Fig. 5. The TIME clusters have different PBM score in CD138+ cells shows that the malignancy of MM affects its TIME.
A Clustering using TIME expression profiles derived 5 clusters in the UAMS-I dataset. B KM-plot of 5 clusters in the UAMS-I dataset. Cluster-4 (C4) has poor PFS than others, and C2 has the best prognosis; C The TIME expression cluster 2 has the lowest PBM scores (average PBM score = −0.19), and cluster 4, characterized by low infiltration of granulocytes and higher infiltration of M1/M2 macrophages in TIME, had significantly higher PBM scores (average PBM = 1.46) than other clusters. The p values of clusters 1, 2, 3, and 5 compared to cluster 4 are 0.009, 2e−04, 4e−05, 0.008, 0.029, respectively. Wilcoxon’s rank-sum test. The heatmap of TIME removed cells with an average of less than 1% per sample, as well as osteoblast, osteoclast, and adipocyte. The values for each cell are presented as a distribution normalized to a mean of 0 and a standard deviation of 1. D MM patients with high PBM scores in CD138+ cells in all samples of the UAMS-I dataset have lower neutrophils, mast cells, resting NK cells, and higher CD8 T cell infiltration in their TIME.