Fig. 1

Overview and basic statistics of the data. A Schematic overview of Canvolution workflow. For the preprocessing part, mutation profiles and gene expression from single-cell RNAseq data are used as input. After generation of the evolutionary tree and calculation of the abundance of each clone, mutational signature scores and transcriptional signature scores are obtained by evaluating multiple cancer-associated signatures and metabolic pathways. With cell–cell interaction information, Canvolution can also generate a ligand receptor (LR) mutation score for each clone. The evolutionary path analysis measures the correlation between the signature score and the tree depth for each path. Similarly, the clonal abundance measures the correlation between the signature score and the size of the clone. B Schematic workflow of the research presented here. Tumor tissue from lung (LC) and chronic myeloid leukemia (CML) patients were used. C Boxplots showing the number of mutations (left panel), the fraction of mutated genes (middle panel), and the ratio of mutation that are assigned as COSMIC (right panel) per clone