Research Paper Volume 15, Issue 12 pp 5445—5481

Figure 5. Quantifying RNA modification patterns according to WM_Score. (A) Sankey plot showing the relationship between WM_Clusters, gene cluster, WM_Score and ACRG molecular subtypes. (B, C) WM_Scores according to different RNA modification patterns and gene clusters in the meta-GEO cohort. P < 0.05 in the Kruskal-Wallis test was considered statistically significant. (D, E) Kaplan–Meier curves showing the overall survival status of WM_Score-low (blue) and WM_Score-high (red) groups of patients in the meta-GEO cohort (D) and TCGA-STAD cohort (E), grouping status of the patients are indicated at the bottom of the chart. P < 0.05 in the Log-rank test was considered statistically significant. (F) Kaplan–Meier curves of WM_Score patients with or without adjuvant chemotherapy. H, high; L, low; ADJC, adjuvant chemotherapy. P < 0.001 by Log-rank test. (G, H) Multivariate Cox regression model of WM_Score, patient age, gender, TNM status, and patient outcomes in the meta-GEO cohort (G) and TCGA-STAD cohort (H) shown by the forest plots. (I) Predictive value of the WM_Score in patients amongst the meta-GEO cohort. The AUC of five, six, and seven-year survival was 0.612, 0.625, and 0.633, respectively.