Improved understanding of the functional interdependencies between the cell's signaling and transcription machinery has profound implications for human health. Dysfunction of this machinery is already implicated in cancer (e.g., transcription factor myc in leukemia, Aurora B histone kinase in colorectal cancer), but a real barrier to further progress is that we do not yet know how the basic machinery works in a simple cell. By comprehensively characterizing and comparing the functional connections between these sets of genes across three eukaryotes broadly positioned across the tree of life (budding yeast, fission yeast, and mouse), more rapid progress can be made toward identifying and potentially even re-programming these defects in human cells, because the components of these processes are widely conserved.
Genetic interaction mapping is an emerging methodology that provides a powerful tool for understanding cellular function. However, despite all of the exciting prior work in this field, several points stand out as remarkable: First, almost all networks to date have been examined under a single static (usually standard laboratory) condition. Biological networks, however, are highly dynamic entities that continuously respond to a host of environmental and genetic changes or are altered more slowly over an evolutionary period. Second, such network changes occur across a wide range of scales, some impacting individual gene interactions, others affecting protein complexes, still others best summarized at the level of broad cell processes. This project addresses these two core issues (network dynamics and scales) by studying how a genetic network is reconfigured under complex species and stresses and by performing these comparisons using tools that capture a network's multi-scale, modular architecture. In the previous funding period, we successfully focused attention on generating genetic interaction maps across two yeasts, S. cerevisiae and S. pombe. We now pursue this analysis in mammals by furthering methodology to generate quantitative genetic networks using combinatorial RNAi in mouse fibroblast cells. Also, in the previous period we developed an approach termed dE-MAP (differential E-MAP), which allowed for the comparative analysis of yeast genetic interaction maps generated across different exogenous stresses. We now propose to leverage this analysis to create the first dE-MAPs in mammals, allowing us to study the evolution of stress response networks across the eukaryotic lineage. Finally, we will further develop a novel systems biology framework to use genetic interaction maps to drive the creation of a data-driven Gene Ontology (GO), which ultimately feeds back to design of subsequent E-MAPs and iterative refinement of GO. This work will be carried out over three Specific Aims: (1) Improvement and scale-up of a platform for high- density quantitative genetic interaction mapping in mammals;(2) Generation of differential genetic networks in response to stress in mammals and yeasts;and (3) Development of a computational methodology called network-extracted ontologies to enable multi-scale modeling and comparison of genetic networks. Successful completion of these aims will generate important community resources, including a large database of genetic interactions and modules in three different eukaryotic species and further development of network-based ontologies as a new multi-scale approach for network analysis. Because this project will involve close coordination between the Krogan and Ideker laboratories, this grant invokes the co-Principal Investigator mechanism. It joins two University of California campuses as well as two California Institutes for Science and Innovation, QB3 and Calit2, which support the PIs in the San Francisco and San Diego areas, respectively.