Structure Prediction Protein Design Biclustering + cMonkey Network Inference Publications          Links + Collaborations People                   Opportunities             Software + Code Teaching            

 

WCG Status Update

WCG Post

HPF2 Update - July 2009

Greetings WCG Volunteers,

We've been working diligently since the last update to compile results for a grand paper describing the research involving WCG protein simulations. We've been folding proteins on the WCG for awhile now, and a handful of papers have been published utilizing our results; but we haven't yet detailed our methods and results in full. As the summer begins, we will be focusing much of our attention on compiling statistics on the accuracy of our structural classifications and molecular function predictions. After we submit the publication for review we will turn our attentions toward improving access to our database so that our predictions can be intuitively used by the community.

In order to properly categorize the accuracy of our structure predictions, we rely on the expansion of the RCSB Protein Data Bank (http://www.rcsb.org/pdb/) and updates to the manually curated SCOP: Structural Classification of Proteins database (http://scop.mrc-lmb.cam.ac.uk/scop/). We only fold proteins that don't have recognizable sequence homology to proteins in these databases, but unless these structures are eventually solved it would be very difficult for us to understand the quality of our predictions. Occasionally, however, a new protein is submitted to the PDB, classified in a SCOP release, and happens to be identical to something we've folded. With proteins like these we can not only examine the quality of the 3D models predicted, but also the accuracy of our predicted classifications. Much of that work had been finished when the SCOP group released version 1.75 in June. While this means we need to re-run much of our analysis, it also means we have a larger set of samples for our statistics.

As an example, here's a protein whose classification we correctly predicted. The structure folded as work unit "bt386" comes from the organism "Thermotoga maritima MSB8" and is deposited in the PDB as 2nzc.


Click here or the image for full size

After we finish this work and submit our publication for review, we'll shift focus towards the database and user interface. We would like to make our data as accessible and user friendly as possible. By integrating our predictions with BioNetBuilder (http://err.bio.nyu.edu/cytoscape/bionetbuilder/), a researcher examining protein interactions in particular processes or developmental cycles in organisms will have access to novel predictions about the character of the proteins in question.

We have a lot of work ahead of us to get together all of the analysis and highlights for the coming publication, but it could not match the collective contribution from the WCG. We'll be working hard to keep the research supported by members relevant and important.
--
Patrick Winters
Bonneau Lab


For those of you interested tracking work units, I've updated our Experiments table (http://homepages.nyu.edu/~rb133/wcg/experiments.html). We're folding the Rice proteome now, and will be doing so for quite awhile.