Fig. 6

The development and validation process of prognostic risk score model for predicting the occurrence of MASLD. (A, B) MASLD onset-related genes were identified through LASSO logistic regression analysis. The optimal penalty parameter lambda was selected using 10-fold cross-validation. (C) A heatmap visualizes the variation in risk scores across different samples based on the expression of 27 genes. (D) Heatmaps display the correlations in gene expression among a selected set of 27 genes. (E) Risk scores were compared across various MASLD stages, including healthy controls (HC), healthy obese individuals (HO), and those with simple steatosis (SS). (F) ROC analysis was conducted to assess the predictive value of risk score. (G) A comparison of risk scores between HC and SS samples was performed using the GSE66676 dataset. (H) ROC analysis evaluated the predictive value of risk score in the GSE66676 dataset. (I) Risk scores were compared between HC and HO samples using the GSE126848 dataset. (J) ROC analysis assessed the predictive value of risk score in the GSE126848 dataset.