Research Paper Volume 16, Issue 9 pp 7774—7798

Figure 2. (A) A bar chart showcasing the “importance” of MPIGs arranged in descending order, derived from a random forest analysis. Greater importance of a variable indicates a larger discrepancy in prediction accuracy between models with and without the variable. (B) A balloon plot demonstrates significant regulatory relationships between WERs and corresponding MPIGs proven by perturbation experiments from RM2Target database. (C) Venn diagrams highlight that out of 28 PIGs identified in malignant cell lines from the RM2Target database through MeRIP-seq analysis, 20 exhibit significant m6A methylation modifications. Among the top 10 PIGs (MPIGs) for importance, 8 are methylated. Hypergeometric distribution tests suggest a propensity for m6A methylation in PIGs associated with prognosis. Further, in the PRJNA733602 dataset, 7 out of the top 10 most important PIGs show enhanced m6A peak differences, underscoring a correlation between prognostic relevance and increased m6A modification levels. (D) A network diagram reveals the regulatory relationships between WERs and their regulated PIGs within the RM2Target database. (E) Balloon plots display the predictive performance of Cox models constructed with MPGs for UC prognosis across multiple datasets, as indicated by the C-index. (F) A nomogram drawn from Cox models based on MPGs in the GSE32894 dataset, illustrating individual patient scores and their corresponding survival probabilities.