1) Annotation of the Intergenic Regions of the Human Genome, focusing on Tiling Microarrays and Pseudogenes

2) Large-scale Prediction of Protein Function, in terms of Molecular Networks

3) Aspects of Structural Genomics:
Target Data Mining, Membrane Protein Interactions, and Analysis of Protein Motions in terms of Packing

In the first area, we had a number of major accomplishments. We identified all the pseudogenes in the mouse and human genomes and compared them. We were also part of a large-scale collaborative effort endeavoring to tile the entire human genome onto microarrays and look for all the transcribed regions. In the second area, we created an analysis of the yeast genome that coupled the dynamics of gene expression to regulatory networks, allowing us to discover transient hubs and the way regulatory network is "rewired" in different cellular states. We also developed rigorous method for mapping protein-protein and regulatory interactions between organisms and a practical web tool, called TopNet, for performing many molecular network calculations. In area three, we developed a methodology for describing structural flexibility in terms of Hidden Markov Models, a sophisticated mathematical formalism. We also developed a way of mining the structural genomics targets to identify those that are most suitable for high throughput structure determination. Finally, we embarked on new research directions, identifying particular aspects of database interoperability related to computer security.

For the following year we have a number of goals. We would like to inter-relate our tiling arrays with our pseudogenes to search for the elusive partially functional pseudogene -- the transcribed pseudogene. We would like to generalize our tool for molecular networks to encompass the prediction, as well as analysis of networks. We would like to extend our molecular motions database to better take into account the way packing constrains motions.

Excel Publication Matrix

Publication Grid