Chemistry Reference
In-Depth Information
independently (antibody, antigen complexes).
They are therapeutic drug targets
and the mechanism of protein association is of considerable interest to drug
discovery.
The two main problems of protein-protein docking methods are the following.
The first is the 'conformational search' used to generate the correct conformation
of the two proteins together. Target identification is the second most difficult
problem. How do we recognize a native conformation among millions of
generated configurations? The increase in computational power allows us to
generate many configurations (including some close to the native conformation).
It is noteworthy that the individual proteins can also undergo conformational
change upon binding.
Scoring functions have been used to solve the problem of recognizing a native or
near-native conformation (among millions generated). For the conformational
search, most methods use a two-stage process. A rigid-bod global/local search is
done in the first step in order to determine the relative position and orientation of
the two proteins. In the second step there is account for side chain and backbone
flexibility
via
refinement of the configurations. In the absence of information on
the complex, fast Fourier transform (FFT)-based algorithms and geometric
hashing can be used to identify binding modes. These approaches use grid point
and scan the space by translating and rotating one grid about the other. A
correlation-type scoring function is used to evaluate the docked configurations on
a grid.
When geometric hashing is used, we describe the protein surface by critical
points. These are used to detect geometric complementarity. Genetic algorithms
and Monte Carlo (MC) simulated annealing are used to explore the potential
regions. In these methods the search is confined to the starting structure. It may be
possible to miss the correct conformation. Restraint-based docking can be used to
direct the search.
Steric overlap is often included in the thousands of complex structures generated
by rigid-body docking. Backbone and side chain refinements are subsequently
required. One way to do this refinement is to use energy minimization (atomic
