About
Exponentially increasing costs of pharmaceutical development – a phenomenon known as Eroom’s law – is a call for innovation. Here, we propose to integrate scattered genetic information on drug targets to guide their selection. During the past decade, the statistical genetic community discovered and replicated thousand variant-trait associations in human populations. Exploring this wealth of data, investigators recently revived the initial hopes for the potential of genetics to enhance drug development: drug targets with genetic evidence for efficacy (i.e. the target gene is linked to a variant significantly associated with the drug indication) have twice as much chance of success in a clinical trial. So far, previous efforts focused on European populations and sex and age aggregated data. In other domains of statistical genetics, such as polygenic risk scores, this sampling bias leads to a loss of accuracy in non-European ancestries. There is also mounting evidence of varying drug effects between sex and age strata. Yet, women and elderly are under-represented in randomized control trials. To document the genetic component of these drug effect variations and address these sampling biases, we will collect sex and age disaggregated data spanning all ancestries. Building upon this representative database, we will design a predictive algorithm to predict drug target efficacy from genetic information. The developed algorithm will be applied on all drug targets yet to be explored by pharmaceutical companies (which represent 2/3 of the 4479 druggable genes). In short, this project aims at summarizing genetic evidence for a drug target potential into a score relevant for all human population. By facilitating the prioritization of drug targets, we hope to transform Eroom’s law back into its well known reverse: Moore‘s law.