A unique malignant cell type per patient tumor encoded in each cancer cell transcriptome

Deciphering shared features between patients through unsupervised analyses of tumor single-cell transcriptomes is hindered by the predominant clustering of malignant cells based on the patients' tumor of origin. In contrast, cancer-associated non-malignant cell cluster according to their cell type (e.g., macrophage), independently of the patient. We investigated the origin of this contrasting clustering behavior using computational analyses and data sampling techniques across 14 cancer types. We demonstrate that tumor-driven malignant cell clustering is independent of technical or computational biases and non-reducible to tumor-specific gene sets. Conversely, redundant information dispersed across the transcriptome encodes a unique identity shared by malignant cells within each patient's tumor. Identity of normal cell types is similarly encoded. Finally, we demonstrate maintenance of malignant cell identity across space and over time. These findings suggest the establishment of a distinct type of malignant cells within each patient's tumor, robustly and diffusively encoded across the entire transcribed genome.