For validation, Luo and co-workers employed a sensitive multicolor competition assay and could confirm SB431542 ∼25% of the primary candidates, most of which displayed specificity toward KRAS mutant cells in a second, albeit

related pair of isogenic lines. Strikingly, with the exception of KRAS itself, none of these genes had been described as an oncogene, supporting the authors’ previous hypothesis that focusing on ‘non-oncogene addiction’ may offer a broad set of promising novel drug targets [ 28]. Instead, the list of KRAS-synthetic lethal interactors included regulators of mitosis (e.g. the kinase PLK1 (Figure 2)), ribosome biogenesis and translation, sumoylation and RNA splicing. The researchers therefore hypothesized

that KRAS oncogene activation may lead to generally increased levels of mitotic stress and predicted that small-molecule inhibitors further disrupting cell division would specifically affect cancer cells. Indeed, clinically approved or experimental Dasatinib concentration inhibitors of cell division selectively impaired the growth of KRAS mutant cells at low doses both in vitro and in xenograft models of cancer [ 26••]. The number of isogenic cell lines available from commercial or academic sources is growing quickly, enabling comparative high-throughput experiments focusing on many genes, pathways and phenotypes [29, 30• and 31]. Yet, cancer cell lines frequently display genomic instability and the targeted modification of individual loci and the subsequent establishment

of cell lines involves stringent selection procedures. Researchers therefore need to carefully evaluate the degree of genetic and phenotypic similarity between cells originally derived from the same paternal line. Significant interactions between loci observed in a specific genetic background can catalyze novel mechanistic insights; their relevance for drug development, requires validation in a broad panel of genetically diverse model systems. The systematic, high-throughput analysis of genetic interactions in mammalian cells has only recently become feasible. Yet, suppressor-screens and enhancer-screens have long been a genetics staple in model organism such as yeast [32], C. elegans [ 33 and 34••] or Drosophila [ 35]. In particular, yeast Osimertinib geneticists have embraced the growth and viability of cells as a general proxy for organismal fitness, a complex quantitative phenotype, and constructed comprehensive interaction maps by systematically generating (nearly all possible) double-deletion mutant combinations [36••, 37•• and 38]. Besides identifying individual synthetic lethal gene combinations, the systematic assembly of hundreds of interaction profiles into large data matrices has enabled powerful correlative analyses to delineate the complex functional networks underlying cellular processes [36••, 39, 40, 41, 42• and 43].