From fertilisation onwards, the cells of the human body continuously experience damage to their genome, either from intrinsic causes or from exposure to mutagens. While the vast majority of DNA damage is repaired and the genome is replicated with extremely high fidelity, cells steadily acquire single nucleotide variants throughout life. Since cells pass these genetic changes on to their descendants, mutations shared between any two cells therefore imply a shared developmental path. In essence, these somatic mutations connect all cells together into one large phylogenetic tree of human development with the zygote at the root.

Reconstructing phylogenies of human development requires readouts of somatic mutations present in single cells. Recently, low-input whole-genome sequencing following laser-capture microdissection has allowed us to reliably call somatic mutations in distinct single-cell derived physiological units, such as colonic crypts and endometrial glands, while retaining spatial information on a microscopic level. In this way, I reconstructed large-scale phylogenies of cells from many different organs of three individuals. These phylogenetic trees recapitulate the early stages of embryonic development and asymmetric cell allocation in the blastocyst, as well as later clonal expansions such as benign prostatic hyperplasia and neoplastic polyp formation.

In a similar way, I also used somatic mutations to investigate the emergence of paediatric cancer, which is thought to be closely linked to aberrations in development. In the context of phylogenetic analyses of tumours, mutations shared between childhood cancers and different normal tissues can shed light on the embryonic lineage of tumours and may reveal the precise juncture at which tumours began to form. Accordingly, I studied the origin of Wilms tumour, the most common childhood cancer of the kidney. I discovered that these tumours often arise from large tissue-resident precursor clones residing in the normal kidney. These embryonal precursors represent an early clonal expansion driven by H19 hypermethylation.

Lastly, using somatic mutations I discovered that the human placenta is made up of large clonal patches of closely related trophoblast cells. Comparing early embryonic mutations between placental lineages and umbilical cord DNA, which is derived from the inner cell mass, revealed that in approximately half of the cases, a trophectodermal lineage shares no somatic mutations with the umbilical cord. Furthermore, in a quarter of cases, the umbilical cord is entirely derived from a progenitor later than the zygote. This indicates a natural early segregation between these lineages and a pathway to generate confined placental mosaicism.

This dissertation as a whole provides a new framework to study normal and aberrant human development from whole-genome sequencing. The ability to reconstruct developmental lineages retrospectively can answer fundamental questions about human development and carcinogenesis.