Examining approaches to target validation and drug repurposing in large scale genomic projects.

PhD Theses Medical.Drug repurposing presents an opportunity to quickly produce new medications in a cost effective manner. This is especially important in rare diseases where patients are frequently underserved. Here, we apply various methods to first select good targets for repurposing. We analyse loss-of-function (LoF) data, and assess its role in informing drug discovery. We achieve this by curating, aggregating and labelling LoF data and then building a model to predict genes that may harbour homozygous LoF with no negative associated phenotypes. We produce a model with a relatively high degree of accuracy and recall (F-score 0.7), generating 442 predicted genes in addition to 1,744 from aggregation. Following this, we assess whether such data could inform drug discovery in collaboration with AbbVie, an industrial partner. Through the study of historic drug data, comparing our LoF labels with data from previous studies detailing the effect of genetic knowledge on drug discovery, and against the loss-of-function observed/expected upper bound fraction (LOEUF) score, a metric of constraint, we demonstrate that this data adds significant value to drug discovery. Finally, we build a database focussing on rare diseases, and use LoF data, in addition to drug data and expertly curated gene panels to nominate candidates for repurposing. This database will be made available for researchers within the GEL community, such that avenues for repurposing can be further explored