CanSig: discovery of shared transcriptional states across cancer patients from single-cell RNA sequencing data

Multiple cancer types have been shown to exhibit heterogeneity in the transcriptional states of malignant cells across patients and within the same tumor. The intra-tumor transcriptional heterogeneity has been linked to resistance to therapy and cancer relapse, representing a significant obstacle to successful personalized cancer treatment. However, today there is no easy-to-use computational method to identify heterogeneous transcriptional cell states that are shared across patients from single-cell RNA sequencing (scRNA-seq) data. To discover shared transcriptional states of cancer cells, we propose a novel computational tool called CanSig. CanSig automatically preprocesses, integrates, and analyzes cancer scRNA-seq data from multiple patients to provide novel signatures of shared transcriptional states and associates these states with known biological pathways. CanSig jointly analyzes cells from multiple cancer patients while correcting for batch effects and differences in gene expressions caused by genetic heterogeneity. In our benchmarks, CanSig reliably re-discovers known transcriptional signatures on three previously published cancer scRNA-seq datasets, including four main cellular states of glioblastoma cells previously reported. We further illustrate CanSig’s investigative potential by uncovering signatures of novel transcriptional states in four additional cancer datasets. Some of the novel signatures are linked to cell migration and proliferation and to specific genomic aberrations and are enriched in more advanced tumors. In conclusion, CanSig detects transcriptional states that are common across different tumors. It facilitates the analysis and interpretation of scRNA-seq cancer data and efficiently identifies transcriptional signatures linked to known biological pathways. The CanSig method is available as a documented Python package at . Statement of significance CanSig is an intuitive computational approach to detect shared transcriptional states across tumors and facilitate exploratory analysis of single-cell RNA sequencing data. ### Competing Interest Statement The authors have declared no competing interest.