Biomolecular Pleiomorphism Probed by Spatial Interpolation

Author: Mirabela Rusu

Primary Advisor:

Committee Members: Stefan Brimanns, PhD (co-author); Willy Wriggers, PhD (co-authors)

Masters thesis, The University of Texas School of Health Information Sciences at Houston.

Abstract:

Multi-resolution modeling is an essential tool in structural biology as it provides an atomic interpretation of low-resolution experimental data. A variety of rigid-body docking techniques were developed, enabling the user to dock an atomic structure into a low-resolution volumetric map. In concurrent high-resolution EM reconstructions one can observe conformational rearrangements that limit the usefulness of pure rigid-body docking solutions. Only a flexible registration of the structural data can provide an accurate model of the system imaged by the electron microscope. Coupled with molecular dynamics simulation techniques, feature- based reduced models were successfully employed to capture the conformational change between the multi-resolution data sets. The complexity of the calculation and the dependency on an accurate parameterization of the biological system prompted us to search for a more efficient alternative to simulation techniques. Global interpolation methods are utilized in other areas such as image-processing and visualization to describe the overall deformation of an object. Combined with a coarse-grained representation of the biological system such methods can provide a flexible representation of the molecular structure. We have compared a variety of interpolation techniques and evaluated the results by comparing with constrained MD simulations. One method, inverse distance weighting interpolation, produced clearly the best results in our test cases. As an alternative to simulation techniques, the spatial interpolation methods are better adapted for non-expert users, as little parameterization is needed, and are suitable even for very large systems. Moreover, our results show that the overall conformational change is captured accurately by these very efficient global techniques.