Fig. 3

5 prognosis genes were identified as by univariate Cox, LASSO, multivariate Cox and multiple stepwise regression analysis and Kaplan–Meier analysis of prognostic genes. (A) Forest plot of univariate Cox analysis. Through univariate Cox analysis, 17 genes were considered to be statistically significant for patient survival. (B) Prognostic genes cross-validation of LASSO regression. The x-axis is log(λ) and the y-axis is degrees of freedom. The dashed lines show the λ values for the smallest error mean and within 1 standard deviation. The top number shows the number of genes. Red dots show error values for each λ, with bars representing confidence intervals. (C) Plot of LASSO coefficient profiles. The x-axis is the log value of λ, the y-axis is the regression coefficient. The curve represents the relationship between the regression coefficient and λ, with the number at the top indicating the variables whose coefficient remains non-zero when λ is chosen. (D) Forest plot of multivariate Cox regression analysis. 9 model genes (SPAG4, BHLHE40, CD79B, STXBP4, DDHD1, FABP5, FKBP1B, TRAM2, PXDN) were identified using multivariate COX regression analysis. (E) The forest map of multivariate stepwise regression analysis. 5 model genes (CD79B, STXBP4, DDHD1, FKBP1B, TRAM2) were identified using multivariate stepwise regression analysis. F. Kaplan–Meier analysis of 5 prognostic progno genes. The K-M curve shows survival time and probability, with red representing high-risk and blue representing low-risk groups. The bottom half displays the risk list with remaining samples at different survival times. From left to right are CD79B,STXBP4, DDHD1,FKBP1B,TRAM2.