|
Abstract:
|
Failure of drug candidates in late-stage development due to safety issues or previously undiscovered side effects is one of the causes of steadily declining R&D efficiency in the pharmaceutical industry. This triggers the -literally- million dollar question which compounds should advance through early phases, in particular during lead optimization, based on limited data available at these stages. One way to address this question and reduce late stage failure can be the use of high-throughput techniques for measuring holistically transcriptional effects of compounds. Identification of compound-induced perturbations in transcriptomics networks with the aim of understanding biology and mechanisms of action is an application that has already been demonstrated in the literature. However, the utility of gene expression profiling for decision-making in early-stage pharmaceutical drug discovery has not yet been demonstrated. This presentation will illustrate to what extent decision-making during lead optimization can be assisted by exploring relations between biological, chemical and gene expression data, and will introduce some of the main statistical challenges accompanied with it.
|
