Skip to main content
. 2021 Oct 2;24(11):103200. doi: 10.1016/j.isci.2021.103200

Figure 4.

Figure 4

Association between the latent features learned by Miscell versus clinical and genomic features

(A) Box plot representation of latent features associated with ICB therapy. SD: stable disease, PD: progression disease, CR: complete response, PR: partial response. Adjusted p values were shown above each boxplot.

(B) Kaplan–Meier survival curves representation of the association between the latent feature and prognosis in TCGA pan-cancer dataset.

(C) Marker genes of the 31th feature and clinical information across 298 urothelial carcinoma patients.

(D) Heatmap representation of the association between the latent features versus clinical features, genomic and transcriptomic alteration signatures, and Hallmarks of cancer and immune-related circuits. Only latent features significantly associated with ICB therapy response and overall survival were shown.