Computational Structural Biology

Even though there are fast development in the competing technologies, such as NMR and de novo prediction, X-ray crystallography is still the single most trustworthy method to obtain high resolution protein structures. Recently the high-throughput screening made the search for the crystallization condition much easier, but the growth of crystals remains a bottleneck and depends very much on trial-and-error. Our goal is to use computational methods to select the best mutation site on the protein surface that increases crystallization propensity, by using only low resolution or predicted structures, and the core problem is the prediction of crystal packing using any given structure.

Our method is based on a combination of data mining and multi-molecular docking technique. The data mining flavor of the method focuses on discovering geometrical and physiochemical patterns of surface features that are responsible for forming crystal lattice, such as the preferred amino acid, secondary structure (e.g. half-Leucine zipper), metal mediator, or even complex topology (e.g. O-ring). It's worth noting that crystal contacts differ from those that are specific and plays functional roles in nature which can also appear to be crystal contacts. Although statistics shows qualitative difference in contact area between the two types, it is not a trivial task to distinguish them.

 However, no matter which kind of interaction the contact is involved in, the same physical-chemical forces are at play. This makes molecular dynamics a good complementary solution.  The problem of searching for optimal crystal packing becomes searching for the energetic minima of conformational space under constraints of crystallographic symmetry. In practice, rigid body docking is used in advance to quickly produce candidate conformations before MD or other flexible docking methods are used as refinement.


Login Form

Institut für Molekulare Biotechnologie Copyright © 2019. All Rights Reserved.

Please publish modules in offcanvas position.

Free Joomla! templates by Engine Templates