Cancer cells arise from normal cells, and may retain transcriptional patterns present in the cell of origin. Recent advances in single-cell transcriptomics have allowed us to profile both normal and cancer cells at a single cell resolution, enabling direct comparisons between the two. Such comparisons can reveal information about which cells cancers originate from, transcriptional programs that underpin carcinogenesis, and differentiation states in cancer. The work presented in this thesis aims to study normal cell signals in human cancer, utilising single-cell transcriptomics. Chapter 1 provides an overview of transcriptomics as a whole, as well as in the context of cancer, with a focus on single-cell transcriptomics. In Chapter 2, I construct a single-cell normal reference of the fetal adrenal gland, and explore the development of the fetal adrenal medulla, from which the childhood cancer, neuroblastoma, arises. I then compare that reference to neuroblastoma single-cell and bulk transcriptomes in chapter 3. This comparison reveals sympathoblasts as the normal correlate of neuroblastoma cancer cells, and allows identification of transcripts present in both neuroblastoma and fetal adrenal medulla, but not adult tissues, making these transcripts an attractive therapeutic target. Chapter 4 presents a comparison of fetal and adult references across three different organs (gut, lung and liver) to bulk and single-cell transcriptomes of adult cancers originating in these organs. This analysis assesses the contribution of fetal and adult cell type signals to cancer transcriptomes. Overall, my work delineates the transcriptional relationship between cancer and normal cells.