Abstract #30

Chemogenomic profiling reveals functional relationships in yeast. Maureen Hillenmeyer^1,2, Russ Altman^2,3, Ronald Davis^1,3, Corey Nislow¹, Guri Giaever¹. 1) Stanford Genome Technology Center; 2) Program in Biomedical Informatics; 3) Department of Genetics, Stanford University.
   Chemogenomics, the use of small molecules in high-throughput genomic experiments, is a powerful tool for identifying drug targets and understanding gene function. We have used this approach to profile ~400 diverse compounds against the complete yeast heterozygous and homozygous single-gene deletion collections. In these assays we measured the growth phenotype (termed "fitness") of each deletion strain in the presence of each compound, to define the gene’s "druggability;" in this context, we explored the limits of the yeast druggable genome. We then considered each individual gene’s fitness profile, or the pattern of growth phenotypes of the corresponding deletion strain in the various compounds. In these terms, we defined "co-fitness" of genes as the degree of correlation of their fitness profiles (i.e. the similarity of growth phenotypes across compounds). We found that genes with a high degree co-fitness are often functionally related; we used these co-fitness relationships to successfully predict gene function. In comparing co-fitness to co-expression, we found that the two assays have different types of interactions and gene functions that they best predict. Our methods and results are relevant to phenotypic profiling in other organisms.

Return to YGM 2006 Home at SGD