Also in collaboration with the Zaki lab,
we have developed a data structure for protein surfaces, capable of
identifying the correct complementary interaction surfaces in a known
protein-protein interaction -- i.e. the protein docking problem.
A boolean data structure has been defined with an eye towards
developing a heiarchical search engine for protein surfaces.
Ongoing research involves adding flexibility into these surface structures.
This follows on previous development of a protein surface calculator,
MASKER, which also uses Boolean masks.
Protein-protein docking prediction from ContextShapes.
The best ranked hit is shown in black, the original unbound ligand is shown in red, and the rank 1 prediction is shown in blue.
click for a closer look