! some Q/A about SnugDock
ref: [http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000644|SnugDock: Paratope Structural Optimization during Antibody-Antigen Docking Compensates for Errors in Antibody Homology Models]
Q: questions from Jarod ; A: answers from Aroop Sircar and Jeffrey J. Gray .
Thanks for the quick replies from Aroop and Jeffrey.
__Q__: As you mentioned in the paper, you rank the docking decoys by the
interface energy scores rather than the total score, how could I read
the interface energy from the score file "aaFR02.fasc"?
__A__: The column header for interface score is "I_sc"
__Q__: In the first paragraph of section 'Global docking', it said "The
starting structures consisted of the unbound crystal structure of the
antigen and the lowest-interface-energy RosettaAntibody homology
model."; But in the other parts of this paper, the
starting structures of antibodies are "lowest-energy RosettaAntibody
homology model" rather than "lowest-interface-energy" ones, and I
don't find the column header "I_sc" in the score file generated by
__A__: It should read lowest-scoring instead of lowest-interface-scoring.
__Q__: can Ensemble+SnugDock run in centroid mode for global docking?
__A__: SnugDock is only applicable in full atom (high-resolution) mode.
__Q__: Have you tested the SnugDock for the complex where the antigen is
very small like a peptide? how do you treat the water molecules during docking?
__A__: We have not benchmarked against peptide targets because, as you note, they have some different physics such as flexibility and more importance for water-mediated hydrogen bonds. Rosetta does not have any tested means for including explicit waters. For peptide docking, I'd refer you to a paper on HIV-protease by Sid Chaudhury in my lab, and a new paper by Ora Furman's lab on peptide docking with Rosetta.