The human genome was first sequenced in 2003, leading to many discoveries. However, the promise of personalised medicine has yet to be realised. In this thesis, I outline key technical and conceptual challenges that have contributed to this mismatch of expectations. I then introduce the three Aims of this thesis that were designed to substantively address these challenges:
Aim 1. Discover the cell types underlying rare disease phenotypes.
Aim 2. Decompose the phenome into latent genetic factors.
Aim 3. Demonstrate and facilitate FAIR practices.
Together, these results provide new insights into the cell type-specific mechanisms of every disease for which genetic data is publicly available (n>40,000). This includes both rare and common diseases, as well as each of their constituent phenotypes. Furthermore, I demonstrate numerous actionable avenues that can be taken to apply these results to personalised patient diagnosis, prognosis and treatment, as well as the development of novel gene therapies optimised for high efficacy and low toxicity. Finally, all analyses to generate this work are included within the thesis itself as a programmatically reproducible manuscript:
<https://bschilder.github.io/thesis/inst/docs/>