Search

Introduction: Acute ischemic stroke (AIS) accounts for the majority of all stroke, globally the second leading cause of death. Due to its rapid development after onset, its early diagnosis is crucial. Objectives: We aim to identify potential highly reliable blood-based biomarkers for early diagnosis of AIS using quantitative plasma lipid profiling via a machine learning approach. Methods: Lipidomics was used for quantitative plasma lipid profiling, based on ultra-performance liquid chromatography tandem mass spectrometry. Our samples were divided into a discovery and a validation set, each containing 30 AIS patients and 30 health controls (HC). Differentially expressed lipid metabolites were screened based on the criteria VIP > 1, p < 0.05, and fold change > 1.5 or < 0.67. The least absolute shrinkage and selection operator (LASSO) and random forest algorithms in machine learning were used to select differential lipid metabolites as potential biomarkers. Results: Three key differential lipid metabolites, CarnitineC10:1, CarnitineC10:1-OH and Cer(d18:0/16:0), were identified as potential biomarkers for early diagnosis of AIS. The former two, associated with thermogenesis, were down-regulated, whereas the latter, associated with necroptosis and sphingolipd metabolism, was upregulated. Univariate and multivariate logistic regressions showed that these three lipid metabolites and the resulting diagnostic model exhibited a strong ability in discriminating between AIS patients and HCs in both the discovery and validation sets, with an area under the curve above 0.9. Conclusions: Our work provides valuable information on the pathophysiology of AIS and constitutes an important step toward clinical application of blood-based biomarkers for diagnosing AIS. [ABSTRACT FROM AUTHOR]

The cell surface hydrophobicity (CSH) is an assessable physicochemical property used to evaluate the microbial adhesion to the surface of biomaterials, which is an essential step in the microbial biofilm formation and pathogenesis. For the present in vitro fermentation experiment, the CSH of ruminal mixed microbes was considered, along with other data records of pH, ammonia-nitrogen concentration, and neutral detergent fibre digestibility, conditions of surface tension and specific surface area in two different time scales. A dataset of 170,707 perturbations of input variables, grouped into two blocks of data, was constructed. Next, Expected Measurement Moving Average – Machine Learning (EMMA-ML) models were developed in order to predict CSH after perturbations of all input variables. EMMA-ML is a Perturbation Theory method that combines the ideas of Expected Measurement, Box-Jenkins Operators/Moving Average, and Time Series Analysis. Seven regression methods have been tested: Multiple Linear regression, Generalized Linear Model with Stepwise Feature Selection, Partial Least Squares regression, Lasso regression, Elastic Net regression, Neural Networks regression, and Random Forests (RF). The best regression performance has been obtained with RF (EMMA-RF model) with an R-squared of 0.992. The model analysis has shown that CSH values were highly dependent on the in vitro fermentation parameters of detergent fibre digestibility, ammonia – nitrogen concentration, and the expected values of cell surface hydrophobicity in the first time scale. [ABSTRACT FROM AUTHOR]