MantisBT - Rosetta
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0000323Rosetta[All Projects] Incorrect Resultspublic2014-02-04 10:412014-02-11 14:14
Reporterjadolfbr 
Assigned Tojadolfbr 
PrioritynormalSeverityminorReproducibilityalways
StatusassignedResolutionopen 
PlatformAll platformsOSAnyOS VersionAny
Product VersionTrunk 
Fixed in Version 
Application(s) Affectedmany
Command Line UsedNA
Developer Options
Fixed in SVN Version
Summary0000323: scoring/sasa.hh reports values by default using radii that were optimized for a no-longer-used scoreterm
DescriptionIn the function calc_per_atom_sasa, 3 radii can be used for the calculation. These include the Rosetta default, LJ radii, and those from the program Naccess. The default is to use the Rosetta default - radii that were explicitly optimized for use in a scoreterm that is no longer used. I don't think the scoreterm was ported to Rosetta3.

This is fine, but many programs and classes (perhaps blindly) use these default radii and should probably use either the LJ or the standard Naccess radii instead as it seems using the Rosetta defaults can produce SASA values differing by 0000011:0000005-10% from LJ/Naccess values.
Steps To ReproduceUse core/scoring/sasa/calc_per_atom_sasa using default setting. Or use the SasaCalculator pose metric. Even better, use pretty much any of the functions within sasa.hh that call calc_per_atom_sasa using default values.
Additional InformationChanging the defaults would change numerous pieces of Rosetta and I'm not sure I'm ready to do that. However, I have reorganized the code and commented it heavily so that this should be known. In reality, sasa.hh should probably be its own class either within core/scoring or as a simple mover with LJ or Naccess radii being the expected defaults.
TagsNo tags attached.
Relationships
Attached Files
Issue History
Date ModifiedUsernameFieldChange
2014-02-04 10:41jadolfbrNew Issue
2014-02-04 11:59jadolfbrNote Added: 0000328
2014-02-11 14:14jadolfbrAssigned To => jadolfbr
2014-02-11 14:14jadolfbrStatusnew => assigned

Notes
(0000328)
jadolfbr   
2014-02-04 11:59   
In addition to this, using Naccess defaults with all atoms will result in the wrong values of SASA due to the Naccess radii being that from Chothia 1976 ('The Nature of the Accessible and Buried Surfaces in Proteins'):

"Hydrogen atoms were not considered individually but included in the van der Waals radii used for the non-hydrogen atoms. The van der Waals radii used, oxygen 1.4 A, trigonal nitrogen 1.65, tetrahedral nitrogen 1.5, tetrahedral carbon 1.87, trigonal carbon 1.76, sulphur 1.85, and water 1.4, were derived from the intermolecular contact distances observed between non-hydrogen atoms in crystals of amino acids"