Fig. 1

Composition of GBMAN’s four main subpopulations. (A) Integration of 127 human GBM scRNA-seq data across 9 individual datasets using BBKNN. A total of 498,747 single-cell transcriptomes were processed and integrated to reduce batch effects and ensure robust clustering. UMAP was employed to visualize the high-dimensional data in a two-dimensional space. (B) Violin plot illustrating marker gene expression for each major cell cluster. The plot highlights the expression levels of key markers across identified cell clusters. The width of each violin represents the distribution density of gene expression. (C) UMAP visualization of four distinct subgroups of neutrophils. Each subgroup is color-coded for clarity. (D) The box plot shows the ROGUE scores of the four subgroups. (E) FeaturePlot showing the spatial distribution of four marker genes across the UMAP embedding. (F) Heatmap depicting the correlation between the four neutrophil subgroups. Hierarchical clustering was performed to assess the similarity of gene expression profiles among subgroups, with colors representing correlation coefficients. (G) The proportion changes of the four neutrophil subtypes in primary and recurrent cases