Storer Symposium on Computational Networks Biology, 10am - 6pm,Sept 25, 2006, Kleiber Hall

To develop therapeutics and agricultural breeding programs, geneticists often seek to identify nucleotide-level changes that cause phenotypes to vary across populations.  Genetic mapping at this level has proven to be a significant challenge, seemingly because most traits are controlled by a network of multiple loci, each with a modest effect.  As a result, geneticists understand relatively little about the molecular genetic principles that govern the impact of a polymorphism on a trait.  To study these principles on a wide scale, we began with a survey of the genetics of a large set of phenotypes—the expression levels of all genes in a cross between a lab strain and a wild strain of the budding yeast S. cerevisiae.  Using classical linkage mapping, we identified loci controlling thousands of transcriptional differences between the strains, and we experimentally confirmed the single-nucleotide differences responsible for large groups of transcript changes.  We analyzed the number of loci that underlie each trait, including a key estimate for the rate of non-additive genetic interaction between loci; here again we experimentally confirmed hypotheses about the molecular basis of interactions.  We also studied the molecular characteristics that distinguish functional alleles from silent ones, in a genome-wide analysis of polymorphisms that affect the expression of individual genes in which they lie.  And in recent work, we have assessed the role of mutations arisen in the laboratory on the differences—in DNA sequence, expression, and macroscopic phenotypes—we observe between yeast strains.  Overall, our results indicate that transcript levels in a single-celled yeast exhibit rich and complex genetics.  Ultimately, the insights from study of this model system may help us predict the impact of mutations on complex traits, including those in higher organisms.

With the appearance of large networks of protein-protein and protein-DNA interactions as a new type of biological measurement, methods are needed for constructing cellular pathway models using interaction data as the central framework.  The key idea is that, by comparing the molecular interaction network with other biological data sets, it will be possible to organize the network into modules representing the repertoire of distinct functional processes in the cell. Three distinct types of network comparisons will be discussed, including those to identify:

(1) Protein interaction networks that are conserved across species
(2) Networks in control of gene expression changes
(3) Networks correlating with systematic phenotypes and synthetic lethals

Using these computational modeling and query tools, we are constructing network models to explain the physiological response of yeast to DNA damaging agents.

Relevant articles and links:

Yeang, C.H., Mak, H.C., McCuine, S., Workman, C., Jaakkola, T., and  Ideker, T. Validation and refinement of gene regulatory pathways on a network of physical interactions. Genome Biology 6(7): R62 (2005).

Kelley, R. and Ideker, T. Systematic interpretation of genetic interactions using protein networks. Nature Biotechnology 23(5):561-566 (2005).

Sharan, R., Suthram, S., Kelley, R. M., Kuhn, T., McCuine, S., Uetz,  P., Sittler, T., Karp, R. M., and Ideker, T. Conserved patterns of protein interaction in multiple species. Proc Natl Acad Sci U S A.8:102(6): 1974-79 (2005).

Suthram, S., Sittler, T., and Ideker, T. The Plasmodium network diverges from those of other species. Nature 437: (November 3, 2005).

http://www.pathblast.org

http://www.cytoscape.org

Acknowledgements: We gratefully acknowledge funding through NIH/NIGMS grant GM070743-01; NSF grant CCF-0425926; Unilever, PLC, and the Packard Foundation.

Sequence polymorphisms affect gene expression by perturbing the complex network of regulatory interactions. Standard methods (e.g., Yvert et al., 2004) attempt to associate each gene expression phenotype with genetic polymorphisms.  This talk describes a novel probabilistic method, called Geronemo, which aims to understand the mechanism by which genetic changes perturb gene regulation. Geronemo automatically constructs a set of co-regulated genes (modules), whose regulation can involve both sequence variations and expression of regulators. By exploiting the modularity of biological systems, Geronemo reveals regulatory relationships that are indiscernible when genes are considered in isolation, allowing the recovery of intricate combinatorial regulation.  By incorporating both expression and genotype of regulators, Geronemo captures cases where the effect of sequence variation on its targets is indirect. We applied Geronemo to a dataset from the progeny generated by a cross between laboratory (BY) and wild (RM) isolates of S. cerevisiae. Geronemo produced novel hypotheses about genetic perturbations in the yeast regulatory network, including transcriptional regulation, signal transduction, and chromatin modification. Our global analysis highlights some of the key differences between the regulatory network of the BY and RM strains.  In particular, our results suggest that a significant part of individual expression variation in yeast arises from evolution in the regulatory regions and the coding sequences of a small number of chromatin modifiers.   Moreover, our analysis suggests an intriguing hypothesis, supported by subsequent wet-lab experiments, which elucidates a pathway associated with p-bodies, a recently discovered protein complex that helps regulate mRNA degradation.

Joint work with Su-In Lee, Dana Pe’er, Aimee Dudley, David Drubin, & George Church.

Biographical sketch

Daphne Koller received her BSc and MSc degrees from the Hebrew University of Jerusalem, Israel, and her PhD from Stanford University in 1993.  She joined the faculty at Stanford in 1995, where she is now an Associate Professor of Computer Science.  Her main research interest is in understanding complex systems involving significant amounts of uncertainty, with a primary focus on biological systems.  She is the author of over 150 refereed publications, which have appeared in venues spanning Science, Nature Genetics, the Journal of Games and Economic Behavior, and a variety of conferences and journals in AI and Computer Science.  She was awarded the Arthur Samuel Thesis Award in 1994, the Sloan Foundation Faculty Fellowship in 1996, the ONR Young Investigator Award in 1998, the Presidential Early Career Award for Scientists and Engineers (PECASE) in 1999, the IJCAI Computers and Thought Award in 2001, the Cox Medal for excellence in fostering undergraduate research at Stanford in 2003, and the MacArthur Foundation Fellowship in 2004.

Altered interactions within gene networks may underlie many interesting evolutionary transformations. To explore this possibility, our group is studying changes in protein:DNA interactions and resulting changes in gene expression in the context of human evolution. We carried out a genome-wide scan for positive selection on 5' flanking regions that are expected to be significantly enriched for cis-regulatory elements. Our results suggest that positive selection on cis-regulatory elements during human evolution has been widespread and has affected somewhat different biological processes and expression domains than adaptive evolution of coding regions. Genes scoring high in our analyses include disproportionately many genes involved in adaptive immunity and diverse aspects of metabolism, as well as genes associated with a variety of uniquely human traits. We have begun to characterize the functional consequences of evolutionary changes within the cis- regulatory regions of several genes associated with cognition and metabolism, focusing on mutations segregating within human populations and those that were fixed during human origins.

Over the past decade several biological mechanisms have been identified that allow organisms to produce relatively constant phenotypes in the face of both environmental and genetic disturbances. These mechanisms could evolve through natural selection acting directly on sensitivity to disturbance or could emerge as a byproduct of selection on other features of the system. Using evolutionary theory we can show that selection will not typically lead to the evolution of genetic robustness, but that environmental robustness can easily evolve. This theory predicts that environmental robustness will evolve in proportion to the fitness cost associated with individual phenotypes while genetic robustness will only evolve at loci that occupy particularly interactive positions in the gene network, irrespective of their direct fitness consequences. These ideas provide a testable hypothesis that is supported by gene expression data from yeast.