Additional file 2: Figure S1. Demarcation of patients using a SNN modularity optimization-based clustering algorithm. A Unsupervised clustering of NSCLC samples with four clusters. B Unsupervised clustering of LUAD and LUSC samples. C Distribution of LUAD and LUSC samples in four different clusters. Figure S2. Associations between NSCLC subtypes and TME features. A Levels of stromal score, immune score, ESTIMATE score, and tumor purity in LUAD (n = 512), LUSC (n = 430), and LASC...