Lasso Regression Model Research Articles

A cuproptosis-related prognostic signature for guiding clinical diagnosis and treatment in uveal melanoma patients

BackgroundCuproptosis, one of the most recently discovered forms of cell death, is induced by the disruption of copper binding to the mitochondrial respiratory acylation components. However, the mechanism underlying cuproptosis in uveal melanoma (UM) has not yet been adequately studied. MethodsRNA and clinical data were obtained from The Cancer Genome Atlas (TCGA) database. Differentially expressed cuproptosis-related genes were identified by R software. A prognostic signature was constructed by applying LASSO regression and Cox regression models. The associations between the signature and the immune microenvironment, overall survival, and drug sensitivity were studied. In addition, qPCR and Western blotting were performed on UM cells and RPE cell lines to verify the expression levels of the genes encoding dihydrolipoamide dehydrogenase (DLD) and dihydrolipoamide S-succinyltransferase (DLST) in UM cases. ResultsUsing a cuproptosis-related prognostic signature, UM samples were classified into high- and low-risk groups. A significant difference in overall survival between the two risk groups was evident. Receiver operating characteristic curves demonstrated that the signature is a reliable predictor of prognosis. Immune cell infiltration, drug sensitivity, and immune checkpoint expression were analysed. Significant immune difference between the two high-risk groups was found, and the high expression of immune checkpoints in high-risk groups suggests significant immunotherapy potential. In addition, drug sensitivity analysis experiments suggest that erlotinib may be a potential treatment for high-risk patients. The results of in vitro experiments confirmed that DLD and DLST had higher expression levels in UM cell lines. ConclusionsThe prognostic signature developed in this study is a reliable biomarker for predicting the prognosis of UM and may serve as a tool for personalised treatment of patients with UM.

COVID-19 related myocardial injury is associated with immune dysregulation in symptomatic patients with cardiac MRI abnormalities.

While acute cardiovascular complications of COVID-19 are well-described, less is known about longer-term cardiac sequelae. For many individuals with cardiac signs or symptoms arising after COVID-19 infection, the aetiology remains unclear. We examined immune profiles associated with magnetic resonance imaging (MRI) abnormalities in patients with unexplained cardiac injury after COVID-19. Twenty-one participants (mean age 47 [SD 13] years, 71% female) with long COVID (n=17), raised troponin (n=2), or unexplained new-onset heart failure (n=2), who did not have pre-existing heart conditions or recent steroid/immunosuppression treatment were enrolled a mean 346 (SD 191) days after COVID-19 infection in a prospective observational study. Cardiac MRI and blood sampling for deep immunophenotyping using mass cytometry by time of flight and measurement of proteomic inflammatory markers was performed. Nine of 21 (43%) participants had MRI abnormalities (MRI(+)), including non-ischaemic patterns of late gadolinium enhancement and/or visually overt myocardial oedema in 8 people. One patient had mildly impaired biventricular function without fibrosis or oedema, and 2 had severe left ventricular impairment. MRI(+) individuals had higher blood CCL3, CCL7, FGF-23 and CD4 Th2 cells, and lower CD8 T effector memory (TEM) cells, than MRI(-). Cluster analysis revealed lower expression of inhibitory receptors PD1 and TIM3 in CD8 TEM cells from MRI(+) patients than MRI(-) patients, and functional studies of CD8 T αβ cells showed higher proportions of cytotoxic granzyme B+ secreting cells upon stimulation. CD8 TEM cells and CCL7 were the strongest predictors of MRI abnormalities in a LASSO regression model (composite AUC 0.96, 95%CI 0.88-1.0). CCL7 was correlated with diffuse myocardial fibrosis/oedema detected by quantitative T1 mapping (r=0.47, p=0.04). COVID-19 related cardiac injury in symptomatic patients with non-ischaemic myocarditis-like MRI abnormalities is associated with immune dysregulation, including decreased peripheral CD8 TEM cells and increased CCL7, persisting long after the initial infection.

Multi-modality risk prediction of cardiovascular diseases for breast cancer cohort in the All of Us Research Program.

This study leverages the rich diversity of the All of Us Research Program (All of Us)'s dataset to devise a predictive model for cardiovascular disease (CVD) in breast cancer (BC) survivors. Central to this endeavor is the creation of a robust data integration pipeline that synthesizes electronic health records (EHRs), patient surveys, and genomic data, while upholding fairness across demographic variables. We have developed a universal data wrangling pipeline to process and merge heterogeneous data sources of the All of Us dataset, address missingness and variance in data, and align disparate data modalities into a coherent framework for analysis. Utilizing a composite feature set including EHR, lifestyle, and social determinants of health (SDoH) data, we then employed Adaptive Lasso and Random Forest regression models to predict 6 CVD outcomes. The models were evaluated using the c-index and time-dependent Area Under the Receiver Operating Characteristic Curve over a 10-year period. The Adaptive Lasso model showed consistent performance across most CVD outcomes, while the Random Forest model excelled particularly in predicting outcomes like transient ischemic attack when incorporating the full multi-model feature set. Feature importance analysis revealed age and previous coronary events as dominant predictors across CVD outcomes, with SDoH clustering labels highlighting the nuanced impact of social factors. The development of both Cox-based predictive model and Random Forest Regression model represents the extensive application of the All of Us, in integrating EHR and patient surveys to enhance precision medicine. And the inclusion of SDoH clustering labels revealed the significant impact of sociobehavioral factors on patient outcomes, emphasizing the importance of comprehensive health determinants in predictive models. Despite these advancements, limitations include the exclusion of genetic data, broad categorization of CVD conditions, and the need for fairness analyses to ensure equitable model performance across diverse populations. Future work should refine clinical and social variable measurements, incorporate advanced imputation techniques, and explore additional predictive algorithms to enhance model precision and fairness. This study demonstrates the liability of the All of Us's diverse dataset in developing a multi-modality predictive model for CVD in BC survivors risk stratification in oncological survivorship. The data integration pipeline and subsequent predictive models establish a methodological foundation for future research into personalized healthcare.

Incorporating informatively collected laboratory data from EHR in clinical prediction models

BackgroundElectronic Health Records (EHR) are widely used to develop clinical prediction models (CPMs). However, one of the challenges is that there is often a degree of informative missing data. For example, laboratory measures are typically taken when a clinician is concerned that there is a need. When data are the so-called Not Missing at Random (NMAR), analytic strategies based on other missingness mechanisms are inappropriate. In this work, we seek to compare the impact of different strategies for handling missing data on CPMs performance.MethodsWe considered a predictive model for rapid inpatient deterioration as an exemplar implementation. This model incorporated twelve laboratory measures with varying levels of missingness. Five labs had missingness rate levels around 50%, and the other seven had missingness levels around 90%. We included them based on the belief that their missingness status can be highly informational for the prediction. In our study, we explicitly compared the various missing data strategies: mean imputation, normal-value imputation, conditional imputation, categorical encoding, and missingness embeddings. Some of these were also combined with the last observation carried forward (LOCF). We implemented logistic LASSO regression, multilayer perceptron (MLP), and long short-term memory (LSTM) models as the downstream classifiers. We compared the AUROC of testing data and used bootstrapping to construct 95% confidence intervals.ResultsWe had 105,198 inpatient encounters, with 4.7% having experienced the deterioration outcome of interest. LSTM models generally outperformed other cross-sectional models, where embedding approaches and categorical encoding yielded the best results. For the cross-sectional models, normal-value imputation with LOCF generated the best results.ConclusionStrategies that accounted for the possibility of NMAR missing data yielded better model performance than those did not. The embedding method had an advantage as it did not require prior clinical knowledge. Using LOCF could enhance the performance of cross-sectional models but have countereffects in LSTM models.

Prediction of antimicrobial resistance in Klebsiella pneumoniae using genomic and metagenomic next-generation sequencing data.

Klebsiella pneumoniae is a significant pathogen with increasing resistance and high mortality rates. Conventional antibiotic susceptibility testing methods are time-consuming. Next-generation sequencing has shown promise for predicting antimicrobial resistance (AMR). This study aims to develop prediction models using whole-genome sequencing data and assess their feasibility with metagenomic next-generation sequencing data from clinical samples. On the basis of 4170 K. pneumoniae genomes, the main genetic characteristics associated with AMR were identified using a LASSO regression model. Consequently, the prediction model was established, validated and optimized using clinical isolate read simulation sequences. To evaluate the efficacy of the model, clinical specimens were collected. Four predictive models for amikacin, ciprofloxacin, levofloxacin, and piperacillin/tazobactam, initially had positive predictive values (PPVs) of 92%, 98%, 99%, 94%, respectively, when they were originally constructed. When applied to clinical specimens, their PPVs were 96%, 96%, 95%, and 100%, respectively. Meanwhile, there were negative predictive values (NPVs) of 100% for ciprofloxacin and levofloxacin, and 'not applicable' (NA) for amikacin and piperacillin/tazobactam. Our method achieved antibacterial phenotype classification accuracy rates of 95.92% for amikacin, 96.15% for ciprofloxacin, 95.31% for levofloxacin and 100% for piperacillin/tazobactam. The sequence-based prediction antibiotic susceptibility testing (AST) reported results in an average time of 19.5 h, compared with the 67.9 h needed for culture-based AST, resulting in a significant reduction of 48.4 h. These preliminary results demonstrated that the performance of prediction model for a clinically significant antimicrobial-species pair was comparable to that of phenotypic methods, thereby encouraging the expansion of sequence-based susceptibility prediction and its clinical validation and application.

18F-Fluorodeoxyglucose Positron Emission Tomography-Based Risk Score Model for Prediction of Five-Year Survival Outcome after Curative Resection of Non-Small-Cell Lung Cancer.

The aim of our retrospective study is to develop and assess an imaging-based model utilizing 18F-FDG PET parameters for predicting the five-year survival in non-small-cell lung cancer (NSCLC) patients after curative surgery. A total of 361 NSCLC patients who underwent curative surgery were assigned to the training set (n = 253) and the test set (n = 108). The LASSO regression model was used to construct a PET-based risk score for predicting five-year survival. A hybrid model that combined the PET-based risk score and clinical variables was developed using multivariate logistic regression analysis. The predictive performance was determined by the area under the curve (AUC). The individual features with the best predictive performances were co-occurrence_contrast (AUC = 0.675) and SUL peak (AUC = 0.671). The PET-based risk score was identified as an independent predictor after adjusting for clinical variables (OR 5.231, 95% CI 1.987-6.932; p = 0.009). The hybrid model, which integrated clinical variables, significantly outperformed the PET-based risk score alone in predictive accuracy (AUC = 0.771 vs. 0.696, p = 0.022), a finding that was consistent in the test set. The PET-based risk score, especially when integrated with clinical variables, demonstrates good predictive ability for five-year survival in NSCLC patients following curative surgery.

Development of a prediction model for incidence of diabetic foot in patients with type 2 diabetes and its application based on a local health data platform

Objective: To construct a diabetes foot prediction model for adult patients with type 2 diabetes based on retrospective cohort study using data from a regional health data platform. Methods: Using Yinzhou Health Information Platform of Ningbo, adult patients with newly diagnosed type 2 diabetes from January 1, 2015 to December 31, 2022 were included in this study and divided randomly the train and test sets according to the ratio of 7∶3. LASSO regression model and bidirectional stepwise regression model were used to identify risk factors, and model comparisons were conducted with net reclassification index, integrated discrimination improvement and concordance index. Univariate and multivariate Cox proportional hazard regression models were constructed, and a nomogram plot was drawn. Area under the curve (AUC) was calculated as a discriminant evaluation indicator for model validation test its calibration ability, and calibration curves were drawn to test its calibration ability. Results: No significant difference existed between LASSO regression model and bidirectional stepwise regression model, but the better bidirectional stepwise regression model was selected as the final model. The risk factors included age of onset, gender, hemoglobin A1c, estimated glomerular filtration rate, taking angiotensin receptor blocker and smoking history. AUC values (95%CI) of risk outcome prediction at year 5 and 7 were 0.700 (0.650-0.749) and 0.715(0.668-0.762) for the train set and 0.738 (0.667-0.801) and 0.723 (0.663-0.783) for the test set, respectively. The calibration curves were close to the ideal curve, and the model discrimination and calibration powers were both good. Conclusions: This study established a convenient prediction model for diabetic foot and classified the risk levels. The model has strong interpretability, good discrimination power, and satisfactory calibration and can be used to predict the incidence of diabetes foot in adult patients with type 2 diabetes to provide a basis for self-assessment and clinical prediction of diabetic foot disease risk.

Analysis and Prediction of Influencing Factors of American Real Estate Price Based on Machine Learning

The American real estate market has always been a subject of widespread interest, with its price fluctuations impacting not only the lives of millions of American households but also exerting profound effects on the nation's overall economic health. This paper focuses on analyzing which factors are more significant in influencing the American real estate market in order to gain a better understanding of the factors influencing this market and to offer insightful information for stakeholders, policymakers, and investors navigating this dynamic market. We apply linear regression model, lasso regression model and random forest regression model to United States real estate data to explore variables that influence the U.S. housing price and find that PPI, M3, and population are the critical variables impacting the housing price most. By comparing the performance of three regression models, we identify the random forest regression model as the optimal model for predicting real estate prices.

Identification and Prognostic Value of m6A-Related Genes in Glioblastoma.

N6-methyladenosine (m6A) is one of the most common forms of mRNA modification, which is dynamically regulated by the m6A-related genes; however, its effect in glioblastoma (GBM) is still unknown. We sought to investigate the association between m6A-related genes (m6A-RGs) and GBM. Transcriptome data and the relevant clinical data were downloaded from The Cancer Genome Atlas and Gene Expression Omnibus databases. The m6A-RGs were identified from differently expressed genes, and COX and lasso regression models were applied to locate the prognosis-related genes. We identified 15 out of 19 m6A-RGs differentially expressed between GBM and nontumor tissues. We identified two subgroups of GBM (clusters 1 and 2) by applying consensus clustering. Compared with the cluster 1 subgroup, the cluster 1 subgroup correlates with a poorer prognosis, and most of the 19 m6A-RGs are higher expressed in cluster 1. Through univariate Cox and lasso regression model, we identified three m6A-RGs, namely HNRNPC, ALKBH5, and FTO, which were used to construct a Cox regression risk model to predict the prognosis of GBM patients. We identified a valuable m6A model for predicting the prognosis of GBM patients, which can provide useful epigenetic biomarkers.

Exploration and verification a 13-gene diagnostic framework for ulcerative colitis across multiple platforms via machine learning algorithms

Ulcerative colitis (UC) is a chronic inflammatory bowel disease with intricate pathogenesis and varied presentation. Accurate diagnostic tools are imperative to detect and manage UC. This study sought to construct a robust diagnostic model using gene expression profiles and to identify key genes that differentiate UC patients from healthy controls. Gene expression profiles from eight cohorts, encompassing a total of 335 UC patients and 129 healthy controls, were analyzed. A total of 7530 gene sets were computed using the GSEA method. Subsequent batch correction, PCA plots, and intersection analysis identified crucial pathways and genes. Machine learning, incorporating 101 algorithm combinations, was employed to develop diagnostic models. Verification was done using four external cohorts, adding depth to the sample repertoire. Evaluation of immune cell infiltration was undertaken through single-sample GSEA. All statistical analyses were conducted using R (Version: 4.2.2), with significance set at a P value below 0.05. Employing the GSEA method, 7530 gene sets were computed. From this, 19 intersecting pathways were discerned to be consistently upregulated across all cohorts, which pertained to cell adhesion, development, metabolism, immune response, and protein regulation. This corresponded to 83 unique genes. Machine learning insights culminated in the LASSO regression model, which outperformed others with an average AUC of 0.942. This model's efficacy was further ratified across four external cohorts, with AUC values ranging from 0.694 to 0.873 and significant Kappa statistics indicating its predictive accuracy. The LASSO logistic regression model highlighted 13 genes, with LCN2, ASS1, and IRAK3 emerging as pivotal. Notably, LCN2 showcased significantly heightened expression in active UC patients compared to both non-active patients and healthy controls (P < 0.05). Investigations into the correlation between these genes and immune cell infiltration in UC highlighted activated dendritic cells, with statistically significant positive correlations noted for LCN2 and IRAK3 across multiple datasets. Through comprehensive gene expression analysis and machine learning, a potent LASSO-based diagnostic model for UC was developed. Genes such as LCN2, ASS1, and IRAK3 hold potential as both diagnostic markers and therapeutic targets, offering a promising direction for future UC research and clinical application.

The divergence and driving factors of corporate environmental information disclosure in China

Corporate environmental information disclosure (EID) has been a widely acknowledged issue in the field of ecological environment research. This paper explores the spatiotemporal evolution characteristics of EID in China from 2008 to 2022, and discusses the industry and regional heterogeneity. It reveals significant differences in EID between heavy and non-heavy polluting industries, as well as among different economic regions, revealing that there is still significant potential for improvement in corporate environmental governance. Further, this paper establishes a two-way fixed effects model and empirically investigates the effects of internal and external driving factors on EID. In addition, Lasso regression model is used to screen the vital driving factors and analyze their relative importance. The empirical results show that enterprise size, management age, operating capacity, supervisory board size, and debt capacity have relatively more substantial explanations for changes in EID, which are more critical driving factors. The contribution of the study is to provide valuable reference for constructing corporate EID indicator system, and to points out the direction for the sustainable development of enterprises in different industries and regions. At the same time, it has theoretical guiding significance for finding the factors that promote EID and strengthening regional environmental governance.

Presenting a Multispectral Image Sensor for Quantification of Total Polyphenols in Low-Temperature Stressed Tomato Seedlings Using Hyperspectral Imaging.

Hyperspectral imaging was used to predict the total polyphenol content in low-temperature stressed tomato seedlings for the development of a multispectral image sensor. The spectral data with a full width at half maximum (FWHM) of 5 nm were merged to obtain FWHMs of 10 nm, 25 nm, and 50 nm using a commercialized bandpass filter. Using the permutation importance method and regression coefficients, we developed the least absolute shrinkage and selection operator (Lasso) regression models by setting the band number to ≥11, ≤10, and ≤5 for each FWHM. The regression model using 56 bands with an FWHM of 5 nm resulted in an R2 of 0.71, an RMSE of 3.99 mg/g, and an RE of 9.04%, whereas the model developed using the spectral data of only 5 bands with a FWHM of 25 nm (at 519.5 nm, 620.1 nm, 660.3 nm, 719.8 nm, and 980.3 nm) provided an R2 of 0.62, an RMSE of 4.54 mg/g, and an RE of 10.3%. These results show that a multispectral image sensor can be developed to predict the total polyphenol content of tomato seedlings subjected to low-temperature stress, paving the way for energy saving and low-temperature stress damage prevention in vegetable seedling production.

Multi-factor normalisation of viral counts from wastewater improves the detection accuracy of viral disease in the community

The detection of viruses (e.g. SARS-CoV-2, norovirus) in wastewater represents an effective way to monitor the prevalence of these pathogens circulating within the community. However, accurate quantification of viral concentrations in wastewater, proportional to human input, is constrained by a range of uncertainties, including (i) dilution within the sewer network, (ii) degradation of viral RNA during wastewater transit, (iii) catchment population and facility use, (iv) efficiency of viral concentration and extraction from wastewater, and (v) inhibition of amplification during the RT-qPCR step. Here, we address these uncertainties by investigating several potential normalisation factors including the concentration of ammonium and orthophosphate. A faecal indicator virus (crAssphage), and the recovery of the process-control viruses (murine norovirus and bacteriophage Phi6), used for quality control during the RT-qPCR step, were also considered. We found that multi-factor normalisation of SARS-CoV-2 RT-qPCR data was optimal using a combination of crAssphage, process-control virus recovery, and concentration efficiency to improve prediction accuracy relative to clinical test data. Using multi-normalised SARS-CoV-2 RT-qPCR data, we found a lasso regression model with random forest modelled residuals lowers the prediction error of positives by 46 %, compared to a single linear regression using raw data. This multi-normalised approach enables more accurate wastewater-based predictions of clinical cases up to five days in advance of clinical data, identifying trends in disease prevalence before clinical testing, and demonstrates the potential to improve viral pathogen detection for a range of currently monitored and emerging diseases.

Development and validation of a clinical decision tool for preoperative micropapillary and solid pattern lung adenocarcinoma of CT≤2cm.

Micropapillary (MP) and solid(S) pattern adenocarcinoma are highly malignant subtypes of lung adenocarcinoma. In today's era of increasingly conservative surgery for small lung cancer, effective preoperative identification of these subtypes is greatly important for surgical planning and long term survival of patients. For this retrospective study, the presence of MP and/or S was evaluated in 2167 consecutive patients who underwent surgical resection for clinical stage IA1-2 lung adenocarcinoma. MP and/or S pattern-positive patients and negative-pattern patients were matched at a ratio of 1:3. The Lasso regression model was used for data dimension reduction and imaging signature building. Multivariate logistic regression was used to establish the predictive model, presented as an imaging nomogram. The performance of the nomogram was assessed based on calibration, identification, and clinical usefulness, and internal and external validation of the model was conducted. The proportion of solid components (PSC), Sphericity, entropy, Shape, bronchial honeycomb, nodule shape, sex, and smoking were independent factors in the prediction model of MP and/or S lung adenocarcinoma. The model showed good discrimination with an area under the ROC curve of 0.85. DCA demonstrated that the model could achieve good benefits for patients. RCS analysis suggested a significant increase in the proportion of MP and/or S from 11% to 48% when the PSC value was 68%. Small MP and/or S adenocarcinoma can be effectively identified preoperatively by their typical 3D and 2D imaging features.

Development of forecasting of monthly SAR time series in river systems: A multivariate data decomposition-based hybrid approach

The sodium adsorption ratio (SAR) is a critical variable in assessing the quality of water resources, and accurately forecasting its time series is operationally valuable. This study developed a hybrid approach using the multivariate variational mode decomposition (MVMD) model for signal analysis, Boruta model for feature selection, pure linear neural network (PLNN), support vector regression (SVR), Lasso regression, and Elman neural network (ENN) models to forecast monthly SAR time series for rivers. Data from two rivers were used to enhance result reliability, and the developed models were compared with corresponding basic models to evaluate their impact. Numerical and graphical criteria demonstrated the significant superiority of the developed models over the basic ones. Among the basic models, the ENN model exhibited the highest accuracy, while the MVMD-Boruta-ENN model surpasses all investigated models. This finding suggested that the ENN model's structure is more suitable for SAR time series forecasting than other basic models. Analyzing the models' residuals revealed lower mean, standard deviation, skewness, and error range in the developed models, indicating their robust behavior in forecasting the SAR time series. Notably, forecasting extreme SAR values holds greater importance than other values. Anderson-Darling and Kolmogorov-Smirnov tests identified the dominance of the generalized logistic and log-logistic (3-parameters) functions in SAR time series. Probability distribution functions were used to estimate extreme values, and the studied models exhibited more accurate estimations compared to the basic models, indicating their enhanced resilience. The consistent patterns observed in comparing developed and basic models for the entire series, as well as extreme values and residuals across the two investigated rivers, emphasized the reliability of the results.

Combining developmental and sleep health measures for autism spectrum disorder screening: an ECHO study.

Sleep problems are reported for up to 80% of autistic individuals. We examined whether parsimonious sets of items derived from the Modified Checklist for Autism in Toddlers, Revised (M-CHAT-R) and the Brief Infant Sleep Questionnaire (BISQ) are superior to the standard M-CHAT-R in predicting subsequent autism spectrum disorder (ASD) diagnoses. Participants from 11 Environmental influences on Child Health Outcomes (ECHO) cohorts were included. We performed logistic LASSO regression models with 10-fold cross-validation to identify whether a combination of items derived from the M-CHAT-R and BISQ are superior to the standard M-CHAT-R in predicting ASD diagnoses. The final sample comprised 1552 children. The standard M-CHAT-R had a sensitivity of 44% (95% CI: 34, 55), specificity of 92% (95% CI: 91, 94), and AUROC of 0.726 (95% CI: 0.663, 0.790). A higher proportion of children with ASD had difficulty falling asleep or resisted bedtime during infancy/toddlerhood. However, LASSO models revealed parental reports of sleep problems did not improve the accuracy of the M-CHAT-R in predicting ASD diagnosis. While children with ASD had higher rates of sleep problems during infancy/toddlerhood, there was no improvement in ASD developmental screening through the incorporation of parent-report sleep metrics. Parental-reported sleep problems are common in autism spectrum disorder (ASD). We investigated whether the inclusion of parental-reports of infant/toddler sleep patterns enhanced the effectiveness of developmental screening for autism. We reported higher rates of difficulty falling asleep and resisting bedtime during infancy and toddlerhood among children later diagnosed with ASD; however, we did not find an improvement in ASD developmental screening through the incorporation of parent-report sleep metrics. In our sample, the standard M-CHAT-R had a sensitivity of 39% among children of mothers with government insurance compared with a sensitivity of 53% among children of mothers with employer-based insurance.

Benchmarking the performance and uncertainty of machine learning models in estimating scour depth at sluice outlets

ABSTRACT This study investigates the performance of six machine learning (ML) models – Random Forest (RF), Adaptive Boosting (ADA), CatBoost (CAT), Support Vector Machine (SVM), Lasso Regression (LAS), and Artificial Neural Network (ANN) – against traditional empirical formulas for estimating maximum scour depth after sluice gates. Our findings indicate that ML models generally outperform empirical formulas, with correlation coefficients (CORR) ranging from 0.882 to 0.944 for ML models compared with 0.835–0.847 for empirical methods. Notably, ANN exhibited the highest performance, followed closely by CAT, with a CORR of 0.936. RF, ADA, and SVM performed competitive metrics around 0.928. Variable importance assessments highlighted the dimensionless densimetric Froude number (Fd) as significantly influential, particularly in RF, CAT, and LAS models. Furthermore, SHAP value analysis provided insights into each predictor's impact on model outputs. Uncertainty assessment through Monte Carlo (MC) and Bootstrap (BS) methods, with 1,000 iterations, indicated ML's capability to produce reliable uncertainty maps. ANN leads in performance with higher mean values and lower standard deviations, followed by CAT. MC results trend towards optimistic predictions compared with BS, as reflected in median values and interquartile ranges. This analysis underscores the efficacy of ML models in providing precise and reliable scour depth predictions.

Statistical modelling of height growth in urban forestry plantations

ABSTRACT Urban plantation dynamics in different topographical and climatic conditions in Odisha were evaluated using linear model selection and regularisation techniques. The main objective was to evaluate how and to what extent the urban plantations respond to various climatic and edaphic conditions. The relationship between vegetation growth and climatic and soil parameters was studied using four statistical learning tools, subset selection, ridge regression, lasso, and partial least squares regression, and their performance was compared to a multiple regression model. The test MSE for the subset selection, ridge regression, lasso, and partial least squares regression models was evaluated to be 16,261.54, 12245.11, 16263.79 and 14,317.21, respectively. Results proved that statistical learning methods, namely subset selection, lasso, ridge regressions and partial least squares regression, were more accurate than multiple linear regression. From the results, it can be safely concluded that temperature shows greater correlation with the growth parameters. Precipitation also plays a vital role in vegetation dynamics. Soil parameters indicate a positive correlation with that of the growth.

HTR2B as a novel biomarker of chronic obstructive pulmonary disease with lung squamous cell carcinoma

Chronic obstructive pulmonary disease (COPD) is often associated with lung squamous cell carcinoma (LUSC), which has the same etiology (smoking, inflammation, oxidative stress, microenvironmental changes, and genetics). Smoking, inflammation, and airway remodeling are the most important and classical mechanisms of COPD comorbidity in LUSC patients. Cancer can occur during repeated airway damage and repair (airway remodeling). Changes in the inflammatory and immune microenvironments, which can cause malignant transformation of some cells, are currently being revealed in both LUSC and COPD patients. We obtained the GSE76925 dataset from the Gene Expression Omnibus database. Screening for possible COPD biomarkers was performed using the LASSO regression model and a random forest classifier. The compositional patterns of the immune cell fraction in COPD patients were determined using CIBERSORT. HTR2B expression was analyzed using validation datasets (GSE47460, GSE106986, and GSE1650). HTR2B expression in COPD cell models was determined via real-time quantitative PCR. Epithelial–mesenchymal transition (EMT) marker expression levels were determined after knocking down or overexpressing HTR2B. HTR2B function and mechanism in LUSC were analyzed with the Kaplan‒Meier plotter database. HTR2B expression was inhibited to detect changes in LUSC cell proliferation. A total of 1082 differentially expressed genes (DEGs) were identified in the GSE76925 dataset (371 genes were significantly upregulated, and 711 genes were significantly downregulated). Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis indicated that the DEGs were mainly enriched in the p53 signaling and β-alanine metabolism pathways. Gene Ontology enrichment analysis indicated that the DEGs were largely related to transcription initiation from the RNA polymerase I promoter and to the regulation of mononuclear cell proliferation. The LASSO regression model and random forest classifier results revealed that HTR2B, DPYS, FRY, and CD19 were key COPD genes. Immune cell infiltration analysis indicated that these genes were closely associated with immune cells. Analysis of the validation sets suggested that HTR2B was upregulated in COPD patients. HTR2B was significantly upregulated in COPD cell models, and its upregulation was associated with increased EMT marker expression. Compared with that in bronchial epithelial cells, HTR2B expression was upregulated in LUSC cells, and inhibiting HTR2B expression led to the inhibition of LUSC cell proliferation. In conclusions, HTR2B might be a new biomarker and therapeutic target in COPD patients with LUSC.

Association between dietary inflammatory index and NT-proBNP levels in US adults: A cross-sectional analysis.

With cardiovascular diseases standing as a leading cause of mortality worldwide, the interplay between diet-induced inflammation, as quantified by the Dietary Inflammatory Index (DII), and heart failure biomarker NT-proBNP has not been investigated in the general population. This study analyzed data from the National Health and Nutrition Examination Survey (NHANES) 1999-2004, encompassing 10,766 individuals. The relationship between the DII and NT-proBNP levels was evaluated through multivariable-adjusted regression models. To pinpoint crucial dietary components influencing NT-proBNP levels, the LASSO regression model was utilized. Stratified analyses were then conducted to examine the associations within specific subgroups to identify differential effects of the DII on NT-proBNP levels across diverse populations. In individuals without heart failure, a unit increase in the DII was significantly associated with an increase in NT-proBNP levels. Specifically, NT-proBNP levels rose by 9.69 pg/mL (95% CI: 6.47, 12.91; p < 0.001) without adjustments, 8.57 pg/mL (95% CI: 4.97, 12.17; p < 0.001) after adjusting for demographic factors, and 5.54 pg/mL (95% CI: 1.75, 9.32; p = 0.001) with further adjustments for health variables. In participants with a history of heart failure, those in the second and third DII quartile showed a trend towards higher NT-proBNP levels compared to those in the lowest quartile, with increases of 717.06 pg/mL (95% CI: 76.49-1357.63, p = 0.030) and 855.49 pg/mL (95% CI: 156.57-1554.41, p = 0.018). Significant interactions were observed in subgroup analyses by age (<50: β = 3.63, p = 0.141; 50-75: β = 18.4, p<0.001; >75: β = 56.09, p<0.001), gender (men: β = 17.82, p<0.001; women: β = 7.43, p = 0.061),hypertension (β = 25.73, p<0.001) and diabetes (β = 38.94, p<0.001). This study identified a positive correlation between the DII and NT-proBNP levels, suggesting a robust link between pro-inflammatory diets and increased heart failure biomarkers, with implications for dietary modifications in cardiovascular risk management.