Figure 1.
Identification of subtypes through clustering and dimension reduction. (A) Uniform Manifold Approximation and Projection (UMAP) dimension reduction and density-based spatial clustering of samples used for training and holdout testing (n = 376), colored according to clusters, with cluster labels informed by Dai et al and Brady et al. (B-C) TARGET samples (n = 265) within our clustering colored according to subtype labels assigned by Dai et al (B) and Brady et al (C).