Fig. 3: Comprehensive analysis of survival prediction and clinical associations using TCGA multi-omics data.
From: GAUDI: interpretable multi-omics integration with UMAP embeddings and density-based clustering
A Bonferroni-corrected p-values for each of the 5 factors identified by multi-omics integration methods as predictive of survival, based on a Cox regression analysis. Dotted lines indicate the threshold for a corrected p-value of 0.05. Results for the remaining five cancer types are available in Supplementary Fig. 2. B Number of factors associated with clinical annotations in each method and the selectivity score of the factor-annotation associations. The dashed horizontal line represents the mean selectivity score. C Kaplan–Meier survival curves for AML clusters with the highest and lowest median survival times identified by each method. The log-rank test is used for p-value calculations, and the sample size for each cluster is provided (N). Source data are provided as a Source Data file.
