechanisms underlying poor responses to vaccines, but also reduced autoimmunity and allergies in low- and middle-income countries.