Research Paper Volume 15, Issue 23 pp 14109—14140
Machine learning algorithm integrates bulk and single-cell transcriptome sequencing to reveal immune-related personalized therapy prediction features for pancreatic cancer
Figure 2. Exploration of the relationships between the regulation of immune cells and clusters. (A) Unsupervised consensus clustering based on 1612 IRGs. (B) The fractions of immune cells between Cluster_1 and Cluster_2. (C) The TIDE score between Cluster_1 and Cluster_2. (D) The differences in expression of common ICBs among distinct clusters. (E) ssGSEA analysis was utilized to estimate the abundance of immune cells. (F) The alluvial plot displayed the relationship between the TIME subtype and other molecular classifications. (G) Heatmap of Cramer’s V statistic reflected the corrections between seven PC molecular classifications.