6-27-2005
Seattle

Update from the Institute for SystemsBiology

Work completed:

We have completed 60-72% of the work we plan to process in this phase of the human proteome folding project. That means we’ve processed a huge amount of work that when combined and post-processed here at the ISB results in 3D-structure predictions for over 53,000 protein domains taken from proteins in all complete genome sequences. Again, these domains are the least annotated sequences out there, so this info should be a major source of insight for figuring out what these proteins do. We have two grids devoted to this project (one operated by IBM, the worldcommunitygrid.org, and one operated by United Devices, grid.org). So far IBM has returned 49 batches of work resulting in 49,000 structure predictions, while UD has returned 14 batches. We are projected to finish the work in december at this rate (the grid has grown a bit slower than we thought it would, so tell you’re friends).

Biology interface:

We have begun the final stage of development for this project, the front end that will display results to biologists. Two programers at the ISB have now begun the process of integrating the results we get from the grid with the myriad other types of biological data using the data visualization and integration platform Cytoscape (cytoscape.org). The functional implications of any given structure prediction are best interpreted in light of everything else we know about a given protein. This front end will allow biologists to efficiently evaluate not just the structure and sequence matches for a given protein, but also its context in the cellular network. Work will begin on a subset of 5 organisms including yeast, human and malaria, but will eventually include all organisms on the grid ... more on this soon.


Richard Bonneau
Sr. Scientist
ISB, Seattle