To elucidate the structure, function, and evolution of large genetic networks is the ultimate goal of functional genomics. Progress towards this goal will not only lead to a deep understanding of all living things, but also to profound insights into the origins of human disease. Recently, the first genome-scale data on a large genetic network has become available, the network of all pairwise protein interactions in the yeast Saccharomyces cerevisiae. This data provides the foundation of the proposed work, a structural and functional characterization of this network, and an explanation of its structure in terms of its evolution via interaction turnover and gene duplication.
Aim 1 is a global characterization of the yeast protein interaction network (YPIN). A graph theoretical framework will be developed and used to address questions such as the following: Is the YPIN a connected network? If not, how many subnets does it have? Does it belong to any known category of graph? Do interacting proteins or proteins in the same subnet have similar spatiotemporal expression patterns? Are genes whose products have many protein interaction partners more or less likely to undergo gene duplication? Aim 2regards the evolution of the YPIN by gene duplications. The majority of genes in the YPIN are members of gene families. At what rate do duplicate genes change their interaction partners or become associated with different subnets after duplication? Preliminary results show that the YPIN is structurally similar to a random network, and that the rate of turnover in protein interactions is high. They also show that it is possible to estimate the rate at which new interactions evolve.
Aim 3 finally integrates all this information to provide a mathematical model for evolution of the YPIN. This model will explain all structural features of the YPIN as a function of the rate of interaction loss and interaction gain, the rate of gene duplications in the YPIN, and the rate of interaction turnover after gene duplication. The information required in the model can be estimated from genomic sequence, gene expression, and protein interaction data. Geneticists and evolutionary biologists have speculated for decades on the structure and evolution of large genetic networks. The proposed work will put this speculation to an end for a key genetic network, some of whose conserved components are also involved in the etiology of human disease.
Showing the most recent 10 out of 27 publications
