Protein folding using Monte Carlo derived techniques and implicit solvation

[Anonymous]

I am looking for publications where Rosetta has been used to successfully fold proteins from sequence and where implicit solvation model has been necessary, could you give me some references?
Thanks

[Anonymous]

References where an implicit solvation model has been necessary? I’m not sure any Rosetta publication will treat the necessity of using implicit solvation models. All Rosetta calculations are done with implicit solvation. Rosetta isn’t set up to do explicit solvation, and comparing implicit versus explicit solvation isn’t really something that a Rosetta-based protocol would do.

The best you’re likely to come up with regarding the necessity of implicit solvation is to find some (non-Rosetta) references which talk about how a particular protein is hard to fold with explicit solvation, and then hope to find a Rosetta paper which shows how the implicit solvation Rosetta model can handle it. (The CASP exercises might be good for that – you could find a group using explicit solvation models, look at the targets they do poorly on, and then see if Rosetta does better.) That’s comparing apples to oranges, though. Rosetta takes a completely different approach to folding proteins than does a molecular mechanics based approach (which is likely what you’d get for a method using explicit solvation). It would be hard to say if it’s the implicit solvation per se which makes the problem tractable, or if it would be Rosetta’s fragment-based approach instead.

Perhaps if you elaborated on the reasoning behind your request I could think of some suggestions which might be useful.

[Anonymous]

I mean, if you can give some folding examples where implicit solvation has been used (not explicit solvation), in terms of SASA plus sigma coefficients. Thanks

[Anonymous]

The model of implicit solvation that Rosetta uses isn’t a SASA based approach. It’s the EEF1 solvation method of Lazaridis & Karplus. (http://www.ncbi.nlm.nih.gov/pubmed/10223287) – at least that’s what’s used for the full atom stages. For the centroid-based stages, the implicit solvation is implicitly included in the (environmentally dependent) statistical potentials which are used for centroid scoring.

Regarding references for examples of that, I’d point you to any of the CASP papers which use Rosetta, which are coming out of David Baker’s lab. (Go to http://depts.washington.edu/bakerpg/drupal/biblio look for papers which mention CASP or folding.) As I mentioned before, the type of implicit solvation that Rosetta uses is somewhat baked in, and there really isn’t any papers which do extensive validation/comparision of the different solvation models used – at least none that I’m aware of which deal with protein folding.

[Author]

Posts