Extended Data Fig. 6: Clinical analyses in TCGA data.
From: Integrating multimodal cancer data using deep latent variable path modelling
a: The heat maps shown in this figure illustrate associations between DLVs, estimated using the full DLVPM model, and clinical molecular and histological characteristics. Quantitative values in each heatmap are the Pearson’s Correlation Coefficient between different clinical types, and DLVs. Heatmap squares that are coloured are significant at the p < 0.05 FWER corrected level. Squares that are not coloured are non-significant (n = 152). b: Kaplan–Meier survival curves for patients stratified into high- and low-risk groups based on risk scores derived from a Cox proportional hazards model n the METABRIC dataset, stratified on RNASeq and SNV DLVs. Patients were divided into high- and low-risk groups based on the median risk score: those with scores above the median were classified as ‘High Risk,’ while those with scores below the median were classified as ‘Low Risk.’ The x-axis represents progression-free interval (PFI) in days, and the y-axis represents survival probability. Shaded regions indicate confidence intervals (n = 1980).
