Kavraki Lab

Physical and Biological Computing

Keywords:

Abstract:

The problem of incorporating protein flexibility in the routine in silico screening
of large databases of small chemical compounds is still an unsolved and hard problem.
The main reason behind this difficulty is the exponential explosion in computational
complexity due to the inclusion of the large number of degrees of freedom that represent
the receptor flexibility. In order to address this limitation several flexibility models for
the receptor have been developed which try to limit the number of additional model
parameters. These models can be roughly grouped divided into five different categories.
These are the use of soft receptors which relax energetic penalties due to steric clashes,
the selection of a few critical degrees of freedom in the receptor binding site, the use of
multiple receptor structures either individually or by combining them using an averaging
scheme, the use of modified molecular simulation methods, or the use of collective
degrees of freedom as a new basis of representation for protein flexibility. All these
flexible receptor models strive to balance an improvement in the accuracy of the docking
predictions with an increase in computational cost. In addition, other challenges such as
the development of accurate solvation models and scoring functions make the receptor
flexibility problem even harder.