Yeast orthologue mapping predicts genetic vulnerabilities in tumour cells
Session type: Poster / e-Poster / Silent Theatre session
Two major challenges for the development of oncology drugs are: defining patient populations that are likely to benefit from single agent targeted therapies, and the identification of drug combinations that can overcome intrinsic and acquired resistance. Genome-wide genetic screens in combination with established drugs have begun to address these issues. However, they are to some extent limited by drug polypharmacology, which may confound results. Additionally, they cannot currently address all pairwise gene-gene interactions in the mammalian genome. In contrast, there is a wealth of genome-wide synthetic lethal interaction data in model organisms such as yeast. Despite their evolutionary distance from higher eukaryotes, many core metabolic and signalling pathways are highly conserved from yeast to man.
We ran computational analyses of yeast synthetic lethal interactions and performed multi-organism comparisons to predict pathway-level genetic vulnerabilities in cancer. We also retrospectively analysed current clinical combination trials to determine whether they could have been predicted by interrogating pre-existing datasets.
Our analysis identified several targets for which drugs are available, and that are already being used in combination in preclinical and early phase trials. We also identified unexpected synthetic lethal interactions between metabolic and signal transduction pathways.
The largest pairwise synthetic lethal screens run in human cells to date still only cover a few hundred genes. While the scale will increase as technology improves, we suggest that in silico analysis of model organism synthetic lethal data can be exploited to predict novel vulnerabilities in cancer, and to identify novel drug combinations.