Lung cancer is one of the most dangerous malignant tumors affecting human health. Lung adenocarcinoma (LUAD) is the most common subtype of lung cancer. Both glycolytic and cholesterogenic pathways play critical roles in metabolic adaptation to cancer. A dataset of 585 LUAD samples was downloaded from The Cancer Genome Atlas database. We obtained co-expressed glycolysis and cholesterogenesis genes by selecting and clustering genes from Molecular Signatures Database v7.5. We compared the prognosis of different subtypes and identified differentially expressed genes between subtypes. Predictive outcome events were modeled using machine learning, and the top 9 most important prognostic genes were selected by Shapley additive explanation analysis. A risk score model was built based on multivariate Cox analysis. LUAD patients were categorized into four metabolic subgroups: cholesterogenic, glycolytic, quiescent, and mixed. The worst prognosis was the mixed subtype. The prognostic model had great predictive performance in the test set. Patients with LUAD were effectively typed by glycolytic and cholesterogenic genes and were identified as having the worst prognosis in the glycolytic and cholesterogenic enriched gene groups. The prognostic model can provide an essential basis for clinicians to predict clinical outcomes for patients. The model was robust on the training and test datasets and had a great predictive performance.

[This corrects the article DOI: 10.3389/fmed.2024.1285067.].

UNASSIGNED: The purpose of this study is to explore whether machine learning can be used to establish an effective model for the diagnosis of Parkinson\'s disease (PD) by using texture features extracted from cerebellar gray matter and white matter, so as to identify subtle changes that cannot be observed by the naked eye.
UNASSIGNED: This study involved a data collection period from June 2010 to March 2023, including 374 subjects from two cohorts. The Parkinson\'s Progression Markers Initiative (PPMI) served as the training set, with control group and PD patients (HC: 102 and PD: 102) from 24 global sites. Our institution\'s data was utilized as the test set (HC: 91 and PD: 79). Machine learning was employed to establish multiple models for PD diagnosis based on texture features of the cerebellum\'s gray and white matter. Results underwent evaluation through 5-fold cross-validation analysis, calculating the area under the receiver operating characteristic curve (AUC) for each model. The performance of each model was compared using the Delong test, and the interpretability of the optimized model was further augmented by employing Shapley additive explanations (SHAP).
UNASSIGNED: The AUCs for all pipelines in the validation dataset were compared using FeAture Explorer (FAE) software. Among the models established by Kruskal-Wallis (KW) and logistic regression via Lasso (LRLasso), the AUC was highest using the \"one-standard error\" rule. \'WM_original_glrlm_GrayLevelNonUniformity\' was considered the most stable and predictive feature.
UNASSIGNED: The texture features of cerebellar gray matter and white matter combined with machine learning may have potential value in the diagnosis of Parkinson\'s disease, in which the heterogeneity of white matter may be a more valuable imaging marker.

Ternary polymer solar cells (PSCs) are currently the simplest and most efficient way to further improve the device performance in PSCs. To find high-performance organic photovoltaic materials, the established connection between the material structure and device performance before fabrication is of great significance. Herein, firstly, a database of the photovoltaic performance in 874 experimental PSCs reported in the literature is established, and three different fingerprint expressions of a molecular structure are explored as input features; the results show that long fingerprints of 2D atom pairs can contain more effective information and improve the accuracy of the models. Through supervised learning, five machine learning (ML) models were trained to build a mapping of the photovoltaic performance improvement relationship from binary to ternary PSCs. The GBDT model had the best predictive ability and generalization. Eighteen key structural features from a non-fullerene acceptor and the third components that affect the device\'s PCE were screened based on this model, including a nitrile group with lone-pair electron, a halogen atom, an oxygen atom, etc. Interestingly, the structural features for the enhanced device\'s PCE were essentially increased by the Jsc or FF. More importantly, the reliability of the ML model was further verified by preparing the highly efficient PSCs. Taking the PM6:BTP-eC9:PY-IT ternary PSC as an example, the PCE prediction (18.03%) by the model was in good agreement with the experimental results (17.78%), the relative prediction error was 1.41%, and the relative error between all experimental results and predicted results was less than 5%. These results indicate that ML is a useful tool for exploring the photovoltaic performance improvement of PSCs and accelerating the design and application with highly efficient non-fullerene materials.

Conventional environmental health research is primarily focused on isolated chemical exposures, neglecting the complex interactions between multiple pollutants that may synergistically or antagonistically influence toxicity, thereby posing unexpected health risks. In this study, we address this knowledge gap by introducing an explainable machine learning (ML) approach with Feature Localized Intercept Transformed-Shapley Additive Explanations (FLIT-SHAP) designed to extract the dose-response relationships of specific pollutants in mixtures. In contrast to traditional SHAP, FLIT-SHAP can localize the global intercept to elucidate mixture effects, which is crucial for understanding the oxidative potential (OP) of ambient particulate matter (PM). Assessing multi-pollutant OP using FLIT-SHAP revealed both synergistic (55-63 %) and antagonistic (25-42 %) effects in laboratory-controlled OP data, but an antagonistic (33-66 %; lower OP) effect in ambient PM. Notably, the FLIT-SHAP approach demonstrated higher prediction accuracy (R2 = 0.99) compared to the additive model (R2 = 0.89) when evaluated against real-world PM samples. Quinones, such as phenanthrenequinone, play a more significant role in PM2.5 than previously recognized. Through this study, we highlighted the potential of FLIT-SHAP to enhance toxicity predictions and aid decision-making in the field of environmental health.

BACKGROUND: Health care-associated infections due to multidrug-resistant organisms (MDROs), such as methicillin-resistant Staphylococcus aureus (MRSA) and Clostridioides difficile (CDI), place a significant burden on our health care infrastructure.
OBJECTIVE: Screening for MDROs is an important mechanism for preventing spread but is resource intensive. The objective of this study was to develop automated tools that can predict colonization or infection risk using electronic health record (EHR) data, provide useful information to aid infection control, and guide empiric antibiotic coverage.
METHODS: We retrospectively developed a machine learning model to detect MRSA colonization and infection in undifferentiated patients at the time of sample collection from hospitalized patients at the University of Virginia Hospital. We used clinical and nonclinical features derived from on-admission and throughout-stay information from the patient\'s EHR data to build the model. In addition, we used a class of features derived from contact networks in EHR data; these network features can capture patients\' contacts with providers and other patients, improving model interpretability and accuracy for predicting the outcome of surveillance tests for MRSA. Finally, we explored heterogeneous models for different patient subpopulations, for example, those admitted to an intensive care unit or emergency department or those with specific testing histories, which perform better.
RESULTS: We found that the penalized logistic regression performs better than other methods, and this model\'s performance measured in terms of its receiver operating characteristics-area under the curve score improves by nearly 11% when we use polynomial (second-degree) transformation of the features. Some significant features in predicting MDRO risk include antibiotic use, surgery, use of devices, dialysis, patient\'s comorbidity conditions, and network features. Among these, network features add the most value and improve the model\'s performance by at least 15%. The penalized logistic regression model with the same transformation of features also performs better than other models for specific patient subpopulations.
CONCLUSIONS: Our study shows that MRSA risk prediction can be conducted quite effectively by machine learning methods using clinical and nonclinical features derived from EHR data. Network features are the most predictive and provide significant improvement over prior methods. Furthermore, heterogeneous prediction models for different patient subpopulations enhance the model\'s performance.

BACKGROUND: Chronic stress is highly prevalent in the German population. It has known adverse effects on mental health, such as burnout and depression. Known long-term effects of chronic stress are cardiovascular disease, diabetes, and cancer.
OBJECTIVE: This study aims to derive an interpretable multiclass machine learning model for predicting chronic stress levels and factors protecting against chronic stress based on representative nationwide data from the German Health Interview and Examination Survey for Adults, which is part of the national health monitoring program.
METHODS: A data set from the German Health Interview and Examination Survey for Adults study including demographic, clinical, and laboratory data from 5801 participants was analyzed. A multiclass eXtreme Gradient Boosting (XGBoost) model was constructed to classify participants into 3 categories including low, middle, and high chronic stress levels. The model\'s performance was evaluated using the area under the receiver operating characteristic curve, precision, recall, specificity, and the F1-score. Additionally, SHapley Additive exPlanations was used to interpret the prediction XGBoost model and to identify factors protecting against chronic stress.
RESULTS: The multiclass XGBoost model exhibited the macroaverage scores, with an area under the receiver operating characteristic curve of 81%, precision of 63%, recall of 52%, specificity of 78%, and F1-score of 54%. The most important features for low-level chronic stress were male gender, very good general health, high satisfaction with living space, and strong social support.
CONCLUSIONS: This study presents a multiclass interpretable prediction model for chronic stress in adults in Germany. The explainable artificial intelligence technique SHapley Additive exPlanations identified relevant protective factors for chronic stress, which need to be considered when developing interventions to reduce chronic stress.

Carbon price is a pivotal element in the carbon trading sector. Accurate estimation of carbon price can offer precise guidance for the carbon market participants. This study introduces a novel prediction model encompassing both point and interval prediction for the carbon price. Firstly, to distill the volatility traits inherent in carbon price, the successive variational mode decomposition is utilized to adaptively decompose the carbon price into regular sequences. Secondly, to obtain the optimal input variables, the partial autocorrelation function and random forest are employed to filter the influencing factors and historical carbon price. Then, to avoid single model constraint, a combination model of categorical boosting and kernel extreme learning machine optimized by the sparrow search algorithm is employed for the point prediction, and the shapley additive explanation is employed to elucidate the model prediction process. Finally, to provide more efficient information, the adaptive bandwidth kernel density estimation is applied to the interval prediction. The data from Hubei carbon market is adopted as a case study, and the results indicate that the mean absolute error, mean absolute percentage error, root mean square error and R2 of the proposed model are 0.1022, 0.0022, 0.1262 and 0.9921, respectively. The historical carbon price, Brent crude oil futures settlement price and European Union allowance futures carbon price have a positive impact on carbon price, and Hushen 300 has a negative impact on carbon price. Compared with the constant kernel density estimation, the proposed model achieves higher interval coverage probability and lower interval width. Thus, the application of the hybrid model can promote the operational efficiency of the carbon market and facilitate the implementation of carbon emission reduction policies.

The use of deicers in urban areas, on runways and aircrafts has raised concerns about their environmental impact. Understanding the ice-melting mechanism is crucial for developing environmentally friendly deicers, yet it remains challenging. This study employs machine learning to investigate the ice penetration capacity (IPC) of 21 salts and 16 organic solvents as deicers. Relationships between their IPC and various physical properties were analysed using extreme gradient boosting (XGBoost) and Shapley additive explanation (SHAP). Three key ice-melting mechanisms were identified: (1) freezing-point depression, (2) interactions between deicers and H2O molecules and (3) infiltration of ions into ice crystals. SHAP analysis revealed different ice-melting factors and mechanisms for salts and organic solvents, suggesting a potential advantage in combining the two. A mixture of propylene glycol (PG) and sodium formate demonstrated superior environmental impact and IPC. The PG and sodium formate mixture exhibited higher IPC when compared to six commercially available deicers, offering promise for sustainable deicing applications. This study provides valuable insights into the ice-melting process and proposes an effective, environmentally friendly deicer that combines the strengths of organic solvents and salts, paving the way for more sustainable practices in deicing.

OBJECTIVE: Although there are various model-based approaches to individualized vancomycin (VCM) administration, few have been reported for adult patients with periprosthetic joint infection (PJI). This work attempted to develop a machine learning (ML)-based model for predicting VCM trough concentration in adult PJI patients.
METHODS: The dataset of 287 VCM trough concentrations from 130 adult PJI patients was split into a training set (229) and a testing set (58) at a ratio of 8:2, and an independent external 32 concentrations were collected as a validation set. A total of 13 covariates and the target variable (VCM trough concentration) were included in the dataset. A covariate model was respectively constructed by support vector regression, random forest regression and gradient boosted regression trees and interpreted by SHapley Additive exPlanation (SHAP).
RESULTS: The SHAP plots visualized the weight of the covariates in the models, with estimated glomerular filtration rate and VCM daily dose as the 2 most important factors, which were adopted for the model construction. Random forest regression was the optimal ML algorithm with a relative accuracy of 82.8% and absolute accuracy of 67.2% (R2 =.61, mean absolute error = 2.4, mean square error = 10.1), and its prediction performance was verified in the validation set.
CONCLUSIONS: The proposed ML-based model can satisfactorily predict the VCM trough concentration in adult PJI patients. Its construction can be facilitated with only 2 clinical parameters (estimated glomerular filtration rate and VCM daily dose), and prediction accuracy can be rationalized by SHAP values, which highlights a profound practical value for clinical dosing guidance and timely treatment.