Predicting Synthetic Lethality

Sharyl L. Wong1, Lan O. Zhang2, Amy H. Tong, Debra S. Goldberg, Oliver D. King, Guillaume Lesage, Marc Vidal, Brenda Andrews, Howard Bussey, Charles Boone, Frederick P. Roth
1sharyl_wong@student.hms.harvard.edu, Harvard Medical School; 2lan_zhang@student.hms.harvard.edu, Harvard Medical School

Synthetic lethality strongly supports the existence of genetic buffering. It arises when mutations to a set of genes cause cell death, while mutations to any subset of these genes do not. Currently comprehensive identification of synthetic lethal relationships in yeast and particularly in higher organisms is largely infeasible. Therefore, predicting synthetic lethality may expedite identification of redundant genes and pathways that buffer an organism from the phenotypic consequences of genetic mutation. Integrating multiple data types including co-localization, correlated mRNA expression, physical interaction, protein function, and sequence homology, we constructed a probabilistic decision tree with which we successfully predicted synthetic lethal gene pairs in Saccharomyces cerevisiae. Furthermore, the gene pair characteristics important in generating predictions may better our understanding of genetic robustness.