Rosetta rebuilding...
Phenixers, I'm curious about using rosetta to rebuild an existing model (phenix.mr_rosetta it seems) [0] What is being rebuilt in rosetta? The whole model? Or is the model divided into chunks and each chunk sent to rosetta? [1] Can a specific region be chosen to be rebuilt? (In the case where most of the model is correct) [2] How are the rosetta models scored ? Specifically, "(d) refine Rosetta models, average density from top 20%, continue rebuilding each Rosetta model using averaged density". What's the scoring function (mapCC, R-Free, LLG??) How's the top 20% obtained? Thanks! F --------------------------------------------- Francis E. Reyes M.Sc. 215 UCB University of Colorado at Boulder
Hi Francis, [0] What is being rebuilt in rosetta? The whole model? Or is the model divided into chunks and each chunk sent to rosetta? The model is being divided into chains, and each chain is rebuilt individually by rosetta (along with an electron density map as part of the rosetta target function). If there is NCS present, then only one copy is rebuilt with Rosetta and it is copied to the positions of all the other copies. [1] Can a specific region be chosen to be rebuilt? (In the case where most of the model is correct) No, not at the moment (I know this would be useful!) [2] How are the rosetta models scored ? Specifically, "(d) refine Rosetta models, average density from top 20%, continue rebuilding each Rosetta model using averaged density". What's the scoring function (mapCC, R-Free, LLG??) How's the top 20% obtained? Rosetta models are scored by LLG in this process. All the best, Tom T
On Mon, Jan 16, 2012 at 8:33 AM, Francis E Reyes
I'm curious about using rosetta to rebuild an existing model (phenix.mr_rosetta it seems)
Do you mean that the model is already correctly placed and partially refined? My guess is MR-Rosetta is going to be too coarse a method for this task. Frank DiMaio in the Baker lab and I are working on actual refinement starting from a broad range of initial models, but it's probably going to be a while before we have anything ready to use. If you think you have a good test case, I'd be interested in seeing it. -Nat
participants (3)
-
Francis E Reyes
-
Nathaniel Echols
-
Terwilliger, Thomas C